Today at Umbra we released sample data for the first time through the Registry of Open Data on AWS.1

We’re starting small with 10 sites around the world updated multiple times a week. And we threw in a bunch of other stuff for fun including everything we’ve posted online to date (if we missed something, let me know).

All of the data is licensed the same way our paid data is licensed - openly. No login required to download. No talkin’ to sales people. No baloney (besides my tweets promoting it). All samples include detected data and most also include complex data and even phase history.2 

If you play around with the data and decide you want to task some fresh stuff, you can sign up to get on the waitlist for early access to our tasking platform, Canopy, here.

Oh, and did I mention the pricing is listed on the website?

To summarize:

1. Open data (I’ll)…

2. Open licensing (be)…

3. Self-service tasking (gosh)…

4. Transparent pricing (durned!)

The last few weeks of teasing our stuff on social media has led to a huge wave of interest from around the world, so please be a patient with us if you can. I promise we’re working as fast as possible to get people set up on Canopy.

If you simply must get your hands on some fresh microwave energy from Umbra, but you can’t stand waiting for direct access, then you’re in luck. Yesterday, SkyFi announced support for tasking our satellites at SXSW, so you’re actually only one swipe of a credit card away from tasking your first SAR collection: app.skyfi.com/explore/task

What’s the Big Deal?

If you don’t work in the satellite imagery industry you might read these opening paragraphs and wonder what the big deal is, because honestly, transparent pricing and open licensing are pretty boring topics. It’s all so obviously where the world is headed, and therefore, where this industry is headed. And yet…somebody’s gotta take the plunge. Might as well be us.3

I honestly, unabashedly, cringe-worthily believe Umbra is going to change the world. Half of that equation is our technology prowess, and to be fair, that’s a really impressive competitive moat.

However, in my opinion, the bigger deal is our courage to approach the market in a novel way. Some have called us unsavory names for taking this approach. I’ve personally been called naive and idealistic more times than I can count. I was only called an idiot once, that I can remember, but it left an impression. If you aren’t pissing off the old guard, what are you even doing?

The truth is that satellite imagery is invaluable for some of the most dire challenges that arise in the world, but it’s under-delivering on its potential. And fixing the business model is the first step toward reaching that potential.

Our strategy, in a nutshell, is to simply not jerk people around. And I think that in the coming years, if we can demonstrate that this strategy not only works but works splendidly, customers will demand better of our competitors and the contagion will spread. The result, I believe, will be a dramatic increase in the productivity of our small sector of the economy and a heck of a lot more benefit to society.

Today is an inflection point, but we are only just barely getting started. We aren’t out to become the best SAR provider - we are hellbent on permanently and irrevocably changing this entire industry for the better. 

From Whining About It, to Doing Something About It

Exactly three years ago, I made the difficult decision that I needed to leave the satellite imagery industry. I was fed up. I realized that I was wasting my time. My head was black and blue from banging it against the same wall so many times. 

For the prior three years, I had been pouring myself into Raster Foundry and Raster Vision, the open source satellite image processing and analytics tools we were building at Azavea.4 Unfortunately, they never took off the way we'd all hoped.

The biggest problem wasn’t a technical one - it was a cultural one. Satellite imagery providers were not equipped to support small clients; they refused to quote standard pricing, often tried to steal our customers, and occasionally even hired us to unwittingly help them steal customers from our competitors! That was especially messed up!5

When I thought I had found my next opportunity outside of the satellite imagery industry, I wrote an open letter of sorts—a breakup letter in my mind—called “The Commercial Satellite Imagery Business Model is Broken.” I figured it might help some of the suckers I was leaving behind.

The response was immediate and profound. 

I was saying the quiet part out loud - repeating cliched observations that industry insiders had been lamenting privately for years. It was the first thing I ever wrote that more than a dozen people read, and the response to it changed my life. 

Of the hundreds of people who reached out to me after that blog post, one of them was particularly persistent. Gabe Dominocielo, the founder of Umbra. He had to follow up with me multiple times, because I was so jaded by previous bad experiences that the first time I went to the website and watched the hype video, I got the wrong impression that it was all vaporware…6

I’ll be honest - even after I met Gabe, I was skeptical. What he was proposing sounded too good to be true (even for “naive” and “idealistic” me). Transparent pricing. Open licensing. API-first development. Essentially, every box on my wish list checked off - and he was even more adamant about it than I was.

I’ll never forget one of my first strategy meetings after joining Umbra. We were debating our licensing approach - I didn’t think they’d go for CC BY 4.0, so I suggested a custom, but very permissive license. And Gabe got so offended. He was furious that we’d even consider a half measure - were we going to flip this industry upside down or was I too chicken??

After that call that I realized I’d found my home. And I haven’t looked back a day since. 

Nothing you see today is the result of any one person. It’s a heroic effort that many dozens of people have worked on tirelessly - in at least a couple cases for almost eight years solid. A healthy portion of the people who built the software at Umbra were also contributors to Raster Vision and Raster Foundry - for a handful of us, this is a second act. And I’m grateful to work with those folks (and the many new close friends I’ve met since joining) every day.

Looking back, the last two years have been kind of incredible:

• Built, tested, and launched five satellites with a sixth set to launch in less than a month.

• Built an automated ground system

• Built Canopy, our self-service tasking and data delivery platform, on top of that ground system

• Built a set of internal tools for controlling the satellites, QA’ing the data coming off of them, and ensuring timely delivery to customers

• Rebranded the company (twice)

• Tripled in size from 30 to 90+

• Recruited and hired half of the C-Suite

• Onboarded our first customers, many of whom had been waiting patiently for years

• Signed a major partnership with Maxar, the leading satellite imagery provider in the world

• And as of today…released sample data for the first time. Pretty dang cool.

I honestly think this team is building a generational business…the kind of thing that I may only ever get one shot at in my career, as much as I hope to do this over and over again. 

There’s quite a lot of work still to do, and I’m not counting my chickens just yet. But, I can see the path ahead of us stretching many years into the future, and I’ll savor every moment of the journey while I can.

What’s next?

It’s all next.

Step 1 was fix the business model. Step 2 is grow the market together with customers using our data to build products that their customers don’t even need to know involves SAR.

Sure, the data we released is exciting to the tiny group of people that follow the commercial SAR industry. Internally, we call them the “First 100,” the most innovative remote sensing analytics firms on the planet. That’s our target audience.

And I hope those 100 organizations are excited. After all, the data is conspicuously high resolution compared to what people are used to, and it’s affordable compared to what others charge.

But for the rest of us, what we are announcing today are just necessary preconditions. We need a movement - a wave of entrepreneurs able to turn this raw material into finished products and services that normal people can get value from.7

It turns out the old hype video I watched when Gabe was first recruiting was real. We really can produce a ridiculous amount of high resolution data per satellite per day. What you see in the sample data folder is not some concerted effort - that’s mostly just what our marketing director does in his (extremely limited) spare time. There is a lot more to come. 

They say a rising tide lifts all boats. Prepare for the flood. Other SAR providers, optical satellite imagery companies, analytics firms, world governments, non-profits, Fortune 500 companies, academic researchers, anyone excited about applying remote sensing data to solve problems - we need all the help we can get to make the most of this new step function change in available, affordable, high resolution SAR. And there’s about to be a ton more of it available this year.

Depending on which category you fall into:

• Learn more about our sample data drop here!

• Get on the early access list to buy data from us directly here!

• Get access to task our satellites immediately from Skyfi here!

• Apply for one of our open positions here!

• Tell us what sites we should add to our recurring collections here!

Welcome to Part II of this riveting series about the impact that “internet in space” will have on the commercial satellite imagery ecosystem. Unfortunately, what I thought was going to only take two installments has now ballooned into three...as the saying goes, I would have made it shorter if I’d only had more time.

If you missed Part I, count yourself lucky. For those of you who feel like punishing yourself, be my guest:

A Closer Look with Joe Morrison

How “Internet in Space” Will Transform the Satellite Imagery Industry: Part I

Disclaimer: this is the perfect disclaimer. It emphasizes that I am speaking as an individual and not in an official capacity as a representative of my employer, Umbra. It notes my glaring financial conflict of interest as an active participant in the very industry I am commenting on. Most importantly, it’s in italics…

Read more

3 years ago · 21 likes · 13 comments · Joe Morrison

This is all the rehashing I’m gonna do on Part I: dozens of companies, including some of the biggest ones in the world, are building “relay” constellations that will enable realtime two-way communication with Earth observation satellites in just the next few years (demonstrations have been made, launches are planned, knives are being sharpened).

I’m not going to speculate on which of the many competitors entering the relay race are going to be successful.1 But I’m gonna assume at least a couple of them will be.

And when they are, I have three major predictions about what will happen:

1. There will be sweeping ground station provider consolidation by the cloud providers.

2. The concept of “direct tactical downlink” will go extinct.

3. Demand for edge processing in space will plummet to near-zero (with a notable exception).

The first and second of those will be covered in this installment; the third will be covered in Part III.

Death of a Ground Station Salesman

Back in my day, you used to have to sit around and wait for satellites to make contact with a ground station before you could give them new commands or get data and telemetry back.

But pretty soon, you won’t have to wait. Relay constellations with radios for commanding and high-throughput laser links for backhaul will soon dot the sky and enable information to travel from the ground to the satellite and back at close to the speed of light.

I always look to Uber and Lyft for how the satellite imagery industry should work. Think about it - the user experience of tasking a satellite to point at a place on Earth should be exactly like calling an Uber.

And if you look at how the ride sharing business has evolved, they will charge you less if you are willing to wait longer for your ride to show up. That’s how “rush delivery” of satellite imagery via relay will likely play out to start. Pay a little less, get your data down a little slower.

Instantaneous commanding, on the other hand, probably won't cost extra, since the data rates are so much lower and can be done via old school RF.2

My guess—and it truly is just a guess at this point—is that the relay constellations are all aiming to be cost-competitive with traditional ground station providers. I don’t think they have to be 1:1 to justify switching whole-hog, but parity makes the siren song irresistible.

All that spells trouble for the legacy ground station providers.

Why would you ever wait to downlink directly to a ground station if you could instantaneously relay it to the cloud? It would be like selecting the 15-minute-delayed Lyft driver when another one two minutes away cost the same.

But it’s even a little worse because of the complexity satellite imagery providers have to keep track of today with ground stations and “passes” over them—why should I have to understand the plumbing used to get my data down at that level of granularity?

I shouldn’t. My data should just appear in my blob storage in the same cloud environment and region I do everything else in. I don’t care the path it takes to get there as long as it’s fast, legal, auditable, and secure. Which is pretty much the same set of objections cloud providers have been working tirelessly to have unassailable answers to for the last decade.

Ok, so if can relay data to and from any compatible dish or ground station in the world, more or less instantly, what does that imply about how ground infrastructure might change?

Well, for one, you don’t need spatially proliferated data centers if you don’t need direct line of site. You may want regional stations to make backhaul over the open internet to certain geographies faster by strategically placing them near important nodes (much like the cloud providers have major hubs all over the world). Someone smarter than me can build the spreadsheet running the math on cost vs benefit for that one…

If you read Part I of this series, you know that ground station site selection is a huge deal, and there are very large international businesses predicated on superior positioning of ground stations for reliable line of sight to common inclinations, like polar orbiting low Earth orbit Earth observation satellites.3

But, if I were getting into the ground station game now, I wouldn’t be packing my extra-warm coat for the tundra-covered stomping grounds of the old guard. I’d be picking out my favorite Hawaiian shirt to go searching for the sunniest stretch of gigabit-speed fiber backbone I could find.

Because the new name of the game is: lasers. Space lasers, to be precise.4

These are the primary tools of relaying mass-quantities of data around the globe. If you think deeply about how a relay constellation works, it starts to hurt your head—these things autonomously spin on gimbals, locate their nearest neighbor, and pulse information across the fabric of spacetime at gigabit speed…

These relay constellations will also use regular old radios to communicate with each other and share smaller data packages (eg commands, telemetry, software updates etc.).

It’s like having a virtual ground station in space, distributed across LEO in thousands of nodes all whipping across the sky at many times the speed of sound.

And they can either send data back down to Earth via radio (low bandwidth, but not generally affected by weather) or laser link (high bandwidth but foiled by clouds).

Now, if you have 3,000+ satellites in your relay constellation, one of them is gonna have a bead on just about everywhere on Earth at all times. And that means wherever the relay constellations choose to put their ground stations, they should always be (indirectly) capable of contacting every satellite compatible with their network as long as they have line of sight to a relay satellite (which, they always should).

So where would I, Joe Morrison, locate my fields of radar antennas? Thanks for asking.5

Vivaaaaa Las Vegas! I’d look for extremely sunny areas near major fiber nodes on the cheapest land I could find. The suburbs of Las Vegas, Phoenix, and Albuquerque top my list for North America.

You could run this same analysis for anywhere in the world. Here’s a good starting point: https://globalsolaratlas.info/map.

Storm Clouds on the Horizon

So far, two of the three major cloud providers have decided to make a run at the ground station establishment: Amazon’s AWS Ground Station and Microsoft’s Azure Orbital Ground Station. Google appears to be partnering with SpaceX on their ground infrastructure, but hasn’t announced a one-size-fits-all service (at least, not yet).

In the words of the indisputable conscious of my generation, Cardi B, the cloud providers are making “money moves.”

Over the coming few years, will these behemoths gobble up the many dozens of independent ground station providers around the world? Probably a few of them. I think most others will die of natural causes in short order.

There will always be a small number of "indy" ground stations that hold out, but I think that will be primarily for uneconomic reasons. Most likely they'll operate systems used specifically for government work that comes with its own set of idiosyncratic political, regulatory, and procedural pressures. It’s not too different from how there will always be national constellations for both comms and Earth observation despite buying commercially almost always penciling as a much better deal.6

Take this all with a grain of salt - people have been making the same (bad) call for a lot longer than I’ve been alive. One of my favorite accounts on Twitter, a small ground station provider in Dundee, Scotland, put me on blast for this prediction. “Rumors of my death have been greatly exaggerated,” as they say.

And the most obvious objection to this newfangled common pipeline for tasking satellites is The One Major Concern…some of you spooky types have been thinking it this whole time…

“Wait!” I hear you calling from behind the dim glow of your phone as you read this incredulously on the toilet, “Won’t it always be more secure to directly downlink to a trusted ground station and therefore there will always be a role for geographically distributed and even mobile ground stations, especially for military users?”

Great question, imaginary interlocutor. And incredibly convenient, as well. Because I’m gonna talk all about that in the next section!

Tactical Impractical Downlink

The concept of “tactical” or “in-theater” downlink is very simple: rather than route tasks and data through a standard ground station network, out to the open internet, and then onwards to end-users, you simply instruct satellites to downlink certain data “directly” to those end users who receive and process the data locally.

The two most common places this pops up in the industry are:

1. “Direct access” programs that allow customers to task satellites and receive data directly during reserved orbits or sections of orbits without the commercial provider knowing much about it at all.

2. “Mobile ground stations” that enable military units to receive data in the field and process it locally where it can be put to use immediately.

Most people have never heard of either, because most people aren’t a top-50 military in the world. Direct access, in particular, is a tried-and-true business model - it’s Maxar’s bread-and-butter, and they oughta know what works, because they’re currently being acquired for a whopping $6.4B.

And tactical downlink is emerging as a topic du jour in military circles where fears of a potential “denied” or “degraded” space environment in the event of a major conflict have prompted people to think about creating more resilient protocols and systems.

In both of these cases, the rationale for all the added complexity is that it’s more secure to directly interact with the satellites than it is through a third party network. Military and intelligence customers care a lot about what’s called “chain of custody,” and ensuring that prying eyes didn’t see what they collected or, worse, intercept and sabotage the data in some clever way.

However, my contention is that modern hashing, authentication, and end-to-end encryption techniques allow for information to flow very securely across the open web. And no, I’m not talking about blockchain, I’m talking about stuff like SHA-3, the Double Ratchet Algorithm and open implementations of authentication schemes like JSON Web Tokens. These are all open standards with ecosystems of paid services built around them for ease of access - there's no need to reinvent the wheel.

Now…even if the data is encrypted at rest, there’s still a trust issue. You might actually want your imagery provider to do quality assurance on the data before sending it to you. You might also want them looking at things like telemetry to help ensure continuity of service and health of the satellites.

But, whether you’re dealing with a direct downlink or a relay through an open network, that element of trust will always be required. My basic point is that, with proper authentication, end-to-end encryption, hashing/provenance services, and other independently audited security measures, the level of trust is really no more elevated for tasking over the open internet than it is for direct access or tactical downlink.

We’re seeing in real time how Starlink is making a huge difference in Ukraine where the importance of maintaining comms is paramount and commercial satellite imagery providers are moving the needle. Ukraine didn’t need to invest $100Ms in mobile ground station infrastructure to be able to take advantage of that - and it can certainly be done securely based on mutual willingness to authenticate, encrypt, and hash data going back-and-forth across the wire (all with free and open technology).

And even in a degraded or denied space environment where jamming and even Anti-satellite weapons (ASAT) have been deployed, you have to think to yourself…will it be easier to establish contact with 3,000+ node comms network or a 30+ satellite imagery constellation…not to mention which of those two is more likely to be rapidly redeployed (the one with manufacturing capability to churn out 1,000s of satellites in short order and multi-purpose uses beyond just ISR).

I concede a few things:

1. I am not a cyber security expert by any stretch of the imagination, nor do I hold a clearance, so I don’t know if there are other reasons why direct and tactical downlink may be necessary (even, perhaps, for logistical reasons e.g. physical proximity to air-gapped environments).

2. Changing the hearts and minds of the world’s military and intelligence leaders happens over decades, not months, and there’s a clear emphasis on direct and tactical downlink across the world that won’t reverse course overnight.

So, while I’m skeptical of the longterm utility of this kind of technology, I think every satellite imagery provider needs to create some offering in this regard to stay competitive. I don’t begrudge them the choice, nor do I envy the effort required to duplicate most of your ground segment for on-premise deployment.

More to come…

Part III will be all about processing at the edge in space - I had intended to include it in this blog, but Part II was already way too long. Plus, I’m back to work this week, and we just launched two more satellites today, so things are…busy.

In this two-part series, I’m going to take A Closer Look™️ at the emerging reality of commercial, ubiquitous connectivity in space. Part I covers the basics of how we communicate with Earth observation satellites today.

Part II (coming soon!) will go into more detail on how internet in space will start to dramatically affect the Earth observation industry in the near future.

For those reading this at or near the date of publication…happy New Year, you hopeless nerds. Say hi to your exasperated families for me—maybe bring this essay up while you’re all waiting for the ball to drop. Trust me, they will love that.

It’s Crazy Any of This Works at All

Have you ever considered how unbelievable it is that satellite imagery even exists at all??

Sometimes when I conjure a map of high resolution satellite image on my phone simply by dancing my dainty little fingers across the screen, I think to myself, “I live in the future.”¹

Now that I work at a satellite imagery company, I’ve become acutely aware of the long sequence of hurdles every satellite image must navigate to arrive safely before your eyes. The more I’ve learned about these hurdles, the greater my sense of wonder has grown that people figured this stuff out in the 50s...

Launch is hard. Manufacturing is hard. Getting stuff built and tested on Earth to work in a zero G, frigid, irradiated vacuum is hard. Summarily: space is hard.²

Because of the challenge of space, there’s another key part of every mission that is criminally overlooked. It’s the very foundation of the modern space technology stack, the bedrock of the two-way communication between humans on Earth and assets in space known as the “ground segment” (“ground” for short).³ You could say…ground is hard 😏.

There’s a monumental change coming to the space segment just around the corner that will make the experience of tasking satellites feel even more magical than it already does—relay constellations. But the effects of these massively ambitious projects will be felt most acutely back on the ground…in a sense, a big portion of the ground segment is “lifting off.”

The Space Development Agency (SDA) transport layer. SpaceX’s Starlink. Amazon’s Kuiper. Inmarsat. Viasat. Kepler. Hedron. Spacelink. Warpspace. The list of major relay constellations being built out seemingly grows longer by the day.

Thousands of these relay satellites are scheduled to launch in the coming few years, and the wave is evident already with Starlink serving as a preview of what’s to come:

In this two-part essay series, I’ll share some strong opinions (loosely held), and I look forward to reading your responses, especially opposing perspectives.⁴

Latency, Lord of the Spies

To understand why relay constellations are gonna be such a big deal for the satellite imagery industry, first you have to understand some basics about what makes satellite imagery companies tick.

There are three key value propositions in the business of satellite imagery: resolution, coverage, and latency.

Of those, internet connectivity in space most directly influences latency, which is just the time between when someone realizes they want a satellite image and when they actually get it.

Latency is tightly correlated with pricing power. There are people who will pay quite a lot of cheese for an image if you can deliver it in 30 minutes without any prior warning. Some of those same people have no use for an image if it’s delivered just a couple hours after it’s captured.⁵

There are no four words in the satellite imagery industry that command suitcases full of cash quite like “high resolution” and “low latency” when strung together in a sentence.  The intersection of these two value propositions enables a category of use cases I call “mobilization,” which I contend is the highest and best use for satellite imagery. Read the footnote if you are feeling lost.⁶

The Waiting Game

Contrary to my wildest dreams, Earth observation satellites don’t have zero marginal cost for new image acquisitions in the same way that SaaS products have zero marginal cost for serving data to new users.

Satellite imagery companies do, however, benefit tremendously from economies of scale due to the way ground station and cloud infrastructure reservations work. They pay real costs to get every image down, which is partly why they are stubbornly trying, and mostly failing, to sell lots of ad hoc archival images (since those images do have close to zero marginal cost to serve).⁷

Unfortunately, low latency is expensive to provide, and rarely necessary. But in an emergency, every second counts.

The brute-force way to reduce latency is to simply launch more satellites—more birds means more opportunities to collect. Waiting around for an “access opportunity” is typically the largest source of latency for customers today.

But as you add satellites, there are diminishing returns to latency. Suddenly, a new bottleneck emerges—the time between collection and delivery. Historically, reducing latency at this step has entailed adding more ground station sites and processing data on board the satellite itself (more on that later).

Due to the very real costs of reducing latency at each of those steps, it’s helpful to break it down into a funnel to more easily assess where data tends to get bottlenecked:

I’m going to mostly ignore ordering and collecting, because they aren’t significantly affected by the advent of internet in space.⁸

Commanding

After you order satellite imagery, the instructions have to somehow get up to the satellite. The most common way of commanding satellites is via radio frequencies emitted from ground stations. Often, these are remote data centers with big fields full of radar dishes that trace the path of satellites as they whiz across the sky in order to send and receive messages.

For polar orbiting satellites (the ones that fly North-South/South-North), ground stations near the poles are most favorably located in terms of available time on target.⁹ The closer you get to the equator, the less frequently polar orbiting satellites will have line of sight to any particular point along that latitude band. But, near the poles, no matter where you are latitudinally (is that a word?), the satellites will pass overhead at a decent angle almost every every orbit.

Orbits typically take about 90 minutes in low Earth orbit. Assuming a perfectly efficient system with no competition for resources on board the satellite or at the ground station, this simple two-ground station setup means that customers placing orders rarely have to wait more than about 45 minutes for the satellites they are communicating with to receive instruction.

What if there’s a satellite passing a target of interest that won’t make a ground station contact beforehand? Like, what if it’s gonna go right over your target in the next five minutes? Remember - it’s actually kind of ideal when a satellite only has a few minutes heads up before an acquisition opportunity, that’s the nature of emergencies. And emergency tasking pays the bills.

In the traditional model of reliance on line of sight with ground stations, you’re simply out of luck. Your ideal satellite will whiz right by your target at 16,000mph in blissful ignorance.

It’s hard to say exactly how much lost revenue potential satellite imagery providers suffer due to the buffer between ordering and commanding, but the larger a constellation becomes in terms of assets on orbit, the more likely it is that at any given moment one or more of them is missing a fleeting (and lucrative) opportunity.

Downlinking

But what’s the point of instantaneous tasking if you don’t also have instantaneous downlink? 

To use a rather personal analogy: I don’t care that you baked the pizza, Dominos. That gives me no satisfaction. I need you to bring it to me. I want to eat the pizza. Eating the pizza is what gives me satisfaction.¹⁰

As with commanding, downlinking relies heavily on line of sight to ground stations for most modern satellite imagery constellations. So, expanding your ground station network is a “twofer.”

But, whereas a command might only constitute a few MBs, a downlink might be many GBs, even when compressed. The vast majority of bandwidth consumed by satellite imagery providers is getting stuff back down to the ground, not sending instructions up. So, every second of contact with a ground station is precious if you’re putting your satellites to good use.

Almost all satellites use radio frequencies to communicate commands, but some use high-throughput “laser links” to beam data back down instead of RF. Also called optical terminals, these suffer from the same affliction as their optical telescope cousins: they succumb to clouds…and most of the world is cloudy most of the time.

One of the most frustrating things that can happen to a satellite imagery provider is a “missed pass” resulting in a traffic jam of data accruing on the satellite’s storage system.

With ground station contact time being such a precious commodity, how do you subsequently prioritize downlink? Do you get the hot-n-ready pizzas out while they’re still fresh or do you downlink on a first-come, first-served basis resulting in potentially more customers being impacted by higher-than-ideal latency?

Optical terminals allowing for high-throughput downlink are certainly promising, but without the ability to relay between satellites to a clear line of sight with a ground station (more on that in Part II), they’re also a bit risky. Most choose to compress their data and send it down via radio frequencies that cut through the atmosphere unperturbed, even though the throughput is often lower.

Processing & Delivering

Alright, so you managed to snap a picture and send it back down to Earth, now you’ve gotta turn it into interpretable information.

The simplest way to do processing and delivery is to send down your raw data from the spacecraft and process them in the cloud.

There is one major alternative to doing processing the cloud: processing at the “edge.”The “edge” can either refer to the satellite itself (“space edge”) or servers at the ground station (“ground edge”).

The benefit of edge processing in space is that when you turn raw samples into structured files, you can compress them much more efficiently. As long as your processor can keep up with the rate of imaging, you don’t have to worry about a queue of raw data building up after a missed pass causing further delays in a processing queue on the ground later. The downside is that…it’s in space. Like, there’s no worse place imaginable to deploy software.

The benefit of edge processing on the ground is that it’s easier to update and more resilient to confounding environmental influence than edge processing in space. More importantly, though, if your ground station is mobile and has edge processing capabilities, it opens up the possibility of “tactical” or “in-theater” downlink (even, potentially, across multiple constellations).

If the satellite downlinks directly to your location and you process the data on the ground, not only is that relatively secure, there’s no need to worry about “backhaul” over a network because you brought the whole damn ground segment with you.¹¹

Backhaul refers to the journey from the ground station to the cloud environment where you’re likely renting a fleet of beefy machines for processing, visualization, and/or egress.

Remember, some of these ground stations are extremely remote on purpose to be near the poles…it can sometimes take a while for them to send data back through the series of tubes known as “the internet.”

What to expect in Part II

Now that you’ve got a pretty solid picture of the status quo, we can talk more specifically about how internet in space will upend it.

In Part II I’ll cover why:

• Ground station providers will face extreme consolidation pressure from cloud providers in the same way collocation data centers have been rolled up in the last decade

• Tactical downlink will slowly (then quickly) become irrelevant

• Edge processing in space is overhyped and edge processing on the ground is under-hyped

Part II is scheduled to drop three days from now. Don’t hold your breath.

¹ In truth, my fingers are actually really big and strong, and not dainty at all. They would surely impress you if you ever saw them. In this case, I’m just saying they’re dainty for comedic effect, and not at all because they’re actually the fingers of a lifelong indoorsman with an email job.

² “Space is hard,” is like an industry-wide inside joke. I’ve had a t-shirt concept in mind for a while... imagine an illustrated astronaut helplessly floating away from an electric bass with an outstretched hand and the caption “Bass is hard.” Would anyone buy that t-shirt? I wouldn’t.

³ I think “ground” is a funny name for something that mostly runs in the “cloud” and literally involves the manipulation of electromagnetic waves traveling through the “air.”

Relatedly…my wife’s grandfather was a pretty smart guy. He helped build some of the first-ever “Earth stations” as they were called back then (much more poetic sounding than ground station in my opinion). He got his start working on a classified Army mission called ADVENT that built some of the first Earth stations for geosynchronous communications satellites in the 60s. Later, he would travel the world leading the build-out an Earth station network for COMSAT.

He retired as an engineering leader at Inmarsat, still today a leading maritime communications provider. I never had the pleasure of meeting him, but I think about him a lot. He’s taken quite a bit of pressure off me - pretty much no matter what I do, I’ll never have the most impressive career in the commercial space industry even among my immediate family members!

Given how hard these things still are today, I get a little sick to my stomach imagining him helping to invent two-way comms with a satellite using mostly pencil and paper in the 60s…like I said, he was a pretty smart guy. 

⁴ I’m kidding, of course. I am entirely uninterested in reading anything you share with me besides obsequious flattery.

⁵ Satellite images are a lot like Krispy Kreme donuts - they just hit different coming straight off the line. Shoutout Krispy Kreme. Shoutout to the assembly line. Shoutout to conveyor belt oven technology. Shoutout Quiznos. RIP Quiznos.

⁶ If you don’t know what I’m talking about, then you probably haven’t read my last essay describing a simple framework for how to understand the satellite imagery industry. If you have read it and you don’t know what I’m talking about, then I’m sorry to be the one to break it you, but…you have extremely poor reading comprehension. Extremely poor. For those of you who have read it, remember it, and are still bothering to read to the end of this footnote anyway…you saucy rascal. You’re one of the real ones.

⁷ For more Thoughts™ on the siren song of archival data sales, check out my wildly unpopular essay, at least among satellite imagery executives, “Open All of the Satellite Imagery Archives.”

⁸ However…I can’t help but take the opportunity in these footnotes to complain about my least favorite of all the phases: ordering. Incredibly, ordering takes anywhere from 60 seconds to 6 months depending on how stubborn and regressive the imagery provider is. Until I worked at Umbra, I had never *directly* ordered a satellite image, I had always negotiated with sales people who placed task requests on my behalf. But after joining up, I was able to experience the sublime satisfaction of clicking a point on Earth, hitting “submit,” and getting a satellite image back without a human in the loop…and brother, let me tell you hhhwhat….once you taste the forbidden fruit, there’s no goin’ back.

⁹ Polar orbit is an extremely popular choice for many satellite imagery constellations for several reasons. For one, since the Earth spins underneath of you as you orbit, you wind up seeing almost everywhere on Earth at some point periodically. Depending on your altitude, there can actually be gaps along the equator that are impossible to image, but that’s often an acceptable tradeoff considering most of the equator is empty ocean anyway. Polar orbit can also allow for “sun synchronicity,” meaning the sun consistently illuminates the Earth below you at regular angles, so shadows and timing are consistent between revisits. And because it’s the most popular inclination, it’s also the easiest and cheapest to get to - when you catch a “rideshare” like one of the SpaceX Transporter missions, you often get “dropped off” into polar orbit. Wayyyy cheaper than a dedicated ride to a mid-inclination that can offer much faster revisit over the mid-latitudes where most of the world’s people live.

¹⁰ Sorry, I’m just a little touchy ever since I learned the pizza tracker is fake.

¹¹ Spoiler alert! Direct tactical downlink is gonna get roasted in Part II. In my opinion, it has about as much relevance to the future of satellite imagery as LAN parties have to the future of competitive gaming.

There are only three basic rackets in the commercial satellite imagery industry: Mapping, Monitoring, and Mobilizing.

I am constantly impressed by the varied and obscure applications of satellite imagery.² Normalized difference vegetation indices. Small baseline subsets of interferograms. Convolutional neural networks. It can all feel a bit…overwhelming.

But at the end of the day, it all boils down to the Three M’s of satellite imagery. Emineminem, if you will.

I use the emineminem framework all the time. I find it both helpful at the micro-scale for clarifying my stance on strategic decisions we’re making at Umbra and at the macro-scale for making sense of industry news.

Dare I call it a mental model? Perhaps I’m being overly generous, since it’s really just a Venn diagram that I scribbled on an iPad on the plane home from Thanksgiving.

Whatever you want to call it, I hope you also find some value in it. I know I have!³

The Three M’s Explained

It’s a common fallacy to believe that the “hard part” of building a satellite imagery company is designing, launching, and operating spacecraft that produce high quality data. It is really hard to do that. At times, it literally involves rocket science.

But, the history of this industry reveals that a good portion of the people who successfully do those hard things still go bankrupt.

Therefore, I think the hardest part of building a satellite imagery company isn’t the satellite imagery part - it’s the company part. You gotta sell stuff to people for more than it cost you to make it. Few companies in the history of this industry have ever done that.

In my opinion, you can’t solve for the challenge of company building if you try to juggle all three of the M’s simultaneously - you must pick one, become the best in the world at it, and build a profitable foundation on top of it before branching out.

Let’s take a closer look at each of the M’s.

Mapping

Mapping is the most mature application of satellite imagery. People have been looking down at Earth from above, taking pictures, and using that to make better maps since the mid-19th century.

It’s also extremely expensive and difficult to support - mappers tend to want very fine resolution, extraordinarily accurate geolocation, and large, seamless coverage areas.

They’re willing to sacrifice timeliness for fidelity. After all, most “foundational features,” like roads and lakes don’t change much year over year.

And yet, for all that effort, it’s also probably the least attractive of the three M’s from a financial standpoint. Lumpy, infrequent, non-recurring sales…what’s not to love?

Not to mention—the competition is staggering. Most of the pixels people actually look at in Google Map’s “satellite imagery” layer are in fact aerial. And the proliferation of consumer drones has further eroded the relevance of satellite imagery for many mapping use cases. For the cost of one tasked 50cm satellite image, you can drive to Best Buy, get a professional-grade drone, and fly your own 5cm imagery before the command has even made it onto the satellite.

When I call for the satellite imagery industry to release its archives into the public domain, the reason I don’t think it’s such a crazy idea is because mapping is a dying line of business in my opinion. Even the aerial companies are moving to cheap subscription models rather than one-off licensing deals (see Nearmap’s recent acquisition as an example).

Monitoring

Now we’re talkin’ baby! Monitoring…just checkin’ in. Keepin’ tabs. Creepin’.

Likely the smallest niche of the group in terms of direct economic value capture, Monitoring is also the fastest growing category of the bunch.

A few billion dollars of VC funding is premised on monitoring becoming a bigger niche than either mapping or mobilization. I personally believe it can replace mapping as the second most important way to make money in the space, although I’ve long felt the process will happen slowly over a decade or more. It’s not that we can’t capture lots of imagery every day…it’s that almost nobody even knows what to look for in it, let alone how to look for it.

I think VC’s love affair with monitoring is less to do with any objective demand signal for it, and more to do with its technical, financial characteristics. Monitoring lends itself to a subscription model (eg $X/km²/month), and that’s a lot more desirable revenue that the one-off purchases inherent to the other two niches.

Predictable, recurring revenue commands a higher valuation multiple due to the accounting principle of the time-value of money—the higher the confidence investors have in future cash flows, the less they discount future cashflows, resulting in a higher present-day valuation.

There’s just one issue…

Monitoring has long been the domain of billion-dollar scientific missions with global coverage and exquisitely calibrated data. Billions have gone into these projects - stuff like Landsat, Sentinel, NISAR.

Unfortunately, due to Physics™️ and the inherent tradeoff between spatial resolution and coverage, it’s tough for commercial providers to sufficiently differentiate themselves. Sure, you can achieve a higher resolution and faster revisit cadence on a lower cost-basis, and you might even be able to reach parity on the spectral/radiometric calibration. But you’re competing with free. Free as in beer and speech.

It will always be tough to justify the (infinite?) premium you need to charge customers for a commercial, global monitoring product. But…crack that nut and it’ll rain B2B SaaS-esque forward revenue multiples from the rafters like confetti.

Mobilizing

I admit it. Mobilizing is the most contrived “M” of the bunch. Hear me out, I think I can make it work…but it’s a little weak.

Mobilizing refers to the phenomenon of a customer reacting to an event by tasking a very high resolution, low latency image to further characterize that event. It could be verifying troop movements in a war zone, or estimating impact during and after a major flood event, or checking to see whether or not a manufacturing facility is currently in use. Whatever the reason - something acute and fleeting is motivating the customer to ask for that satellite imagery.

The key is that the imagery offers sufficient spatial resolution to see human-scale changes and is delivered quickly enough to inform a rapid response (ergo…”mobilization”). This use case is the backbone of every profitable satellite imagery company in history.

People often ask, “What is the killer application for satellite imagery?” Well, this is it. And I mean that very literally. Targeting is an example of a mobilization use case.

Mobilizing doesn’t naturally result in recurring revenue, because it’s inherently uncertain and reactive. However, it can be structured in a recurring way, sort of like a retainer on future capacity (e.g. the “direct access” programs that sell dedicated orbits to customers). How convenient!

Customers looking for this product are less price sensitive than their monitoring and mapping counterparts - it’s hard to assign a value to a piece of information that can literally save a life if delivered quickly enough and with high enough fidelity.

And unlike mapping, there often isn’t a viable alternative - you can’t easily fly a survey plane over denied airspace or use a drone in a hurricane.

A Quick Aside…

It’s nice to be back. I haven’t written for fun in over six months. Let me know what else you are curious about that I could cover in future essays (simply replying to this email goes directly to my personal inbox).

My only other half-written essay in the drafts is about something far more esoteric (how internet in space will affect the satellite imagery industry). So, I’m open to ideas.

Thanks for reading and, above all, stay saucy my friends.

¹ This is the first time I’ve ever put a footnote in the disclaimer, but it’s for an important reason. I think “saucy” has the potential to become my signature adjective. I intend to liberate it from its cultural imprisonment to the phrase saucy minx, a small-minded and bizarre bench role for a player with adjectival hall-of-fame potential.

² Originally I wrote, “sundry and abstruse” instead of “varied and obscure.” In honor of my decision to show restraint, I’m taking this footnote as an opportunity to remind you that, yeah, I went to a liberal arts college. I read books on rare occasions. My vocabulary is significantly above average, even among college graduates (but only if you include the for-profit online schools in that calculation).

³ In fact, I hope for a lot more than just this “mental model” being useful to you. I hope for a lot of things. I hope you find joy and happiness in your life. 

I also hope you share this essay with one person who you absolutely despise, preferably someone with no connection whatsoever to the satellite imagery industry. And, I hope when you share it, you include a note like, “You have to read this, you will love it. There’s one footnote in particular that made me think of you.” 

Maybe you’re even the person that received a link to this essay along with that provocative note. In that case, I hope that you are currently reading this footnote with growing confusion and resentment. I hope you can reflect on why someone would despise you enough to share this essay with you. 

Most of all, I hope you can see this moment for what it is. I want you to imagine your life sprawling out before you on all its splendor and preciousness — you are at a fork in the road. Along the righthand path, should you choose to take it, you will find peace and reconciliation by forgiving the person who shared this essay with you. Along the lefthand path, you will find a new career in the satellite imagery industry. 

Choose wisely. But should you venture down that sinister way into the tangled wood off to the left…welcome, friend. We’ve been expecting you.

Earlier this month I put forth a meandering argument as to why satellite imagery providers should open their archives to the world for free, and I was totally floored by the response I got to it.

Lots of people reached out to say that they agree—and not just customers, people working at the satellite imagery companies I was poking at, too! I’d like to expound those ideas.¹

What Happens When You Give Away Your Archive

Have you ever looked at a chart of the academic publications containing papers that reference Landsat over time?²

That’s a hockey stick, baby. Ex-po-nen-tial.

Raw number of publications per year probably isn’t a perfect proxy for utilization, so let’s look at downloads:

Why does that chart begin in 2009? Because, in 2008 USGS first instituted the policy of giving away Landsat data for free (first with Landsat 7, then the rest of the archive in 2009).³ It probably felt like a crazy choice at the time after billions of dollars of investment in the program… but the effect was immediate and extraordinary. Roughly a 100-fold increase in downloads in a decade.⁴

But even that’s not the whole story!!

In 2010, Google’s Earth Engine project was unveiled with the entire Landsat archive already loaded in. By 2015, AWS had ceremoniously added the Landsat archive to its free “Earth on AWS” repository. Two years ago Microsoft joined the party with the launch of their “Planetary Computer” initiative, which also hosts Landsat data. Hardly anyone I know downloads Landsat scenes anymore - for planetary scale analysis, you tend to “bring the algorithms to the imagery.”

It’s hard to quantify the value of Landsat, but the last time USGS tried in 2017 they estimated that it produces $3.45B in value to society annually. There’s a second important conclusion buried in that study: if they tried charging for the data, that value would likely vanish in the blink of an eye:

Here’s the horrible paradox of satellite imagery archives: they are simultaneously the most tangible, compounding asset satellite imagery constellations create over time… and yet the vast majority of users are categorically uninterested in paying for access to them.

A Moral Argument for Giving It All Away

If Landsat creates billions in annual value to society annually…imagine what the value to society would be if Maxar, Airbus, and Planet released their archives! Those repositories contain data with up to 100x the spatial resolution and 10x the temporal revisit of Landsat…

And just like Landsat data is more valuable due to the availability of Sentinel-1 and Sentinel-2 data to combine it with, releasing commercial archives would have a multiplicative effect.

But to a lot of people in the industry, the idea of just giving away the archive is… unthinkable? Ludicrous. Self-sabotaging, impractical, unsustainable. In a word: stupid.

However, few would disagree that opening up commercial archives would lead to an explosion of new research and entrepreneurship that would greatly benefit society writ large, just as the opening of Landsat’s archive precipitated a little more than a decade ago. To be fair, there are some people who disagree that opening up access would have any net effect—I’ve drawn a helpful diagram to describe that type of person:

While I do not have access to their books, I feel comfortable speculating that Maxar, Airbus, Planet and the rest do not rely in any meaningful way on revenue from data collected 5+ years ago.

And yet, those petabytes of high resolution images are tragically gathering dust in data centers owned by Amazon (in the case of Maxar) and Google (in the case of Planet and Airbus).⁵

Meanwhile, if you listen to the executives of these companies speak publicly, they talk enthusiastically about the irreplaceable role that their data has to play in quantifying the effects of climate change, documenting the horrors of human rights violations, disrupting illicit trade networks, and mitigating a litany of other highfalutin problems facing humanity.

I agree with their assessment! Every week I meet more and more entrepreneurs, researchers, and philanthropists working tirelessly to convert raw satellite data into products that predict, preempt, and pacify human suffering and inequity.

My point is this: if you’ve generated archival data that is being kept in a state where it doesn’t produce significant shareholder value, and yet you know it has the potential to create immense societal value, do you have an ethical duty to release it to the world?

In my opinion, the obvious answer is YES, provided you can do it sustainably, which I believe you can.

The Economic Argument for Lettin’ It Rip

I’ve talked to a lot of smart people in this industry, including current and former executives at dozens of satellite imagery companies. Almost no one disagrees that “old” imagery should be easier to gain access to. In my opinion, it’s just a matter of degrees—maybe for one provider it’s ten years and older, and for another it’s only 90 days and older.

People disagree about the way the data should be licensed, the way access should be screened (due to security concerns), the leverage that cloud providers should have in whatever solution is offered, and other details. But all of those can be overcome with a little consistent effort. And yet, they haven’t been. What gives?

My theory as to why this data remains locked away and largely unused boils down to three reasons:

1. Reputational Risk | Doing something no one has ever done creates serious, potentially career-ending professional risk. The risk-reward ratio on a personal level for these executives makes advocating for this strategy fairly unappealing. Even if a CEO believes in this idea, they have to face a board who almost certainly won’t find it compelling at first glance. No one ever got fired by the board for not giving their archive away.

2. Credibility | Part of the narrative these executives have sold investors is that their unique proprietary advantage includes unique access to an archive no one else has. They may even assign a real dollar value to it as an intangible asset on their books. Giving the archive away would basically require admitting you aren’t going to be the all-things-to-all-people-AI/ML-overlords you may have once pitched investors on and outdated raw data is actually worth what people pay for it: practically nothing.

3. Motivation | Executives are motivated by the incentives designed by their boards. No one is getting a fat bonus for *checks notes* giving stuff away. For the publicly traded companies, especially, boards would seemingly rather see revenue increase a little every quarter than stomach two quarters of declining revenue in exchange for two subsequent years of rapid growth. Opening archival data is a 10-year bet on the creativity of entrepreneurs to build an entirely new economy of monitoring products. In other words, it’s a distraction from hitting month-end sales goals.

If you agree with me that we are undergoing a tectonic shift from the mapping paradigm of old to the monitoring business case of the future, then your attitude toward releasing old data is probably similar to mine.

I believe giving away archival data will result in several benefits that ultimately drive more demand for timely subscription and tasked data far in excess of the lost revenue from archival sales. Those benefits are:

1. Market expansion | Getting access to data requires dealing with a seemingly endless queue of sales people who are trained to look for what worked in the past, not what will work in the future. Open archives would allow new prospects to skip the inane sales back-and-forth and bring the cost of experimentation to ~zero, both in terms of time and money. It’s hard to quantify the true attrition rate of an average satellite imagery sales process - I would bet the vast majority of leads never even bother to subject themselves to the first sales call. Open archives de-risk purchases of fresh data by providing an accurate baseline to understand performance historically, which results in faster product iteration, more products getting launched, and more companies getting started. In a mapping-centric model, old data may jeopardize fresh data sales, since foundational data like roads, buildings, and lakes don’t change that much in the course of a few years. However, in a monitoring-centric model, giving old data away actually increases demand for fresh data, because if you’re monitoring something you want to know what’s happening now, not what happened a year ago.

2. Fused datasets | Magic happens when you combine datasets. A “fused” product or derived dataset based on multiple sources is often more accurate, more resilient, and more easily generalizable than something built on a single input. Opening the archive would allow for novel fused datasets that for certain use cases would be more valuable than the sum of their parts. Researchers and philanthropists, and even some for-profit companies, would gladly donate their work back into the open. Over a three year period, which one do you think would produce more value for your company: your current IRAD initiatives or the cumulative creativity of the remote sensing community?

3. Organic marketing | Whoever first invented the map attribution requirements was a genius. I feel like DigitalGlobe is still a more recognizable brand than Maxar because of the many years of repetitive exposure on nightly news graphics. Just because you give the data away doesn’t mean you can’t require proper attribution. “© 2022 Satellite Imagery Co.” is our industry’s version of “Sent from my iPhone.” Every clever viral product and news story built on top of open archival data will have a legal obligation to credit their sources.

If, like me, you believe opening old data will steadily increase demand for fresh data (where most of the profit comes from already), then you are left with one major hurdle: cost of service. It’s true that serving petabytes of data to customers all over the world incurs real costs. I suspect the cloud bill for some of these providers is already in the low-seven-figure-per-month range. So, how can you afford the spike in usage that would result from releasing your data?

There are two obvious options to me:

1. Strike deals with the cloud providers to host the data for free. You can still charge for downloads/egress (i.e. “requester pays”) and the cloud providers get to keep the upside of all the computation that occurs on their platforms.

2. Expand upon existing government initiatives to broaden access to archival data to the public. Either that’s revising the terms of defense contracts like whatever the next version of EnhancedView, or doubling down on civil gov’t deals like NICFI’s deal with Planet, NASA’s Commercial Smallsat Data Acquisition Program, or ESA’s past commercial data procurements.

Personally, I would pursue both simultaneously. Maybe option one yields results in the short term and option two makes it sustainable for the long haul.

Some have argued that giving cloud providers even more leverage over the future of remote sensing is a bad idea. I agree to an extent—Google likes to deprecate things that millions of people use daily for sport. Corporate charity is a fickle mistress.

However, since the majority of cloud provider’s profit allegedly comes from renting out compute, not storage, it’s probably a profitable deal for them even if they cover the hosting costs. And, because there are at least three major cloud providers who have already made huge commitments to remote sensing, competition is likely to incentivize good behavior.

I’m not too worried about giving cloud providers more “leverage.” It’s tough to imagine a scenario much worse than the status quo. As an industry, we have collectively chosen complacence in the face of terrific need.

It’s past time: open all of the satellite imagery archives.

¹ How can you tell someone got a liberal arts degree? We use “expound” instead of “expand upon” just to make our writing a little less accessible. Eat that, STEM nerds. Go read a really old book for once in your life, losers.

² If you don’t know what Landsat (NASA/USGS) and Sentinel-1/Sentinel-2 (ESA) are, they are scientific Earth observation missions involving large, exquisitely calibrated satellites that capture imagery of the entire globe on a regular cadence. Landsat is celebrating its 50th anniversary this year. In fact, there have been 9 different satellites bearing the Landsat name over those five decades—Landsat 9 was launched last year! The data they produce is open and hosted freely for anyone to access. Pretty wild.

³ According to NASA’s website, this led to a 60-fold increase in downloads. I suspect that’s hugely out of date. Look at the slope of the line between 2008 and 2012…then look at the slope between 2012 and today.

⁴ The same trend holds for Sentinel-1 and Sentinel-2 in case you’re thinking it’s a unique phenomenon (also courtesy of @sakkesarjakoski):

⁵ Maxar made a huge spectacle of migrating to AWS by literally driving a semi-truck of data into their maw, Planet sold off a large minority stake to Google in exchange for switching from AWS to GCP (and the SkySat née Terra Bella née Skybox assets), and Airbus is a featured case study on GCP’s blog.

This is how I would summarize the commercial satellite imagery industry today: a bunch of raccoons (defense contractors) dressed up in trench coats (Silicon Valley startups) pretending to be humans (SaaS products).

When I see a company trying to turn commercial satellite imagery into a one-size-fits-all derived data product and selling it as a subscription product, it feels like greeting an old friend. He’s still down on his luck. He thinks this time is gonna be different. Oh no…he’s gonna ask me for money again, isn’t he?

When it’s a competitor falling into this trap, part of me kind of enjoys it…

…but most of me feels frustrated. After all, I am not rooting for people to fail. We’re building an industry together, not playing a zero-sum game.

As always, I could be totally wrong. Only time will tell.

Error 404: Customer Not Found

Allow me to put it more succinctly: selling derived data as a subscription product does. not. work. I don’t care what it is. The juice is never worth the squeeze.¹

Count cars. Count airplanes. Count ships. Segment land cover. Track oil inventories. Estimate biofuels. Measure water levels. Etc. Etc. Etc.

Now! You can give that data away for free and people may use it—in fact, I think derived data streams (“information feeds”) could be a force multiplier for accelerating realization of the value trapped in massive imagery archives. There are many notable examples of derived datasets being given away for free to much fanfare.²

But sell derived data feeds? Straight up? As a subscription?

Even if it did work (it doesn’t), I’ve written before about how satellite imagery companies entering the analytics business is not my favorite thing in the world in general. There are many reasons that I hold this opinion, but principally:

• It’s hard to be a great analytics provider, harder yet to be a great satellite imagery provider, and nearly impossible to be great at both simultaneously

• You wind up competing with your customers, and that’s, like, not the vibe 💅

• Satellite imagery is inherently applicable to many problems, so you’re fighting gravity to try to focus on just one while getting pulled in a thousand other directions

But it gets worse, because satellite imagery companies are not the only offenders here. In fact, I see lots of new analytics startups pursuing the dangerously seductive strategy of building a new type of dataset and then closing their eyes and hoping customers will magically show up to start buying it.

“Tell me,” I whisper to them, “The customers who want your derived data feed…are they in the room with us right now?”

“No one wants pixels, they want insights” is wrong

There’s an almost ubiquitous cliché in the satellite imagery business that goes something like, “Nobody wants pixels, they want insights.”

It sounds really wise. After all, 99% of the people who can benefit from satellite imagery do not need to ever see it directly to receive that benefit (and wouldn’t even know how to open the file if you sent it to them).

It turns out some people do just want pixels and they spend hundreds of millions of dollars a year on satellite data, so that part of the aphorism is wrong.

Even worse, the second part is wrong, too. It turns out nobody wants insights, either!

What people want is for you to solve their problems for them. And people’s problems aren’t magically solved when you confront them with an “insight.” There’s still action that needs to happen as a result of that insight for value to be created.

Actions are facilitated by applications. I use applications to get real work done every day:

• Using Slack to send YAKOTHGIF (yet another King of the Hill GIF) to a coworker

• Using Spotify to spend half an hour finding the right playlist to blast while I fill out an expense report for three minutes

• Using Excel to shade in the cells so it makes a big smiley face

Without an application, a data feed often winds up in a bad spot: you can produce the right information at the right time with no way to monetize it.

There’s no such thing as a generic algorithm

Ok, so the main reason not to launch data feed products is that there is no demand for them, and therefore by definition they are a waste of time and resources (unless you’re trying to release the data for free in order to help the world or grow the industry or whatever).

But even if there were demand, you run into a second insidious problem. It turns out everything related to satellite imagery analytics is harder than it sounds. Everything. If you want examples, I’ve linked to some relevant blog posts in the endnotes.³

In my opinion, every supervised machine learning model is hopelessly biased by the intent of its creator(s). Namely, it inherits the bias of its training dataset (both geographic and semantic).

Unfortunately, the world does not neatly fit into a clean, universal taxonomy—at least not one that is specific enough to be practical in every day matters. So, you have to make editorial decisions. And that creates bias.

The most exaggerated form of delusion about generic algorithms is when companies release “change detection” products as stand-alone, all-purpose services. What constitutes meaningful change? You can’t answer that in advance of understanding a particular customer’s needs, and therefore you can’t train a supervised model to detect it a priori.

The most practical approaches I’ve seen to subscription data feeds are always custom. You can “fine tune” a generic “pre-trained” model with a relatively small number of examples (1-10K) and get incredible results this way. It’s a compromise, and it certainly does not produce 90% margin SaaS-like revenue, but at least it works and some people will pay for it.

An alternative (and far better) business model

Here’s the thing—the people launching data products are on the path to success. They’ve just pulled off at a rest stop thinking it was their final destination.

Underpinning ALL of my favorite companies in this industry is a custom data feed. Arturo has an algorithm that can predict relevant property characteristics for insurers using satellite or aerial imagery. Cloud to Street can predict flood risk and map flood depth using synthetic aperture radar from space. NCX can offer an efficient marketplace for carbon credits by verifying what suppliers tell them using their pre-computed dataset of carbon stocks. Upstream Tech uses water levels inferred from satellite imagery to predict hydropower generation.

What’s different about these companies? They didn’t stop at the underlying data feed. They build applications that their customers use to take actions and solve problems. If you use an application every day, you don’t mind paying a subscription for it.

Even satellite imagery companies can take this approach. It requires laser-focus on a single industry, which is quite difficult for satellite companies to do. On the whole, we’re a greedy bunch—we tend to want to have our cake and eat it, too.

The problem with data feeds is that they’re easier to work with than raw data, but not by enough. Anyone savvy enough to build an economic indicator application on top of your ship detection algorithm is very likely to also be savvy enough to just… build their own ship detection algorithm (and optimize it for their needs over the idealized generic model).

Plus, application companies like to build their own algorithms/data feeds because they view it as part of their proprietary advantage—it’s incremental investment that yields asymmetric defensibility.

A big industry with a handful of customers

I believe an underlying assumption that motivates people to keep launching these ill-fated data feed products is that they (rightly) believe there is massive, untapped potential in commercial satellite data and the industry is set to grow quite quickly as we collectively find ways to unlock that value.

I can understand the leap in logic from acknowledging not many people work with satellite data today to reasoning that launching new products that are more refined should lower the barrier to entry for new customers. More customers = growth.

However, if you believe the revised credo, “customers don’t want insights, they want applications,” then you might come to the same conclusion as me: the satellite imagery industry will always be small. Only a handful of companies will ever solve their own problems with custom applications built on data feeds—the rest will happily pay someone else to just solve the problem for them.

Crucially, I don’t mean commercial satellite imagery always be a small industry in terms of the dollars that flow through it—just small in terms of the number of firms that constitute it.⁴

¹ Ok, ok, there are exceptions to this rule. VRICON comes to mind - they spun out from DigitalGlobe, made a derived dataset of global elevation data, and then were re-acquired by the same entity that spun them out at the first opportunity. I think they have one huge customer in the U.S. Army and that on its own is enough! Ursa Space’s oil inventory product may be another major exception; they outlasted Orbital Insight which used to have a competing product but eventually shuttered it. However, these are notable exceptions to a simple rule that holds up surprisingly well to the many, many varied attempts at this business model over the last 20 years.

² Global Forest Watch is my favorite (forest cover loss). California Forest Observatory is another phenomenal example powered by my friends at Salo Sciences in partnership with Planet. I’m also partial to the USFS Wildfire Risk app that my former colleagues at Azavea made along with Headwaters Economics. I could go on, and on…Esri/Impact Observatory land cover data, Meta (Facebook) Global Population Estimate, and on, and on…

³ I’ve written up a bunch of examples of how a seemingly simple task like “count cars” becomes excruciatingly complex very quickly in practice. Is a truck a car? Ok, then is a food truck a car? Ok, maybe we should have said, “vehicles.” Cool, then is a boat a vehicle? Is a bicycle? No, ok. Well, then is a motorcycle? And on, and on, and on. And if that weren’t hard enough, the misdirection carries all the way down into the very “quantitative” metrics we use to describe “accuracy” which are just as opinionated as the models themselves. It’s a mess.

⁴ For a sense of what these firms tend to look like, here’s a thread on the topic:

Ayyyy, what’s up nerds? It’s been six months since I last sent out a newsletter. I bet you were starting to think it wasn’t ever gonna happen again! LOL, just kidding. You didn’t even notice that I haven’t been sending out any newsletters, did you?

…Don’t answer that.¹

Something happened yesterday that inspired me to dust off my proverbial sequined leotard and get back out on the dance floor🕺. That thing that happened? Tomorrow.io announced it’s going public via a SPAC. If you don’t know what a SPAC is, go figure it out before returning to finish this post (…and then explain it to me, please).

I know of Tomorrow.io because back in February of this year they announced they’ll be launching their own constellation of radar satellites (and then promptly raised $77M to get ‘er done). They popped up on my radar (😉) again in August when my good friend and fellow knucklehead satellite imagery industry commentator, Aravind Ravichandran, announced that he was joining them as their Director of Strategy. Way to go, Aravind! Wasting no time at all!

Here’s a super underrated perk of SPACs vs. traditional IPOs: instead of a bone-dry SEC filing with dozens of pages of soul-sucking legalese, SPAC deals have to share a slide deck laying out the case for their fundraise publicly on their website. These decks are, I assume, designed for an audience of investment professionals like bankers and pension fund managers. So, naturally, they’re written at about a 7th grade reading level and includes lots of pretty pictures and colorful graphs. Perfect for me.

If you’d like to leaf through Tomorrow.io’s investor presentation yourself, the source material is freely available at the link below. Careful, though, every page is marked “proprietary and confidential” so don’t tell anyone I sent you:

Tomorrow.io SPAC Deck

Why is Tomorrow.io Interesting?

You might think a weather prediction startup’s central claim would be that they’ve invented a more accurate way to predict the weather.² Nope. Unnecessary. It turns out most people in the world don’t even have decent weather prediction available to them. That’s what Tomorrow.io is trying to solve—not improving the best models, but making the basics available for everyone globally.

I view this thesis as a specific expression of a very common trend I see in the mapping world. People who live in wealthy, developed countries tend to think that the whole world is already mapped. Like… can’t most people just use Google maps to navigate where they want to go? Nah. Most of the world is explored but it isn’t, like, understood. Turns out weather works the same way.

I was particularly struck by this slide in Tomorrow.io’s deck:

How do they plan to bridge this massive gap? Satellites, baby. How else? They’re launching a bunch of ‘em. According to the deck, at least 32 are planned with the first two going up in 2022. Yeah, I said 2022. I just double checked and that’s…next year? Holy mother of God. That’s really, really soon. I’ll touch on the audacity of that prediction more in the next section.

The investor deck takes quite a bit of care to highlight the value of “active radar” sensors vs. passive radar for the purposes of weather prediction. The “active vs. passive” distinction refers to sensors that measure the specific photons they intentionally beamed down at Earth vs. measuring photons that serendipitously arrived at the sensor from other sources. The consequence of an active radar system, the deck claims, is that it allows for “direct precipitation measurements” whereas passive systems don’t.

I’m unsure of what that exactly direct measurement means in this context, since I was under the impression that passive microwave sensors can also be used to estimate precipitation, but to be honest, the technology angle here is way over my head - I literally work for an active radar satellite company and have no idea how this technology compares to what we’re doing. The distinction between “direct” and indirect measurement is something I need to learn more about. Aw, hell. Who am I kidding? I don’t understand weather prediction and probably never will.

The Rabbit in Tomorrow.io’s Hat

I think the most interesting aspect of Tomorrow.io’s story is that they didn’t start out as a satellite/space company. They’re a software company first and foremost and have built some truly exceptional applications. I mean, just look at their API documentation. It’s beautiful!

The question I keep asking myself…can a small company with consumer-grade software in its DNA really transition into a vertically integrated hardware company? And not just any hardware company—a space hardware company? Hardware is already extremely hard. But hardware that has to work in the cold vacuum of space is hardware on hard mode. It’s hard hardware that has to be hardened. Hard, Hard, Hard.

My (Twitter) friend Dominic Edmunds, CEO of PlanetWatchers, put it fairly succinctly recently:

Brian Chesky (CEO of AirBnB) recently discussed this challenge of stretching a company from one domain to a new one in service of the mission during a podcast interview. My favorite bit was when he recalled advice he had received from a college professor who told him, “Brian…you can do anything you want in your life, just not all at the same time.”

This observation doesn’t mean that Tomorrow.io can’t have its cake and eat it too. It just means that if they are successful they’ll be incredibly hard to compete with. The most valuable companies in the world are both hardware and software companies—Apple, Tesla, Amazon, Google, etc. And I believe the best satellite data providers will necessarily have excellent software. In this regard, Tomorrow.io already has a leg up on the competition.³

But…how are they gonna pull this off? Seriously. In the deck, they specifically point out that they’ve designed their own radar payload. The company I work for has *also* designed its own radar payload, so I have a fairly concrete understanding of what that looks like. Admittedly, I think it’s gotten easier to design, integrate, launch, and operate satellites over the last ten years. I’d estimate that streamlined launch services, off-the-shelf buses and navigation systems, increased access to talent, and increased access to venture capital has brought the challenge down from an intimidating “dedicate your entire life to this” level to a much more reasonable “dedicate three decades of your life to this” level.

The idea that they’ll be launching their first satellites barely more than two years after deciding to start designing them while also running a data-as-a-service business the whole time is such a stunningly ambitious goal that I simply cannot imagine even writing it down let alone committing myself to it. If (and when) they pull it off, I will be slow clapping from the sidelines shaking my head in utter amazement.

They seem aware of what they’ve bitten off, and kudos to them for acknowledging it in their investor deck. In fact, they even put it in bold type:

Money, Money, Money

SPACs are about one thing and one thing only: good storytelling. So it’s not surprising the financials are often kinda the last thing in the deck. But, let’s be honest, they’re also the most interesting part of the deck.

Tomorrow.io’s investor deck explains that the company commands a valuation of 11.1x estimated earnings in 2023 which, conveniently, are about 6x the estimated earnings for this year of $11M. Revenue is expected to balloon to $472M in five years time while margin on that revenue casually expands from 71% to 87%. They project $276M in free cash flow in that fifth year (2026). That business would be worth a lot more than $723M or whatever they’re pegged at today. In fact, that would be one of the best businesses in the history of businesses.

Here’s the surprising thing to me. They’re sitting on $102M TODAY. Even before they do this fundraise! All-in-all, the SPAC is expected to inject another $420M (nice) into the business. There’s a real chance that when the dust settles they’re gonna be sittin’ on half a billion dollars of dry powder. All I’m saying…the trade show swag is gonna be absolutely lit.

Whether these numbers are sensible or not is not for me to comment on. I don’t own individual stocks in companies I don’t directly work for, and I won’t be changing that policy any time soon. I’ll say this as someone with absolutely no skin in the game:

• The problem they’ve articulated (lack of satisfactory weather forecasting in most of the world) is undeniably enormous

• They are a real company with a real product that makes real revenue and is growing real fast

• If they pull of the constellation, that’s gonna be a deeeeeeeep competitive moat that few will even attempt to swim across

Half a billion in cash money, though. My goodness, gracious! Papa’s got a brand new bag!

A Pattern for the Future of the Earth Observation Industry?

SaaS startups are very regularly faced with the annoying question, “what’s stopping Google from doing this?” In the space industry, you just have to substitute SpaceX in that sentence if you want to sound smart. Every space tech entrepreneur has polished their answer to that one annoying question into a smooth reflective pebble of wisdom that gracefully avoids implying that they’re safe because, realistically, what they’re working on is so obscure it’s literally not worth SpaceX’s time to even consider.

The more troubling question is…what’s preventing every tech-savvy data company in the industries you sell into from moving up the stack? Why not own their own private constellation and throw you away like an old pair of gym socks? Tomorrow.io is doing it! Who’s next??

Before joining Umbra, I would have told you this will probably be a huge trend. It makes so much sense—the same barriers to entry that are toppling for startups launching constellations are making it easier for existing companies to consider dabbling.

But before Umbra, I didn’t fully appreciate the insanity of trying to moonlight as a space tech company. Reading through Tomorrow.io’s deck, one thing is clear: this is not a side hustle. They are pivoting the entire company to a new identity and going all-in on satellites. In fact, coincident with the announcement they were launching a constellation they completely rebranded the company from Clima Cell to Tomorrow.io.

It takes a deep-down-feel-it-in-your-bones level of commitment to go after a vision like Tomorrow.io’s. How many companies are willing to fundamentally change who they are in order to launch a satellite constellation successfully? I would guess very few.

So, I slept easy last night. I think Tomorrow.io is potentially on to something very, very big. And I am rooting for them. I think they’ll be successful precisely because it’s hard to pull off. They’ll become a famous counter-example, an exception to the rule (like almost all extremely valuable companies tend to be).

P.S. Congratulations to Planet for getting listed on the New York Stock Exchange this morning! Just another week in the satellite imagery industry.

¹ I’ve actually been writing way more than I’ve ever written in my whole life. Unfortunately, it’s mostly emails.

² But just in case they do want to make their weather forecasting better…you know how old people with hip replacements always feel a storm coming way before it happens? Tomorrow.io should incorporate real-time feedback from assisted living communities across the country into their models. That’s a million dollar idea.

³ You can even compete with free data if it’s cumbersome enough to access. Have you ever tried to download data from a NOAA website, for instance?

Video games were a huge part of my childhood. I grew up during the golden age of console gaming.¹ In elementary school, the first Spider-Man game I played on Playstation looked like this:

By the time I was graduating from college it looked like this:

My favorite games have always been open world adventures. I love the feelings of creativity and serendipity they illicit. And they all share one core feature in common germane to this newsletter: a map.

Gamers Invented Modern Web Maps

Web mapping nerd-historians often cite Keyhole as an important company in the history of web mapping.

It’s a story as old as time: a bunch of software engineers tried to make a video game, failed, and then accidentally revolutionized a boring industry while simply trying to salvage their game’s technological carcass.²

In Keyhole’s case, they originally tried to build a gaming engine under the moniker “Intrinsic Graphics.” When that idea fell apart, they couldn’t find a buyer, so they spun out a new company to focus just on the mapping component of their platform, Keyhole.

In-Q-Tel stepped in to finance the spinout (technically, the National Imagery and Mapping Agency, NGA’s predecessor, funded them through IQT). IQT is the non-profit venture capital firm that serves the intelligence community.

About a year after getting involved with IQT, Keyhole was acquired by Google and became the foundation for Google Earth. The acquisition turned out pretty well, I think. They spent a few million bucks, put John Hanke in charge of maps, and built a multi-billion dollar line of business and household name.

Hanke is a genius of the highest order—he now runs Niantic, makers of Pokemon Go, the most popular augmented reality game of all time. So he wound up with his video game startup after all.

It’s interesting to speculate about how today’s video games might affect the future of web mapping the way Intrinsic Graphics affected the last couple of decades.

My hunch is that there’s as much to learn from the plumbing as there is to learn from the window dressing. Cyberpunk 2077, for instance, has a fairly stunning 3D map (although, to be fair, it is also a cartographic disaster):

There are two major technological inventions happening in gaming that I am optimistic will eventually make their way to boring old web maps:

Game Engine Graphics

Unreal Engine 5 was announced last year and is siiiiiiiiiiiiiiiiiiiiick.

The new Microsoft Flight simulator uses an in-house game engine that is textured with all of Bing’s aerial and satellite imagery and is so convincing that it’s actually kind of unnerving:

Mapbox and Cesium are neck-and-neck working on making 3D mapping awesome in a modern SaaS context, but there’s still a ways to go before the maps they render look that good. They’ll get there, though, and I’m excited for it.

Compression

In video games, a 2D pixelated surface is called a “texture” while in mapping it’s called a “raster.” But they’re essentially the same thing. One of my favorite small companies is Binomial, which makes GPU-optimized compression algorithms and associated tools for shrinking HD textures. I hadn’t actually looked at their home page in a couple years before writing this blog post…and now Esri is a big logo on the homepage. See? I’m not as dumb as I look.

Two Predictions About Where We’re Headed…

UX Design Will Matter More than Ever

As the tech chasm closes between video games and boring B2B SaaS software, the remaining gulf will still be aesthetics.

One of my favorite companies of the last few years is Unfolded. A gaggle of brilliant engineers, data scientists, and designers walked away from their comfortable Bay Area tech salaries at Uber to build the most exciting mapping interface I’ve ever played with.

Cyberpunk 2077 looks like it could have been built in Unfolded. I mean, just look at this stuff:

Some of you may recognize the name Unfolded because they were in the news recently when Foursquare acquired them for an undisclosed (and probably ungodly) sum.³ A classic case of good things happening to good people.

The reason Unfolded is worth paying up for even though they’re only a few years old is simple: you get a tool that pumps out 10-minute-maps visually on par with 10-year $300M AAA gaming studio projects.

We’re Due for an (Artificial) Reality Check

If I grew up in the golden age of consoles, kids today will say they grew up in the golden age of AR. One of the first things I ever wrote about publicly online was Facebook’s acquisition of Mapillary, which it turns out was all about AR.

Sean would know. He recently sold his AR mapping company, Pixel8earth, to Snap. Mapillary’s former COO, Janine Yoong, is his coworker. And they both work with Randy Meech, whose Places of Interest mapping startup StreetCred was also acquired by Snap earlier this year.

They are building a freaky mappy Frankestein’s monster over there. If you’re working in Corp Dev at Snap: congrats, you have exceptional taste.

It would seem we’re in the midst of a massive AR mapping arms race. Apple recently updated their maps with a stylized aesthetic so smooth it could pass for a new release of Beautiful Katamari.⁴

Not to mention the work they’re doing turning everyone’s iPhones into survey grade mapping equipment (with or without the LiDAR upgrade):⁵

Meanwhile NVIDIA, the premier GPU chip makers, acquired “HD Mapping” company DeepMap last week. Maybe it’s about autonomous vehicle navigation, or maybe it’s about realistic AR maps, or maybe it’s about both.

What’s interesting to me about AR as a new interface is that it’s so new that even video game developers are struggling to figure out how to work with it. Watch this space—the new design patterns and technologies that emerge will slowly but surely make their way into the rest of the mapping industry.

By the time the innovative stuff happening in AR gaming makes it into enterprise mapping software, I’ll probably be an unbearable old man complaining about how easy kids have it now that they can just download the world’s knowledge directly into their implanted brain chip.

But at least I’ll be able to tell them to brain-search for this newsletter in their brain-index and then tell them, “See? I saw this coming.”

¹ My parents refused to buy me a video game console. I begged for a Nintendo 64 for years to no avail. Eventually, though, I got my way. In fifth grade, I entered a Daughters of the American Revolution essay contest (yes, seriously) and took home first prize—$500 of cold, hard Revolutionary cash. I bought myself a Playstation 2.

² Slack has the same origin story. When his video game project Glitch failed, Stewart Butterfield laid almost everyone off and then launched the company’s internal chat client as an enterprise communications product…somehow it worked.

³ Yes, Foursquare is still around. Much ink has been spilled on their miraculous transition from consumer app darling to enterprise API. It’s a cool story and worth reading if you aren’t already familiar with it.

⁴ Great minds think alike…I incidentally found the tweet below while searching Twitter to see if there’s a new Justin O'Beirne essay on the Apple Maps update (there is not).

⁵ Thanks to Doug Thompson for always pointing me to the most interesting underground AR stuff happening across the web. One of my favorite Twitter follows. Go follow him.

Helium is one of the most exciting companies I’ve come across in recent memory. They are building a peer-to-peer network of wireless hotspots around the world. The day I learned of Helium (April 2, 2021), they had just crossed 25,000 active hotspots globally. 10 days later, as of this posting, they’re over 26,400.

I recently purchased a hotspot so that I can support “The Peoples’ Network,” as they call it. It’s hard not to feel like a movement is afoot—25,000 metaphorical middle fingers to the major telecom companies blanketing the land in dirt cheap wireless connectivity. It’s like Napster for HAM radio nerds.¹

My hotspot hasn’t arrived yet, and probably won’t for at least a few more months. Helium-compatible “miners" (more on that term later) are currently back-ordered to the tune of hundreds of thousands of units. Approved manufacturers are desperately trying to ramp up production as message boards fill up with impatient customers.

…but why?

These are more-or-less $300-500 hobbyist radios that serve no purpose besides participating in the Helium network. Yet somehow demand has reached such hysterical heights that people are selling their old units for thousands of dollars on Ebay…

What on God’s green Earth is going on here?

Hint: it involves cryptocurrency speculation.

First, a Deeper Dive into What Helium Actually Is

I think of Helium as two things:

1) a globally federated wireless network composed of internet-connected radios that provide “LongFi” coverage and data backhaul for the “internet of things” (IoT) and;

2) the company that enables it.

LongFi is sort of like WiFi but specifically designed for low-bandwidth, long-range connections of the sort sparsely distributed IoT sensors can use to send data back to wherever their hive mind lives on the internet. A quick example:

Imagine you run an urban bike sharing service. Life is difficult for you already, because you run an urban bike sharing service. You can barely afford to bathe yourself on a regular basis, let alone spare enough change to cover a huge “machine to machine” cellular plan for your entire fleet of undignified beach cruisers. So, instead, you decide to attach cheap GPS sensors to them and connect to The Peoples’ Network via LongFi. If you had 1,000 bikes and set them to update their location once every 10 minutes, and you were able to compress your over-the-air GPS ping to less than 24 bytes (with best practices you can comfortably fit a lat/lon pair into just 6 bytes), a full year of service would run you about ~$525 for the entire fleet. All 1,000 bikes. Every 10 minutes. Insane.

Helium’s hardware looks a lot like a WiFi router. Each “hotspot” sends and receives messages using a radio transmission protocol abbreviated LoRaWAN (short for Long Range Wide Area Network). LoRaWAN relies on an RF modulation technique called “chirping” to carry information-dense signals over long distances with relatively low power consumption.²

Each hotspot is a node in a global network of similar transceivers. The primary function of hotspots is to receive information from the field and pass it along to its final destination via the internet (a process often referred to as “backhaul”). Most hotspots use Wifi or ethernet to connect to the internet, although some use cellular networks. The most effective hotspots with the broadest coverage are usually positioned in high, unobstructed places like the roof of a house. Makes sense.

There is a small but growing roster of independent manufacturers of hotspots. Helium (the company) initially manufactured the devices under their own brand and allowed for DIY setups, but they’ve since ceded hotspot manufacturing to a few partners and locked down security (including putting the kibosh on homemade rigs). You can check out all of their limited options here: https://www.helium.com/store.

What makes the hotspots “special” is a little bit of logic and some cryptographic keys that allow them to join the Helium network and verify their credentials.

Which brings us to the software component of Helium and the place where things really start to get interesting. Helium (the company) built out the actual logic of how each node would communicate with the rest of the network while remaining relatively fault tolerant. I imagine that was (and continues to be) a fairly challenging programming challenge unto itself—although wireless network architecture is fairly well-trodden territory, that doesn’t make it easy. Helium’s code is also “almost entirely” open source and worth poking around.

But the secret sauce, and the whole reason I’m excited about them, is the extremely clever incentive mechanism built directly into the Helium protocol that motivates participants in the network (i.e. the humans) to act out of self-interest in a way that constructively benefits the network overall.

And now, we get to talk about everyone’s favorite dinner party topic: cryptocurrency.

Ballin’ on the Blockchain

I am no great fan of blockchain technology. There’s an old saying in tech, “It is better to remain silent and be thought stupid than to start talking about blockchain and remove all doubt."³

And yet, Helium is the first (and so far only) example of a blockchain-based project that has made me wonder…maybe this actually makes sense? The transactional token associated with the Helium blockchain is abbreviated HNT (Helium Network Token). Participants in the network earn HNT by behaving in ways Helium has decided to reward. The hotspots, once active, “mine” (earn) HNT by doing one of four main activities:

• Challenge - Any hotspot in the network can issue a “challenge” to any other hotspot in the network. The challenge travels over the internet and arrives at its destination (or doesn’t). Challenges earn a teeny, tiny bit of HNT (0.95% of the overall tokens accrued any given day). They set the table for the next activity.

• Emit a Beacon - A hotspot that gets challenged emits a beacon that should be detectable by nearby hotspots if coverage is good. Emitting a beacon is worth a little more than issuing a challenge, but not much—hotspots emitting beacons gobble up about 5.31% of the overall pie. So far, we’ve proven that the challenger and the beacon-producer are both connected to the internet. We’ve only reached the “necessary but not sufficient” part of the checklist.

• Witness - Remember the remake of Mad Max? If a War Boy huffs paint before yelling “witness me!” in the open desert, and no one’s around to hear him, does he make a sound? A hotspot issuing a beacon is like one of those post-apocalyptic bald freaks screaming at the top of their lungs. Other nearby beacons that hear the shrieking and log it are called “witnesses,” and they get compensated a lot for lending their ear. A whopping 21.24% of all HNT mined in a day goes to Witnesses.

• Transfer Data - Yeah, the whole point of the network, right? Beacons that actually facilitate a data transfer are compensated for their efforts handsomely. 32.5% of the pie is reserved for them.

There are technically some other ways to earn HNT, but that’s for you to find out as you spiral deeper and deeper into this godforsaken rabbit hole.

The proportions tell the tale—witnessing and data transfer are the two main things that are rewarded and they are both primarily about receiving data rather than transmitting it. In that bike share example I shared above, the company isn’t sending firmware updates over the air. They just want to know where the dang bikes are. That’s what LongFi is great for. Therefore, nodes in the network that can listen to large areas consistently should be rewarded.

Now’s as good a time as any to point out the two huge asterisks:

1. (Almost) no one is actually using the network right now. So data transfer rewards are almost non-existent. I had a hard time finding any official figures, but on a recent podcast one of Helium’s early investors quoted a figure of ~$200/week of data credits currently being used…that’s a staggeringly low number. I don’t think anyone is earning much from facilitating data transfer currently.

2. Since (almost) no one uses the network, the fastest way to earn HNT is to witness other screamin’ beacons. That means if you set up a hotspot in a coverage desert where, theoretically, it’s providing a ton of valuable additional capacity to the network, it is barely rewarded at all because it has no one to witness. In the future, if the network becomes popular, these nodes will make money from providing the valuable utility of data transfer, but for today, there’s no incentive plant new flags in far-flung places unless you can bring a bunch of friends along with you.

The algorithm for rewarding witnesses is carefully designed to reward optimal geographic density—ideally, no nodes are closer than 300 meters to each other and four nodes fit into one 5km² hexagonal cell.⁴ You can use this handy tool to check if you’re in the “sweet spot”.

For a sense of just how much money active nodes are passively earning at the moment, have a poke around Helium’s coverage map. You can click on any node and see historical statistics—this random hotspot in a quiet suburban Austin neighborhood near where I live raked in 8 HNT today. As I write this, 1 HNT sells for $16.50 (up from around $4 a month ago). So that one hotspot is cranking out $132/day equivalent in imaginary internet money for whoever had the foresight to set it up. Last month, the most active hotspot in Austin earned just shy of $13,000 worth of HNT.

Cryptocurrency Speculation Has Its Merits

I already mentioned (almost) no one actually uses the network. Ultimately, without usage, this is all just another example of a speculative cryptocurrency bubble and it will, eventually, pop.

But what’s special about this particular bubble is how it’s channeling the attention of tens of thousands of people to personally invest in infrastructure that will power a more connected world. Despite acting out of rational economic self-interest, by the end of the year, Helium will have grown into a massive wireless network with hundreds of thousands of nodes. That’s a pretty great side effect of speculation. I wish more memecoins had similarly lofty missions.

If you believe the IoT revolution is only in its early days, and you think The Peoples’ Network will be able to provide a cheaper, better service than its large corporate competitors, then it’s easy to see how usage could increase over time. But if you are skeptical of either or both of those assumptions…maybe don’t buy a hotspot.

Even if you do believe those things, think twice before spending your precious fiat on Helium gear. As more nodes come online, the fixed number of HNT that are mined each day will be split among a bigger and bigger crowd. 10x more hotspots are backordered right now than are connected to the network, so you can reasonably expect a roughly 10x dilution in rewards for each participant once those all come online.

And it gets worse—in August (around the time a lot of these devices are planned for delivery), there will be the first of a regular series of every-other-year “halvenings,” when the monthly supply of new HNT rewards will be cut in half (from 5 million to 2.5 million the first time).

One mitigating factor in all of this is the price of HNT. In the last month it has more than quadrupled from $4 to over $16. As new supply becomes more constrained, that price could continue to climb because of scarcity. Additional data transfer would also cause upward price pressure. It’s difficult to say how much of that is already “priced in” to the $16 number. The network did not get 4x more valuable over the last 30 days. The price action screams speculative bubble and could collapse at any time. Or it could quadruple again next month and never look back. I have no idea and neither does anyone else.

Connect All the Things

Helium, and IoT in general, is mostly bound today by a lack of imagination. It’s a lot like satellite imagery analysis in that sense—there are ton of cool and reasonable sounding ideas but very little practical and valuable stuff happening in real life.

Almost every conceivable sensor has already been pre-fabricated in a form factor that already supports the LoRaWAN protocol and is available for purchase today. Water, pressure, temperature, motion, vibration, you name it—the lego bricks are in the big box, all mixed up, waiting to be stacked into something beautiful.

I look forward to seeing what folks come up with—new businesses, art installations, fun side projects, everything. In a world where data plans can costs just a few cents per month with no minimum commitments…a whole lotta cool stuff is going to get built.

¹ I use the phrase “HAM radio nerds” affectionately. My father keeps a collection of working 19th century telegraph keys, has known Morse code almost since he could spell, and holds an Amateur Extra License from the FCC (the highest of three permissions tiers). My grandfather was a radio operator in a bomber during World War II and bonded with my Dad over electronics tinkering when he was still a little boy. Radios were a big part of my childhood, too—Dad kept a CB radio in the truck that he used to talk to truckers on the highway. He also strung a massive 100+ foot antenna between the pine trees above our house with a homemade slingshot so that he could communicate with people all over the world with his telegraph keys. At one point, we built a working crystal radio together when I was really young, which I felt very proud of at the time. He used to take me with him to HAM radio meetups and competitions all over North Carolina, and while I didn’t exactly catch the “bug” (I am not a licensed amateur radio operator), I will always have a soft spot in my heart for the strange and wonderful world of HAM radio nerds—my father among them.

² Chirping gets its name from the wave pattern formed by the modulation—when translated to audible frequencies, it sounds like a bird’s chirp. Regular readers of this illustrious publication know that I work for a company that is launching a constellation of Synthetic Aperture Radar (SAR) satellites (Umbra). Chirping is a technique often employed in SAR systems as well. In SAR, chirping allows for more energy to be sent down at Earth and subsequently captured by the satellite’s antenna as it bounces back—more energy translates to better spatial resolution per unit of power. So, it’s somewhat like LoRa in that chirping is an energy-efficient way to transmit complex data over long distances.

³ Blockchain is a tricky topic; I think it’s an interesting application of cryptographic research that is rarely practical. At Azavea, my previous employer, they implemented the concept of a “breakable toy” in our professional development plan for apprenticing engineers. It was typically an instructive exercise that tangibly illustrated an important programming concept while allowing engineers to learn by doing. Bitcoin is like an unbreakable toy. As a paradigmatic, conceptual implementation of a “blockchain,” I think it’s beautiful. As a store of value or a digital currency, I think it’s a massive speculative bubble with extremely limited utility. I own a pathetic amount of Bitcoin, not because I think it’s intrinsically valuable as a currency or asset, but because I need a hedge against a scenario in which it goes up forever, and I have to live with regret that I didn’t buy any.

⁴ Helium is using Uber’s open source hexagonal indexing scheme, H3, to define its optimal density. Why not just use a regular square grid? Probably because hexagons are just way cooler. The person who first told me about H3 was Sina Kashuk, now the CEO of Unfolded (an exciting startup based on yet another open source project from Uber, Kepler). He will undoubtedly be pleased to see it in the wild. I still don’t get it. Sorry, Sina!!!

Five years ago, a gaggle of VC-backed startup CEOs sat on conference stages in blazers and t-shirts smugly parroting each other with lines like, “Satellite imagery is quickly becoming commoditized. Analytics built on top of satellite imagery is the real opportunity.”

In the years since, a lot of those CEOs have lost their job or their company or both. Most of the high profile CEOs from the ~2014-15 golden age of geospatial analytics have either stepped down or been forced out.

To be fair, by the end of 2016, I was drinking the same Kool-Aid as the rest of ‘em. We all thought we were on the cusp of a revolution in the commercial earth observation industry. 2017 was going to be the big breakout year when a flood of cheap, awesome, civilian satellites would finally hit the market. All the charts looked like hockey sticks back then, but I never stopped to wonder why the axes weren’t labeled…

☝️ Thank you, BlackSky, for the pity “like.” I owe you one.

Planet was on track to hit “Mission One” in 2017 (they did) and the second Sentinel-2 bird was planned for launch (it did). AstroDigital launched its first satellites. A record 104 satellites were launched in a single day that year (the majority of them remote sensing satellites) and Planet even acquired TerraBella out of the jaws of Google where it wasn’t doing anything for anybody.

Fast forward to today and we’re still “on the cusp of a revolution in the commercial earth observation industry.” Maxar and Airbus are launching big constellations this year, Planet continues to add to the largest constellation of earth observing satellites ever assembled, Satellogic and BlackSky are nipping at their heels with big globs of money to spend, and of course the SAR Wars are heating up.

As exciting as that all sounds, very little has *actually* changed for consumers of satellite imagery. It’s still ridiculously difficult and expensive to buy high resolution, tasked imagery. Investor expectations are at all-time highs as reflected in the valuations of growth stage companies going public via newfangled Special Purpose Acquisition Companies (SPACs) at generous multiples of projected EBITDA five years in the future.

It feels a little like 2016 all over again. If this time is going to turn out differently, it’s going to take more than just money (although the money will certainly help). It will require delayed gratification and very longterm thinking from the industry as a whole. We have to think of this as a positive-sum game or else we’ll all be chasing bigger and bigger slices of a pie that stubbornly refuses to grow.

And even when we’re successful at making high resolution, timely satellite imagery affordable and accessible to a broad audience…it won’t be a commodity product. At least, not in my lifetime.

If Satellite Imagery Were Becoming Commoditized, You Could Buy It

I was recently listening to Aravind Raves’ fantastic podcast that aims to “demystify space technology." His guest took a brief respite from flogging his “space manufacturing” startup to make a lazy argument I’ve heard people espousing for the last five years: earth observation is a crowded, low-margin market of increasingly undifferentiated companies.

Oh yeah?

If it’s such an efficient market with limited prospects for disruption (and therefore outsized returns), tell me…how do I go about buying a satellite image today? How much does it cost? How long will it take for my order to be filled? How will the data be sent to me? Let me take a crack at it:

• Hopefully you have already established close relationships with sales people at one of the major satellite imagery suppliers. If not, you’re gonna have a bad time. Either way, get ready to send some emails.

• Price depends on the company, last quarter’s sales numbers, this quarter’s sales numbers, the clout of the sales person, and the astral alignment of your horoscope. You’re about to spend anywhere between a few hundred dollars for an archival image to several thousand for a tasked image. That is, if they let you buy anything at all.

• I would budget a few weeks to a few months for this adventure if I were you.

• Are you familiar with File Transfer Protocol?

The truth is that supply is still very constrained at the highest resolutions, and that’s mostly what people are actually willing to pay for. Despite the fact that lots more satellites are getting launched this year and next, it’s not at all clear to me that the cost basis of most of those constellations will allow them to monetize their imagery at a significant discount to what’s currently being offered (especially if international governmental demand for that imagery continues to outpace production).

Not to mention it takes a special kind of investor to listen to a CEO pitch them on the idea of leaving millions on the table in the short term in order to grow the market over the long term. To the companies going public and joining the quarterly earnings rat race: I salute you. Godspeed.

So, yes, the cost of launch is going down. And so is the cost of manufacturing. But the cost isn’t, like, zero. These are still big-ass hunks of metal that have to be flung into space on the nose of even bigger-ass hunks of metal and then kept there, against the pull of gravity, for years at a time. Space-based Earth observation is a capital-intensive business where even the very best inevitably lose contact with (or control of) their assets in space. Those costs all get factored into the price of an image.

I personally did not appreciate how difficult it is to build and launch satellites before joining Umbra. It’s one thing to talk about it, and a totally different thing to feel the pressure of delivering on it. Every time I see a successful launch, I don’t feel envious—I feel excited. With each successful mission, we take one baby step closer to a competitive, efficient market. We’re all walking a tightrope of technical risk where the line between financial reward and ruin can come down to a single roll of the dice.¹

Satellite Imagery Suppliers are Like Dairy Farms

I grew up a few miles from a small dairy farm in North Carolina, Maple View Farm. When I was a kid, my Mom would schlep me up the long gravel driveway to the Nutter family’s farm house where a refrigerator waited, sweating in the summer heat, full of the most incredible chocolate milk you can imagine. The stuff is so thick you’d break your neck diving into a pool of it.

The Nutters employed the honor code back then—you dropped your cash in a basket, took your glass jugs of milk out of the fridge, and then returned your empties the next time you visited.

Later on, Maple View opened an ice cream parlor at the top of the hill overlooking the farm. There is no better ice cream on the planet. You can sit on the front porch of the ice cream parlor in one of their rocking chairs and look out across cornfields and pastures, waffle cone in hand, ruminating on everything and nothing. Maple View does not just sell milk and ice cream; they also sell peace and stillness.

Anyway, you probably see where I’m going with this. Today’s satellite imagery companies are a lot more like Maple View Farm than they are like Dairy Farmers of America.²

A big part of the reason satellite imagery is not a commodity is that, unlike raw milk, it’s pretty damn hard to figure out how to produce. This is one of those rare cases where the phrase, “It’s not rocket science,” can’t be employed, because it literally is rocket science.

To extend the analogy, let’s start by imagining satellite imagery is hot milk straight from the teet. Analytics are refined dairy staples, like milk and butter. And software to interpret those analytics is ice cream. The best ice cream isn’t all about the dairy that goes into it; it’s at least as much about the other ingredients you mix in. But the quality of the dairy still determines the quality of the ice cream. Maple View makes the best ice cream on the planet in large part because they also make the best milk on the planet. That same dynamic is why satellite imagery providers who build out analytics teams will always have the upper hand on their customers and partners when competing head-to-head.

Like Maple View, satellite imagery providers are happy to sell their excess inventory to customers who subsequently create competing products. The world is a vast place and it’s not a zero sum game; there’s more demand for milk than Maple View can squeeze from its prodigious heifers. But if one of those customers were to start crowding into Maple View’s territory with their own ice cream parlor franchise…or if, suddenly, one of their customers launched a new product that became extremely popular…Maple View could always just cut them off or raise prices or undercut them with their own copycat product, or some combination of the three. After all, they own the cows. Go raise your own if you don’t like it.

Not All Satellite Imagery is Created Equal

Even if access to satellite imagery had gotten significantly easier over the last five years (it didn’t), the idea that satellite imagery was becoming commoditized would still have been wrong. I’ve heard people employ the phrase, “a pixel, is a pixel, is a pixel,” to color the idea of satellite-imagery-as-commodity. It sounds reasonable enough—it shouldn’t matter which company produces the pixels as long as they meet the spatial, spectral, and temporal requirements of the customer.

This past week on Twitter has been a great opportunity to compare apples to apples. Everybody and their cousin scrambled to get a shot of the Ever Given, the massive ship stuck sideways in the Suez Canal gumming up world commerce. Here’s a great tweet from my friend Ignacio (go follow him on Twitter) with a few such images conveniently placed side-by-side for easy comparison:

Look at how different they all look! Compare those shots to this one from Maxar:

Something you should know, too—anything that makes it to social media has been tortured to death to make it as high-contrast, low-blur, Instagram-worthy as possible. This is what everyone looks like at their best.

Among the providers referenced above, Satellogic wins “most improved” in my book. But Maxar and Airbus, the big incumbents, still have the goods.

For this type of use case (is the big ship still stuck in the tiny canal?) a pixel is a pixel, sure. But if subsequent questions come up like, “how many excavators do they have on site?” or, “Which class of tugboat are they trying to pry it loose with?” suddenly the imagery doesn’t seem so fungible.

In my opinion, the uses for satellite imagery with the most untapped potential tend to be about monitoring change. Any time you’re dealing with a time series of satellite images, you wind up facing down some of the more esoteric characteristics of satellite imagery, like absolute positional accuracy, co-registration, sensor calibration, and incidence angles.

In fact, there’s a whole subdomain of remote sensing research dedicated to the pursuit of making satellite imagery “analysis ready,” and it’s a concept that prominent geospatial analytics firms like Descartes Labs have more or less staked their brand on.

Just like store-bought milk has to be pasteurized and packaged and stored at a steady temperature below 40ºF, satellite imagery must be processed before it’s ok to ingest. And even at that point it’s pretty boring…like a glass of milk. Plus, just like no amount of processing will turn goat’s milk into cow’s milk, no amount of processing will turn low resolution data into high resolution data or one phenomenology into another.³

Less of an Oddity, but Still Not a Commodity

There’s a saying in finance that goes something like, “Being right about something but getting the timing wrong still just means you are wrong.”

I personally believe that high resolution, low latency, low cost satellite imagery is finally going to become available to civilians in the immediate future. I am working every single day to help make that happen, alongside a cadre of brilliant, highly motivated colleagues (not to mention a growing list impressive partners, some of whom are also launching satellites).

Truthfully, I can’t help but harbor more than a little fear that I am falling victim to the same delusions that blinded me in 2016. A handful of people much smarter than me have told me point-blank that the approach we’re taking at Umbra is naive and misguided. I’m not worried about being embarrassed if it doesn’t work out, but I don’t want to feel like a fraud, either. I am determined to prove them wrong.

Luckily, I’m not in this thing to get rich quick or help flip a company in a few years. For one thing, I am 100% convinced that satellite imagery is an irreplaceable tool in the fight against climate change, which is the central issue that will define my generation’s paragraph in the annals of history.

How else, besides using satellite imagery, can you:

• Verify human activity and natural capital on a global scale, even at low fidelity?

• Characterize the extent and severity of natural disasters as they unfold in real time anywhere on Earth?

• Monitor conditions on the ground in places made inaccessible due to geography or political regime?

• Inventory land use and track its changes over time?

How can we make planetary-scale decisions unless we have planetary-scale data to inform them? I don’t think satellite imagery is the answer to the climate crisis, but I think it’s going to be a necessary ingredient of any answers that emerge over the coming decades.

Even when we make it up onto the mountain top where high resolution, timely satellite imagery grows on trees…that doesn’t mean it’s going to turn into a commodity business. Some systems will be more suited to monitoring and others to foundational mapping; some to long-term change detection and others to instant verification; some to classifying stuff and others to counting things.

There will be visible light sensors, and infrared sensors, and ultraviolet sensors, and hyperspectral sensors, and radar sensors, and on and on. None of them will be “analysis ready” without a tremendous amount of specialized work, and the synthesis of these data into novel information will be an active field of research for as long as I’m living and long after I’m gone. Newer, better, smaller, cheaper, tougher, smarter systems will continue to be developed ad infinitum, and someone will always have a technology edge, however briefly.

So next time you hear someone say that satellite imagery is becoming commoditized, try asking them, “Oh yeah? What about Maple View Farm?”

¹ I got the chance earlier this year to interview Walter Scott, the founder and CTO of Maxar. Refining that conversation into a blog post will take a while; I didn’t anticipate starting a new job right around the same time I interviewed him. But one of the big takeaways from that opportunity was just how close Maxar came to insolvency. Their first two satellites failed. By the time they were on the platform counting down to launch for their third bird, ten years had already passed and $42M had gone up in smoke. They were hundreds of millions of dollars in debt. If QuickBird had blown up in the sky in October of 2001, the largest and most successful American satellite imagery company today would likely not exist. But it didn’t blow up!

² The largest dairy producer in the United States is a farming co-operative comprised of thousands of independent producers. I like my milk like I like my Scotch: blended.

³ Unless you are just having fun, stop trying to use style transfer to make Planetscope look like SkySat or make SAR images look like optical images. It’s a fun party trick, but there are so many more interesting applications of machine learning that feel under-explored, like replacing traditional interferometry methods with deep learning or using GANs to make seamless mosaics out of patchwork to bring the cost of producing basemaps down.

A lot of folks don’t realize that Elon Musk got his first big break as an entrepreneur building mapping software. He and his brother Kimbal started Zip2 in 1995, which provided phonebook-style business listings along with embedded maps for a little wow factor. Part of me is envious of the days when your business plan could be as simple as, “we’re gonna put the phone book on the computer.”

After raising a few million dollars in 1996, Elon was removed as CEO and replaced by Rich Sorkin (who is now the CEO of yet another mapping startup, Jupiter Intelligence). In 1999, near the height of dot-com hysteria, Zip2 sold to Compaq for $307M.

The moral of the story is simple: if you wanna get rich, start a software company focused on maps.

I was perusing Elon’s tweets a while back when I stumbled across an interesting response from a guy named Robert Scoble. Since encountering him, I’ve learned that Robert is a professional internet provocateur. He stumbles into interesting tech or tech-adjacent topics, and like a drunk relative toasting the happy couple at a wedding reception, proceeds to embarrass himself with a confidence that is truly awe-inspiring.

We got into a fleeting disagreement about whether or not Tesla’s mapping work was soon going to eclipse Google or Apple. It all began when I glimpsed this exchange:

“Tesla will soon crush Google Maps,” Robert proclaims.

On the rare occasion I read something so ponderously stupid that it hurts me physically, I try to use it as an opportunity for reflection. So, I got to thinking…why do I find this claim about the future of mapping supremacy so objectionable?

Below, I talk it out.

Imagining the Future of Autonomy

I’ll be honest. Elon Musk bothers me.

One of the reasons I’m uncomfortable with Elon as a role model is the way he has poisoned public discourse about autonomous navigation. He sells Tesla customers something called “full self driving capability” for $10,000 as an optional upgrade. You could be forgiven for thinking that would get you a car that can drive itself…fully. Autonomous navigation is traditionally rated on a scale of 0 to 5, and Tesla’s famed auto pilot ranks as a 2.

When it comes to claims about the impending self-driving car revolution, Elon simply cannot help himself. He has no idea how close or far away Tesla is from achieving fully autonomous navigation, but he is always certain it’s next year and has been since 2016.

“I really do not see any obstacles here,” Elon said of the path to full autonomy a couple weeks ago. Interpreting that statement in the best faith possible would require assuming that he meant to finish the sentence, “…because I’m not paying any attention at all.”

And yet…he’s a phenomenal recruiter. Tesla’s “Senior Director of AI” is a former Stanford researcher named Andrej Karpathy. His Software 2.0 talk is a glorious manifesto on the way to build organizations around modern best practices for machine learning. If it’s possible to achieve fully self-driving cars with the state of technology today, I bet Tesla will figure it out. But I’m really, really unconvinced it’s possible.

One of my favorite researchers on the topic, Dr. Missy Cummings of Duke University, had this to say about the idea circulating last year that Tesla “robo-taxis” would be pulling out of driveways everywhere by 2021:

Frankly, I don’t think we’re very close to self-driving cars based on my experiences trying to get deep learning algorithms to trace buildings and roads in satellite imagery. It’s not that the algorithms are weak; on the contrary, they’re exceptionally impressive (and tireless, to boot).

But! They are fragile. Specifically, they are vulnerable to “edge cases”—decisions that are rare, often unintuitive, and inherently unpredictable. Think about all the time you’ve spent riding around cars—there were probably one or two times when you really weren’t prepared for what happened and it scared the living daylights out of you. A family friend once had a deer jump off a bridge and kill itself on his windshield as he passed underneath; show me that in your training data.

What I mean is this: even if we can get 98% of the way to full autonomy this decade, the remaining 2% is so high-stakes that it doesn’t really matter. It’s an existential problem, in my opinion, for people looking to dedicate their career to *this* particular application of machine learning when there are so, so, so, so many other interesting applications for the technology that are better suited to occasionally fallible automation.

Assuming I’m Wrong (and I Often Am)…

Ok, so I just said we aren’t close to self-driving cars. I hope I’m wrong. And assuming I am, let’s think through how self-driving cars will function from first principles:

• I don’t need a perfectly detailed 1:1 map of the world in order to drive. In fact, I don’t need a map at all. Just based on road signs and a basic understanding of geography, I can navigate pretty well even over long distances.

• Autonomous vehicles, if they reach parity with humans, will at least have the same level of agency and competence when navigating. They may augment that ability with better, faster, further-reaching data to achieve better-than-human results. But if the 5G connection gets interrupted and they’re on their own…they should be able to keep functioning at a reasonable level.

• Mapping the world in low definition is a hard enough task that no one has done it exhaustively yet. And for the groups like Google and the National Geospatial Intelligence Agency (NGA) that have tried, their annual budget is in the billions of dollars. It’s just not logistically feasible to get high definition data of everywhere on earth all the time updated regularly.

• Therefore, as the technology behind autonomous navigation improves, I would expect car makers to focus their differentiation on directly-observed, in-car navigation systems supplemented by outsourced map data of varying levels of fidelity.

• By Tesla’s very own admission, they’re focusing on a “vision-based” approach to navigation and counter-positioning themselves against Waymo, who they say is taking a “map-based” approach. Robert Scoble—if you’re reading this—watch the video linked in the prior sentence.

That’s a longwinded, terrible-use-of-bullet-points way of saying: Tesla isn’t going to “crush Google Maps,” and they aren’t even trying to. Even if they were trying to, there’s another big problem: most of the raw data directly observed by sensors on their cars is probably protected by personal privacy laws (even if all Tesla plans to do is use it for R&D purposes).

The California Consumer Privacy Act (CCPA) went into effect last year and is the American analog to GDPR for car manufacturers. Like it or not, California is a large enough “sub-market” within the United States that whatever regulations are enforced there tend to affect the way cars are manufactured everywhere else across the country. That law is not messing around: according to CCPA, you even have a right to contact Tesla and make them delete data they’ve collected about you.

I think the upshot of CCPA, and others like it to come, is that car makers are facing an enormous, uncertain liability if they collect troves of personal data (and I would imagine that video of your driving habits is very personal and sensitive). Therefore, they’re greatly incentivized to work with 3rd-party vendors who either simulate the data (“synthetic training data” as its sometimes called) or go to great lengths to secure direct ownership of the data and provide it to car makers cleanly.

The same thing has happened in every other facet of the mapping industry: eventually, the collection of basic, foundational data is competed down to a low-margin services business and the Big Money™ is made in applications that license that data.

The Finish Line

I believe the autonomous vehicle investment bubble is popping. There are already some signs…in December of last year, Uber divested of its self-driving unit for ~$4B (including forking over $400M in cash to help the new owner), which was down from its prior valuation of $7.5B just a year prior. If anyone has a strong incentive to figure out self-driving robo-taxi fleets, it’s probably Uber.

Autonomous vehicle folks are looking ahead to a long, slow slog of incremental progress under the weight of Elon’s hollow promises. But it’s not all bad—the insane investment in the space over the past five years is still likely to drive acceleration in technologies like edge computing, 3D reconstruction and depth inference from video, and LiDAR sensor manufacturing.

What’s not going to change is the fundamental way we map the world or the players who dominate that pursuit. Autonomous vehicle manufacturers may provide an economic engine that incentivizes broader and more frequent collection of foundational map data, but they don’t want to weigh down their absurdly optimistic valuations with scores of software engineers and field techs rebuilding mapping infrastructure that is already on offer for fairly affordable prices.

Tesla won’t crush Google Maps. Apple on the other hand…

After a little over five years, I’m leaving Azavea. Some day, I’ll write more about my experiences there. I loved every day at Azavea. Parting ways has been bittersweet, to say the least.

Alas, it’s time for a new adventure. Later this month, I’ll start a new position as Head of Product Success and Impact at Umbra, a satellite imagery startup.

I’ll talk a lot about Umbra later on in this post, but first, allow me a brief digression.

Why Commercial Earth Observation is Such a Dysfunctional Industry

I caught the writing “bug” earlier this year when I published a diatribe on the commercial satellite imagery industry, “The Commercial Satellite Imagery Business Model is Broken.”

In my job at Azavea, I bounced between the roles of customer, vendor, and partner to many of the major US- and EU-based Earth observation companies. More often than not, those experiences left me feeling disrespected and frustrated (even as someone who was actively working to help them sell more imagery).

As I wrote in the original post, the industry’s brokenness is the logical outcome of two complementary economic forces:

• The historical capital requirements to build, launch, and run satellite constellations (and the subsequent governance structures that arise from aggressive fundraising)

• The near-monopsony of the US DoD, and the stagnant handful of firms that have made up the large majority of the commercial market to date (i.e. Google Maps, Apple Maps, Bing Maps, etc.)

So, why dig myself deeper into an industry that has caused me so much heartache and that does not appear, to me, to be growing quickly?

Because we’ve only just crossed an inflection point in the cost to serve timely, high-fidelity satellite imagery that will dramatically expand the number of customers who can profitably incorporate information derived from satellite imagery into their decision making processes.

At least, I think.

Limitations of Optical Imagery

Even if the big players expressed a genuine interest in selling data into commercial markets, they probably couldn’t afford to, which is the central lament of my earlier piece on the subject. Speaking specifically about folks like Airbus, Maxar, BlackSky, and Planet, who mainly sell optical imagery, they’re up against some existential constraints.

Capacity

Optical imagery capacity is artificially limited by incessant cloud cover and available daylight hours—a great majority of the world is obscured at any given moment by one or both of those two confounding variables. The consequence is that much of the time that optical imagery satellites spend orbiting the Earth is a total waste.

Limited tasking capacity drives the price of imagery up—you may think you’re paying for imagery, but the satellite imagery providers are thinking in terms of time. The more average passes around Earth that are required to complete a single order, the fewer orders each satellite can process, and the more it must cost at a minimum for all orders.

The satellite imagery market today has settled into a lopsided economic equilibrium entirely concentrated in two buckets: customers that can reserve a lot of capacity upfront (e.g. the military) or customers that are extremely flexible about the timing of image captures (e.g. commercial customers building satellite imagery basemaps). All others need not apply.

Physics

The laws of physics are working against these optical imagery providers. Satellites orbit at a high altitude above Earth, even though they’re in what’s called “low Earth orbit.” These instruments tear across the sky at 15,000+ mph while suspended hundreds of miles above the Earth’s surface, and yet the pictures they take can resolve street markings and car windshields on the ground below.

The lenses that high resolution satellites use are enormous and so are their housings—WorldView-3, for instance, is over eighteen feet tall and clocks in at over three tons…

Big satellites translate to big build and launch costs. Worldview-3 cost about $600M to get operational (even before the costs associated with financing those activities). Maxar’s next generation of satellites, “Legion,” will be smaller and cheaper (about one third the size of WorldView-3), but they still anticipate the constellation will cost hundreds of millions of dollars. That’s a lotta cheddah.

Demand

The “medium” resolution persistent surveillance experiment that Planet has been running for the last decade has not paid off (at least, not yet). The central idea is a very compelling one—if you take a picture of the entire landmass of Earth once a day, you don’t have to deal with the opportunity cost that drives tasking-oriented companies to keep the price of satellite imagery so high. A variable cost suddenly becomes fixed, and you can build a different sort of business model—one that’s more like a software business with infinitely small marginal cost to serve each new customer.

But, the imagery is more than ~10x lower resolution than the kind of stuff WorldView-3 produces. And most people don’t want that.

Planet’s acquisition of the high resolution constellation SkySat (née Terra Bella) from Google, and their subsequent decision to drop the orbital altitude in order to achieve ~72cm resolution (vs. 90cm originally), conveys a clear message. I strongly believe that for commercial purposes, it’s super-high-resolution-or-bust.

It suggests a demand curve that is not, shall we say, evenly distributed across spatial resolution:

Here’s a whole tweet storm on the topic about it if you want go dig deeper:

The punchline: optical imagery providers today don’t support monitoring use cases for anything short of a military budget.

Enter: Umbra

Umbra has designed and manufactured a very interesting type of satellite that leverages a technology called Synthetic Aperture Radar (SAR) to measure the Earth. Earlier this year, I wrote at length about why this genre of Earth observation is so exciting compared to optical imagery, and if you’ve never heard of SAR, it’s probably worth checking out that article since I’m not going to go in depth here. To crudely summarize:

• SAR renders “images” of Earth by beaming down radar pulses and measuring the response. This technique images the Earth equally well during day or night. SAR can also “see” through clouds. As a result, every image a SAR system captures of the Earth is relatively unobscured, and which gives SAR constellations much greater freedom to maneuver around the globe for efficient tasking vs. their optical counterparts.

• Unlike optical imagery, SAR doesn’t require a large lens to resolve images. Rather, it uses an antenna, which can be folded up during launch. That means you can squeeze higher resolution from a smaller package, which in turn means you can manufacture and launch these satellites relatively inexpensively.

• There’s also a technique called interferometric SAR (InSAR for short) that can measure extremely subtle changes in elevation over time based on the phase shift of the radar return. You know how police sometimes use radar guns to tell how fast you’re going? That’s a depth finding application for radar. InSAR works like a depth finder, allowing analysts to precisely measure which parts of the world are getting closer or further from the satellite (“uplift” vs. “subsidence”). If you build a house, the roof is closer to space than the foundation was when you started, which shows up as uplift. If an aquifer drains, an entire valley might literally sink a few centimeters, which shows up as subsidence. It’s a nifty feature of SAR that is very difficult to replicate with optical imagery and allows for relatively straightforward change detection (called “coherent change detection” when specifically describing InSAR-derived changes).

Umbra is not the first startup to think of commercializing SAR satellites. In fact, they are the underdog (at least on paper):

• ICEYE, a Finnish startup, launched their first prototype satellite in January of 2018. They currently have 2 satellites on orbit, with 16 more planned. They’ve raised over $150M, with $87M of that coming just a few months ago in a Series C financing.

• Nipping at their heels is Capella Space, an American startup which just launched their first operational satellite in August of last year. They plan to build a 30-satellite constellation over the coming years and have raised over $80M from top-tier investors (and I would not be surprised if a Series C or some alternative financing is imminent now that they’re streaming data down).

• Earlier this year, PredaSAR announced that they raised $25M in seed funding (what do these designations mean anymore?) and swiftly packed their leadership team with retired military generals in a bid to position themselves as an up-and-coming defense contractor. I love the name—it’s very…earnest.

Betting on the Underdog

If you search around for Umbra, you won’t find much. Founded in 2015, the company was self-funded for years and kept a low profile as its founders, David Langan and Gabe Domincielo, stayed quietly focused on delivering early versions of what would become their marquee satellite design. They don’t make public declarations about launch dates, or how many satellites they plan to launch, or really much at all in the way of forward-looking statements. I like that. A lot.

They have a lot to be proud of, despite opting not to call much attention to themselves. The technology they’ve built is ridiculously impressive. Weighing in at about ~65kg, you could probably check an Umbra satellite on an international flight. Yet, despite the small size, the forthcoming constellation of satellites, once on orbit, will be able to capture images of the Earth at resolutions below 25cm GSD.

There are two important innovations that set them apart from the competition:

1. Umbra will be able to capture insanely high resolution images using a remarkably tiny satellite. They’ll use a proprietary radar that allows for greater bandwidth, resulting in images rendered at a higher spatial resolution. Specifically: their radar generates a signal greater than 1,200MHz of bandwidth (as opposed to ~300MHz for ICEYE and ~500MHz for Capella despite those satellites being significantly heavier).

2. Their other big innovation is described in their publicly issued patent. It has to do with the antenna design itself—not only does the >10m² surface fold up into an extremely tiny package, but it’s also designed with a shock-absorbing rib/mesh combination that dampens interference caused by the satellite’s in-orbit maneuvering (resulting in a higher average throughput of images per day per satellite, since less time is wasted waiting for the antenna to stop vibrating).

A Nascent Market

I consider myself a fairly competitive person. I want to crush the competition. But I’m not referring to Capella or ICEYE—in fact, I deeply admire both of those organizations. I hope they’ll be closer to allies than foes, because SAR is not a “winner-take-all” market anymore than sneaker manufacturing is.

The “competition” in SAR’s case is a massive, untapped market of people who have no idea they would benefit from SAR and would find this blog post even more tedious than you do. As with most “innovative” technology, the biggest competition is the status quo:

• Every bank that sends staff to construction sites once a year instead of checking on progress remotely once a week;

• Every zoning and tax office waiting for people to call in with tips about unsanctioned building activity instead of monitoring for change proactively;

• Every counter-party to a contract guaranteeing land will remain undeveloped and hoping for honest self-reporting;

• Every aid and relief organization trying to understand the situation on the ground during a major flooding event with no access to real-time satellite imagery due to clouds;

• Every journalist working to expose human rights violations on land they’re disallowed from visiting;

• Et cetera

The broader technological trend Umbra is but a small participant in is the inevitable digitization of everything. Marc Andreessen’s opus magnum Why Software is Eating the World came out when I was a freshman in college, although I wasn’t paying attention at the time. In the years since, I’ve heard a similar sentiment echoed countless times—perhaps my favorite articulation of the idea is from an interview with Michael Mayer, founder of an automated coffee subscription business called Bottomless Coffee:

Anytime you can make something important legible to computers, you can change the world. If you look at all of the revolutions that happened after smartphones got significant penetration, it was because suddenly people's locations were legible, and the world around them was legible with cameras.

I’m about to get highfalutin, so bear with me. I believe the most urgent gap in human knowledge at the moment is a high fidelity, global, timely understanding of land use and land cover change. As I wrote earlier this year, I am grateful to work in an industry with a unique role to play in the coming climate crisis:

Understanding and managing land is where I believe geospatial technology will play its most highly leveraged role over the coming decades. We’ve been cataloging the Earth’s contents in increasingly fine-grained detail for millennia — now we’re going to be forced to make dramatic choices about its composition over the next decade or face entirely avoidable and horrifying consequences.

I don’t care about putting satellites in space because it sounds cool. I care about it because it’s the only way I know of, logistically, to gather the information necessary to make informed decisions about land use on a truly global scale.

Risks

There’s a lot that can go wrong. Understanding the key risks of any endeavor, and being able to articulate how you’re mitigating them, is part of the fun of working at startups. Here are some I’ve thought a lot about:

Technical

I don’t know how to build satellites. I’m now going to work for a satellite imagery company. I have no more understanding of how SAR satellites are manufactured than I do of how artificial neural networks work—I’ve seen the output firsthand, and I’ve read about the design, but I still couldn’t personally make one with an unlimited amount of time.

When trying to understand Umbra’s technical risk, I called a few people who knew the industry inside-and-out and were willing to talk to me off the record. They all said the same thing: Umbra has a spectacularly talented and experienced hardware team, and there’s a genuine innovation at the core of their work. That was enough for me.

Regulatory

The largest US-based satellite imagery provider, Maxar, first asked the U.S. Commerce Department for permission to image the Earth at 25cm resolution in 1999 (back then they were called EarthWatch Incorporated). They finally granted permission in 2014. That’s the kind of regulatory uncertainty that can suffocate a startup.

Space-borne satellite imagery is subject to regulation from a laundry list of agencies because of the sensitivity of the data:

• NOAA officially regulates commercial Earth observation through the Commercial Remote Sensing Regulatory Affairs (CRSRA) office.

• The rockets used to get satellites to space are licensed by the Federal Aviation Administration (FAA) as are the commercial launches themselves via the Office of Commercial Space Transportation (which is, for reasons beyond me, abbreviated AST).

• In order to send information down from space, you need to have radio frequency (RF) allocated for your communications, which is regulated by the Federal Communications Commission (FCC).

• And even once you’ve navigated the gauntlet of the first two agencies, because you are trafficking data that can be used for military purposes, you are subject to International Traffic in Arms Regulations (ITAR), which is enforced by The Directorate of Defense Trade Controls (DDTC) under the U.S. Department of State.

• But wait! There’s more! The U.S. Department of Commerce has authority to regulate foreign exports of satellite imagery, usually acting through the Bureau of Industry and Security. This authority is directly spelled out in the original Land Remote Sensing Policy Act of 1992 that created the possibility for an industry to exist in the first place.

That’s a lot of red tape!

Successful commercial Earth observation companies coming out of the EU, China, Russia, Japan, India, and elsewhere are suddenly making commercial Earth observation much more competitive. And the DoD is not about to start relying on foreign-based companies for sensitive military intelligence. So, in part because of strategic national security interests, and in part because of pressure from U.S. politicians who understand the U.S. is trending toward losing its edge in commercial space technology, I feel that we’re more likely to see U.S. regulators motivated to work with American remote sensing firms rather than against them. But…who knows?

Financial

While preparing for this job, I interviewed a dozen or so people in leadership positions at commercial satellite imagery companies (past and present). The number one theme from those discussions: Earth imagery is, historically, a low-margin business that costs a lot to get into. Many high-profile companies in the past have gone bankrupt or been forced to sell, and almost every firm with staying power has at some point flirted with insolvency.

Umbra isn’t very public about its fundraising compared to it’s rivals, and it’s true that they have not raised as much money as their higher-profile competitors (although, Umbra did quietly file with the SEC for their recent $32M in new funding). I don’t mind this. In fact, I like it. I believe Umbra has charted a path to profitability by creating an incredibly efficient system. A single Umbra satellite will be able to collect more high resolution data than anyone else on a per-unit basis, meaning lower capex (and more margin). In the end, it all comes down to unit economics, and Umbra will have the best in the business.

Umbra has resisted the temptation of building big sales and analytics teams before even launching a satellite, which not only reduces their overhead but also allows them to partner with great analytics firms without the messy conflict of interest other satellite providers have to navigate.

Profitability equates to indestructibility. And historically, in this industry, the spoils go to the last company standing, not the one that raises the most money or does the most press.

Market/Demand

Have I mentioned that I don’t personally believe there is any significant commercial market for SAR yet? I would bet that civilians spend more on decorative cheese board sets than they do on SAR imagery at the moment.

However, the defense market is fairly mature—the U.S. military has been consistently using this technology since the 50’s! The data is nothing new to them, even if procuring imagery from commercial vendors is. Therefore, the short-term success or failure of every single U.S. SAR startup, in my opinion, will hinge entirely on their attractiveness to the U.S. DoD.

If, in the next 3-5 years, a solid foundation of DoD revenue does not materialize, I think the domestic SAR market is likely to collapse under the weight of its own astronomical expectations. If it does materialize, however, I think there’s room for at least a few companies to prosper. Among those, the companies who have the conviction and patience to continually invest in growing the commercial market will be the big winners over the next decade.

In 10 years, the commercial SAR market will either be billions of dollars annually, or it will still be vanishingly small. I doubt we’ll wind up in the middle.

I’m betting my career on the emergence of a massive commercial market. And I believe very strongly that Umbra is the best positioned company in the market to ride that wave.

Ethical

I asked on Twitter recently if anyone knew of a good researcher studying the ethics of civilian Earth observation. I mostly got crickets…

The power to task a satellite (almost) anywhere on Earth and capture a super-high-resolution image comes with an enormous responsibility. There are so many nefarious uses for the data, it’s not even worth listing them out. The obvious implication is that if anyone is allowed to task the imagery that can afford it, then some will use it to surveil others for illegal or immoral purposes (e.g. if you can see when someone is or isn’t home by whether or not their car is parked outside, you can use that information to stalk, burgle, attack, or extort them).

While the privacy concerns associated with high resolution imagery are often overstated, occasionally to a ludicrous degree, the underlying threat of weaponized surveillance is a serious one.

I believe, perhaps naively, that information wants to be free. I believe that Umbra will make access to information about the world freer than it is today, and I believe that’s a worthy mission to dedicate one’s life to. I hope to directly influence the open data and charitable giving policies the company plans to enact, and I cannot wait to see the countless, inventive uses for SAR imagery to come out of the next few years to the benefit of humanity.

If you know of resources on the subject of privacy, surveillance, and ethics, I would welcome any help researching the topic. So far, I’ve found a lot of material on the internet and mobile phones, but not much on Earth observation.

Having My Cake and Eating It, Too

There’s one other detail I neglected to mention. Joining me from Azavea will be Chris Brown and Matt Williams, two of the most talented people I’ve ever worked with. Chris was the lead software engineer on my team at Azavea, and Matt was the lead designer at the company. The three of us will have a chance to help Umbra build the best data distribution platform in the industry, and I can’t wait for the challenge.

Life is short, and in years like 2020 it can feel all the shorter. I am grateful that through luck and happenstance I will have a chance to continue working with two people I love, admire, and trust deeply. But, I also look forward to making some new friends at Umbra—hopefully ones that, like Matt and Chris, I would be content to keep working with for the rest of my career.

To a new chapter, a new year, and a new dental insurance provider. Ad astra!

If you work at the intersection of maps and software, at some point in your life you’ve probably heard yourself muttering some version of the following analogy to a stubbornly confused family member:

You know Google Maps? What I do is, like, build little pieces of Google Maps over and over for people who need them but can’t just use Google Maps because they’re not allowed to for some reason, or another.

Of course, you never quite put it that way, because it would hurt too much to admit it to yourself so plainly.

People have half-heartedly compared their profession to Google Maps at family gatherings for well over a decade, because since inception, Google has consistently created the best consumer-facing maps in the world. And I’m not just talking about coverage—they are top of the class in turn-by-turn navigation, geocoding, satellite imagery layers, street level imagery, place data, and on, and on.

The title of today’s newsletter is an allusion to a blog post from a few years ago, Google Maps’ Moat, by Justin O'Beirne.¹ He demonstrates, quite convincingly, that Google has a huge head start over Apple in terms of foundational data (e.g. roads, buildings, parks, etc.), especially in the continental United States. But in the years since, Justin has posted update after update showing Apple’s steady climb toward parity.

Even Google’s crown jewel, Street View, is under siege—a new initiative from Apple called “Look Around” is pumping out street-level panoramic imagery at a furious clip that’s becoming furiouser and furiouser by the week. Here’s an excerpt from Justin’s last dispatch:

Fear and Loathing in Mountain View

It’s hard to knock Big G. They’re still the biggest and best—a Morgan Stanley analyst estimated that Google Maps generated almost $3B in advertising revenue in 2019 alone. But I suspect we’re at the tail end of the golden era for Google Maps. They appear, to me, to be acting from a place of fear and conservatism rather than innovation.

Back in 2012, I had not yet been born (I am, in fact, only six years old). But I am a student of history. Apple Maps launched that year, and it was received so terribly that Tim Cook decided to write a personal apology for the state of the app. He even went so far as to recommend that iPhone users download Bing Maps while Apple sorted itself out…

In that moment, Google’s strategy of investing ungodly sums into Google Maps and keeping all of the data proprietary paid off in spectacular fashion. Apple had pie all over their face and Microsoft had tethered their consumer map product to the Bing brand, which was saddled with a comically inferior reputation in the search domain. Even if Bing Maps had somehow been comparable with Google Maps out of the gate, the average consumer would have associated the brand with being a knock-off of Google anyway. That year, and for many years afterward, Google looked unstoppable.

Then things started to take a turn around 2018, a year after Justin O'Beirne’s seminal summary of Google’s “moat.”

See, Google Maps is not just an app on your phone. It’s also a suite of developer tools that power countless other applications that are used by millions of people every day. And that part of the business is known as the Google Maps Platform (but most of the time I hear it referred to as the Google Maps API).

In 2018, Google inexplicably decided to self-sabotage their enterprise maps business by raising their prices ~1,400% overnight. The only time in my life that I’ve ever felt envious of commission-based sales people was in the wake of that announcement — I would have really liked to work at Mapbox as an order-taker that following quarter.

Still, today, you need an MBA with a specialization in Mind Games to understand the Google Maps Platform pricing scheme, and it helps to have a joint law degree to navigate their terms of service. In fact, the ToS are so Draconian, they’re the subject of investigations by the House Subcommittee on Antitrust, Commercial, and Administrative Law. My favorite tidbit from their recent report, Competition in Digital Markets quotes an anonymous Google Maps Platform customer:

Several developers stated that Google Maps introduced greater licensing restrictions as it gained a stronger market position. One noted that Google’s control over what now serves as a key mapping technology has allowed Google to call all the shots. “We license Google Maps and it’s essentially a contract of adhesion. It’s full of restrictions and we aren’t able to negotiate any changes,” the developer said.

These are the actions of an organization that is annoyed they ever let people become their customers. And for what? A few hundred million in fees, maybe? That’s likely nothing compared to the billions in advertising revenue from consumers searching for the best hamburger in town. So why not jack up the prices, lock down the data, and let ‘em churn, baby, churn?

The trouble is, Google isn’t the only game in town anymore. If they keep alienating their customers and pursuing a proprietary data strategy at all costs, they’re going to continue to lose ground to competition while spending more than ever just to tread water.

FAAMm is Coming

It’s not clear yet that mapping is big business, but it’s quite clear that big business is mapping. The most popular thing I’ve ever written is a piece about the incredible year OpenStreetMap is having, largely because of enormous investment from Facebook, Apple, Amazon, Microsoft, and Mapbox (FAAMm).²

Google’s slight advantage in terms of navigation and foundational features may well be overtaken by the collective effort of their deep-pocketed competitors. Google’s remaining edge in street level imagery, place data, and 3D data all seem weaker than at any moment in their history. Challengers are crowding in from all sides:

• With regard to street level imagery, I’m keeping an eye on what Facebook will do with Mapillary, the slew of well-funded HD mapping companies popping up (e.g. DeepMap, CARMERA), and Apple making a concerted effort with Look Around.

• Facebook’s Pages dataset seems like a natural rival to Google’s Places API if they ever decide to make a go of it. Yelp hates Google for many reasons (including an acquisition deal that fell through more than a decade ago) and would make for a suitable acquisition target for any of Google’s many competitors looking to jump-start an assault on their place data dominance.

• If Facebook ever did want to challenge Google with place data, they might consider partnering with someone like AWS. In fact, AWS just launched its own mapping service, Amazon Location, loaded with data from Esri and HERE.

• Meanwhile Mapbox’s recent launch of Mapbox GL JS v2.0 is taking direct aim at global 3D data, which Google won’t even deign to offer via their Maps Platform.

• And let’s not forget about Microsoft getting serious about their “Planetary Computer” and affiliated AI for Earth initiative, which are gearing up to be the stiffest competition Google Earth Engine has faced to date.

• Meanwhile at an office down the hall, Azure Maps has been humming along quietly since 2018, beating AWS to the punch by over two years. And Microsoft’s consumer app, Bing Maps, has been investing heavily in automated foundational feature creation, having not only openly licensed over 100M AI-generated building footprints but *also* the model and toolkit used to create them.

How Google Will Regress to the Mean if Nothing Changes

One of my favorite blog posts of all time is Steve Yegge’s missive, “Why I left Google to join Grab.” In it, he claims the primary reason he is leaving is because:

First, they’re conservative: They are so focused on protecting what they’ve got, that they fear risk-taking and real innovation. Gatekeeping and risk aversion at Google are the norm rather than the exception.

OpenStreetMap (OSM), over the next decade, has the potential to do to Google what Android did to Apple: dramatically grow the overall market while drawing a clear line between the larger “open” ecosystem and the smaller “proprietary” one. The difference is that while Google Maps probably makes billions annually, it also probably costs billions to maintain, leaving it wallowing in a low-margin no-man’s land compared to its big brother Google Search.

What gives Google Maps an edge over other experiences? Today, I would argue the three pillars of its comparative advantage are Places, Street View, and 3D data. But in a few years, I think it will mostly just be Places. And shortly thereafter, it may well wind up with no remaining edge beyond convenience and brand loyalty.

I don’t believe Google can maintain its leading position in the software mapping world for another decade without pivoting to embrace the same open ecosystem that FAAMm is using to challenge them. Imagine, as my friend Sean Gorman recently wrote, a world in which the 3D globe was a public good (or, by extension, if all foundational features and places of interest were treated that way). What would that mean for the competitive landscape? ³

Rather than a question of who has the resources to generate the best private datasets and protect their proprietary nature, the battleground of competitive advantage would shift to user experience, convenience, and creativity. And, in my opinion, Google has an edge there over FAAMm—the more Google Maps can tie into GSuite and the rest of their prodigious ecosystem of popular applications, the more they can differentiate on superior user experience and less they have to rely on superior data.

If Google doesn’t start taking the Google Maps Platform seriously, they’ll slowly but surely find themselves alone on an island of inferior, less frequently updated, and expensive-to-maintain proprietary data. A new generation of innovative apps built on top of OSM will feast like piranhas on a cow treading water.

In my opinion, there are three major strategic decisions that could position them to maintain their leadership position into the next decade:

1. Ride the OSM wave instead of fighting it. Foundational feature data is becoming commoditized, but Google could make it collectivized. By combining efforts with OSM more broadly, Google could reduce its cost basis moving forward and free up resources for competing at what they’re really great at: building great software.

2. Loosen up licensing restrictions on the Maps Platform generally. Allow customers to digitize features visible in Street View imagery, for instance. Grant permission to users who want to mix-and-match your geocoding service with a beautifully-styled basemap from Mapbox. You’re stifling countless creative uses for your services while fighting against an irresistible force that the core of your search business is built upon: information wants to be free.

3. Aggressively price services to grow the overall market. Google Cloud Platform (GCP) is likely an outsized share of the future of your business—Microsoft and Amazon clearly understand this about themselves. But when it comes to maps, you have the lead. By mis-pricing your maps API, you’re taking a cracked door and slamming it wide open for your competitors. And for what? Do you really need the incremental revenue today?

¹ Who is Justin O’Beirne? I am confident he worked at Apple based on publicly available patent applications in his name, but I don’t think he still works there or else they’d probably be pissed at him for writing publicly about their work (they are notoriously secretive). He keeps an incredibly low profile—no LinkedIn profile that I can find, no Github profile, and perhaps most confusingly, a Twitter profile that was deleted earlier this month after I mentioned it publicly (but which had no tweets, a blank profile picture, and thousands of followers). I love a good mystery!

² Sorry, Mapbox, but you don’t get a capital letter. Sure, you’re a billion-dollar juggernaut and a massive company by any normal measure, but your conspirators are each valued at over one trillion dollars.

³ The recently launched Placekey initiative deserves a shoutout here. I still don’t understand it (sorry, Auren!), but it provides at least some evidence that there is widespread appetite for more openness and shared standards around common geospatial datasets.

Five decades ago, Jack Dangermond and his wife Laura started a boutique consultancy with a rather grandiose name: Environmental Systems Research Institute. Over time, “E-S-R-I” has simplified to just “Esri.” I’ve always thought it was a beautiful sounding name, like a forgotten Egyptian god or the name you might get stuck with if you were born on a hippy commune.

Speaking of names — “Jack Dangermond” sounds like an assumed alias. It’s just a little too perfect. While researching this piece, I was surprised to learn that he came by it honestly. His parents Alice and Peter Dangermond immigrated to the United States from the Netherlands and settled in Redlands, California where they started a tree nursery in 1945 (the same year Jack was born).

Esri is a “lifestyle business” founded by a husband-and-wife team, headquartered in the same sleepy town where Jack grew up, and run profitably without the aid of venture capital or debt financing. Today, Esri is what happens when a lifestyle business averages 30% annual growth for 50 years in a row. One thing leads to another and — whoops — you are now a billionaire. Not a bad… lifestyle.

What is Esri?

Esri is a software company that helps governments, companies, research institutions, and non-profits make maps.

In fact, Esri is one of the largest privately held software companies in the world. It sells a vast array of software for handling geospatial data collectively referred to as a “Geographic Information System” (GIS). In recent years, they’ve taken to describing themselves as working on “location intelligence” and “spatial analytics,” but I suspect that’s an attempt to muddy the water for new challengers like Tableau and CARTO more than a sincere effort to describe what they do. They didn’t invent the term “GIS,” but they are nearly synonymous with it today.

Esri’s history officially begins in 1969, but the current software-focused incarnation of the firm began taking shape in the mid-1970’s when they first took on digitally-focused projects. Their transition from a consulting firm to a software firm wasn’t formally cemented until 1981, the year they released ARC/INFO. 

The first version of ARC/INFO was written in FORTRAN and was only accessible from the command line. Esri targeted government and academic institutions as early adopters of the software, and by 1983 had managed to win contracts with the University of Maryland and the Illinois Department of Natural Resources among a dozen or so other clients. 

Arc/Info (as it was later styled) crossed 100 “seats” in 1986. By 1997, that number had skyrocketed to over 100,000.¹

The firm is probably most famous for its ArcGIS franchise, the generational successor to Arc/Info. Since then, the core desktop licensing business has remained largely undisturbed except for the introduction of ArcGIS Online in 2012, Esri’s first “cloud” offering. Today, they offer so many different products, it’s hard to keep track of all them.

Esri is a formidable business. Their official stats page lists some eye-popping numbers including: ²

• 51 years in business

• 350,000+ customers

• 4,000 employees

• 49 offices worldwide

Patience, Rewarded

No debt. No private equity. No strings attached.

In a financial climate where technology companies are routinely raising hundreds of millions of dollars, sometimes before even launching their first product, Esri stands entirely apart. 

Below, I’ll get into some of the interesting strategic choices Jack and Laura have made while building the business, but I think the most important thing to understand about Esri is its capital structure. The Dangermonds retain unfettered control of the business and have surrounded themselves with a leadership organization of die-hard lifers. They’ve chosen to reinvest 30% of annual revenue (not profit, revenue) into R&D year, over year, over year. That’s a stupefyingly large bet on themselves that probably wouldn’t be possible if they were focused on returning cash to investors or catering to the quarterly attention span of the public markets.

Entrepreneurs of my generation are inundated with examples of technology companies that took a different path. The narrative that venture capital is necessary to build a business of meaningful scale is disturbingly dominant among my peers who have started a business or are considering it. 

“Consulting” has become a dirty word among ambitious young people I know, and “services businesses” are routinely scoffed at for not being “scalable.” Never mind that tens of trillions of dollars of American wealth is tied up in such businesses run by retirement-age executives desperately grasping at a succession plan.

I’ve even heard people describe themselves as having “bootstrapped” their business because they only raised a few hundred thousand dollars from “friends and family.”³ That’s part of my motivation for writing about Esri — I want people to know that they’ve been presented with a false choice between building a big business that has real impact in the world and a self-financed one they can retain complete control over.

I’m not saying Esri-like outcomes are even remotely likely for self-financed companies. I’m simply pointing out that they’re possible, which I think a lot of people just starting out in business genuinely don’t realize.

How The Beast was Built

The Dangermonds made several strategic choices while building Esri that I feel are instructive.⁴ Think of these as a Cliff Notes™ version of their story — most of the truly fascinating stuff is in the details, but unfortunately, I don’t know the details. 

Below are some of the priorities Jack and Laura set forth for the business that helped it quietly become a multi-billion dollar behemoth:

Become a Category King

I mentioned earlier that Esri is nearly synonymous with their software “category,” GIS. The term “category king” comes from one of the only marketing books I’ve ever purchased, Play Bigger. The basic thesis is: enduring companies market the value of their category, not themselves, and in doing so become synonymous with it. 

In Esri’s case, they promote the value and utility of GIS as a concept, to great effect. In fact, if you Google, “history of GIS,” the top hit is a page on Esri’s very own landing page. It’s not every day a company website outranks Wikipedia.

Focus on Overlooked Markets

Esri is now famous for what I’m gonna start calling pedagogical capture. I noted earlier that universities were a large portion of their initial client base with ARC/INFO, and they still are. 

Now “GIS” is its own field of study. You can get an undergraduate or even a Master’s degree in GIS. Esri is the overwhelmingly dominant software of choice for professors in the field. And for a surprising proportion of recent graduates of GIS programs, the only way they know how to express the analytical techniques imparted upon them in school is by working with Esri tools. Like I said, pedagogical capture.

The pejorative industry term for recently graduated Esri-reliant GIS analysts is “button pushers.” They’re becoming rarer by the day, but there are still a few remaining pockets of the world where you can get a job working in Esri software all day long (without knowing how to code or even write basic SQL queries). They’re mostly concentrated in the public sector, the other strategic market Esri focused on early and consistently.

Speaking of the public sector, pour yourself a stiff drink and Google this exact phrase:

Esri Enterprise License Agreement filetype:pdf

While other people see city, county, and state governments as slow, fractured, services-intensive, unprofitable cottage industries where great software companies go to settle for mediocrity…Esri sees dollar signs.

Imagine this common scenario: an undergraduate with a GIS degree can’t find a job, so they join a Master’s program in GIS (decades later, they will find themselves prone to random bouts of crying). Soon thereafter, they pop out the other side with $100K in debt, and by the grace of God, they manage to land a job at a local county GIS office that pays $40K/year. A few years later, Esri’s license is due for renewal — a whopping 5-year commitment of $1.5M. Esri tosses in free tickets to their annual user conference for the entire office…and a deal gets done, no questions asked.

Which brings us to our next strategic priority.

Build a Community

Esri was one of the first software companies to perfect the“user conference.” Today, over 16,000 people attend their annual event. Lanyards abound.

But perhaps the most striking thing about “Esri UC” is the exhibition hall. Esri has built an enormous and thriving partner network. By focusing on growing the industry as much as growing the business, Esri has put themselves at the center of a miniature economy revolving entirely around their software.

Expand by Franchising

One of the most controversial decisions Esri ever made was to aggressively partner for distribution outside of the United States. Personally, I can empathize with the decision — the difficulty of selling overseas, especially to foreign governments, is almost universally underestimated by young companies. Having international distributors allows Esri to maintain its focus on product development and domestic sales (where they have the most competency) while growing and diversifying their revenue with minimal cash investment. I’m not sure if “franchising” is technically the right term, but it captures the spirit of the thing.

The reason the decision is controversial is because resellers of Esri software abroad have…strange incentive structures. They don’t have the overhead cost associated with developing the product, and they typically have exclusive rights to sell the software within their territory. Add those two things up, and you get a strong incentive to sell software for the highest amount possible per deal (because the logistics of servicing each deal are where most of your costs come from). The result is a distorted market, where Esri products routinely cost more in low-income countries than they do in America. That’s messed up.

None-the-less, the model seems to work for them. Esri distributors are as far-flung as Afghanistan, Argentina, Iceland, and New Zealand. On their “Quick facts” page, Esri claims, “most national governments” use their software — a feat that I doubt would be possible without the global footprint their distributors provide.

Embrace Professional Services

As someone who has competed, on occasion, with Esri sales reps, I know their playbook a little. To land new accounts, they prefer not to discount the software so much as throw in free licenses of complementary products, especially if it’s a large non-profit. Sometimes, if they’re really desperate, they’ll even throw in a basket of professional services time to help customize their tools to the customer’s liking or integrate them with internal tools.

A year or two later, once they’re settled in, the bill for an annual license for those freebies will arrive in the mail. I believe in the startup world they call this strategy “land and expand.”

These days, professional services is looked down upon as a low margin source of revenue that only scales linearly. But what Esri realized early on is that in “Enterprise” software, you almost can’t over-invest in landing the initial deal and ensuring it goes well. Professional services, for them, is getting paid to lock a customer in even more deeply. And that’s how you build a huge business — sure, you add new customers, but most of all, you retain your install base and grow revenue per customer over time.

There are Worse Tyrants

Esri often gets a bad rap in the GIS world because it has so few natural competitors. It looms large over a tiny backwater of an industry that it practically created from scratch. 

And it deserves a lot of the criticism it gets — if you’ve ever gone head-to-head with Esri for a deal, you know how fiercely (and unfairly) they’re willing to fight. They also seem to have a habit of bad-mouthing open standards that compete with their own products (until they inevitably decide to support them and pretend they did all along) among other annoying traits.⁵

But I can’t think of a more benevolent pseudo-monopoly in software. Esri’s commitment to supporting environmentalism and civic health is genuine. If they yanked all of their desktop licenses overnight, the world would shut down. Literally. It’s not like their business model involves farming personal data for advertisers — they’re building mission-critical software for clients doing meaningful work.

Most importantly, they’re doing it all on their own terms. They aren’t at risk of a hostile takeover from an activist investor. They don’t have any loans to default on. Their client base is composed of the slowest moving and most loyal organizations in the world. When you add it all up, I think it’s safe to say: Esri can’t be stopped.

¹ Source: https://gisuser.com/2007/06/esri-arcinfo-arcgis-arcview-25-years-in-the-making-a-time-line/. As an aside, a good “dad joke” for all you software people out there: next time someone quotes you in terms of “seats,” ask them if you can get a discount because so many of your coworkers use standing desks. It won’t go over well, but it’ll be satisfying — I guarantee it.

² Source: https://www.esri.com/en-us/about/media-relations/fact-sheet

³ That term — “friends and family” — has never sat right with me. You can’t control who your family is, and unless you grew up in an affluent area, there’s not much chance you have friends who are accredited investors…For all but the most fortunate people, raising a few hundred thousand dollars requires going far, far outside of their close (or even distant) circle of friends and family.

⁴ I refer to them as a team, because while Jack is undoubtedly the “face” of Esri, Laura has always been instrumental in running the business despite her low profile.

⁵ Cloud Optimized Geotiff vs. Meta Raster Format comes to mind as a recent example for the nerds out there.

Maybe someone forwarded you this email. First, be sure to write them a hand-written letter thanking them for being such an incredible friend (and slip in a $20 for good measure). Second, please consider signing up to receive more posts like this each month from A Closer Look:

Subscribe

The news yesterday that the new version of Mapbox GL JS will be proprietary shook me.¹

I am not a zealot. I do not feel entitled to others’ intellectual property, even if they’ve given away their ideas in the past. I know from experience how exhausting, thankless, and exploitative the work of creating and maintaining open source software can feel.

Until yesterday, I was still clinging to a few shreds of romantic optimism about open source software businesses. Mapbox is the protagonist of a story I’ve told myself and others countless times. It’s a seductive tale about the incredible, counterintuitive concept of the “open core” business model for software companies. In a piece I wrote last year that no one read, I defined the open core strategy thusly:²

“Open core” businesses offer a free, open source version of their software and a paid version with additional proprietary features that would be a pain to replicate […]

For the best in-depth analysis of the concept of the “open core” business model I’ve come across, I highly recommend reading Joseph Jacks’ blog post Open Core — Definition, Examples & Tradeoffs.

The whole idea is insane. No one believes it could possibly work when they first learn about it, and yet dozens of companies like Elastic, D2iQ (formerly Mesosphere), MongoDB and Cloudera have all managed to achieve valuations in the billions of dollars by pursuing this batshit-crazy, reverse-psychology, let-it-all-hang-out strategy. Or, at least, they were open core businesses at some point…maybe not so much today. More on that later.

Today, we’re gathered here on the internet to mourn the death of the open core business model. We’re here to tell stories of the before-times, to reminisce about how smart we thought we were. We went against consensus, and we were wrong. Because, open core is dead.

Cloud killed open core.

Back to Mapbox

In the case of Mapbox GL JS, Mapbox had previously decided to openly license the first two versions of the their browser-based map renderer (the same one that powers Snap Maps, the New York Times, and CNN among myriad others). Ever since the initial release in 2014, it’s been incredibly popular among web developers. Once you know what you’re looking for, you start to see it…everywhere.

I personally experienced the power of Mapbox GL JS when my team at Azavea started building GroundWork, a tool for labeling satellite imagery. Using functionality from that library, GroundWork supports freeform drawings of complex geometries. The resulting shapes are cartographic — they’re projected onto a real location on earth, not just suspended in imaginary 2D space. It’s the type of feature that feels obvious and straightforward but is in fact extraordinarily difficult to engineer from scratch.

Even simple-looking shapes drawn with a freehand technique can contain thousands of individual vertices. Pretty quickly, you fill up your screen with hundreds of thousands of vertices worth of shapes and…oh, poop-and-a-half.³ Your browser crashes.

Mapbox GL JS helps to circumvent that problem by summoning help from the graphics card on your machine. There’s no way we could have built that feature within our budget and time constraints without piggybacking on Mapbox’s tens of thousands of hours of hard, low-level engineering work.

Mapbox’s decision this week to release a brand new version of Mapbox GL JS and keep it proprietary stunned me. Not only was v1 already a wildly popular open source library, but Mapbox’s reputation as a prolific creator of open source software is a cornerstone of their whole identity. Describing what makes Mapbox special without mentioning “open source” is like trying to describe chocolate milk to an alien without using the word “liquid.”

Mapbox still has over 800 open source projects to their name and has consistently ranked in the top 40 organizations worldwide in terms of Github activity on public repositories. They employ (and have previously employed) some of the highest-profile open source software contributors in the world, not just the mapping industry.

So…what the dickens is going on here?

Open Core is No Longer a Tenable Business Model

Mapbox’s choice to keep v2 of Mapbox GL JS proprietary is a strong signal. But what precisely the signal is indicating is…less clear. Nonetheless, I’ll bite.⁴

I think it means we’ve reached the end of an era. I think it portends the death of “open core” as a viable business model. My suspicion didn’t start with Mapbox, but it has concluded with them.

To understand how a company like Mapbox decides to release a proprietary update to one of its most popular open source libraries, I think it helps to start by studying at the cautionary tales of two kindred spirits, Mongo DB and Redis Labs.

Ben Thompson wrote an excellent summary in his piece AWS, MongoDB, and the Economic Realities of Open Source almost two years ago. To summarize his summary: if you give your secret sauce away for free, and it gets popular enough, cloud providers will inevitably spin up competitive services using your very own code against you. They will ruthlessly, unapologetically, shamelessly bludgeon you with a rubber chicken of your own fashioning. They’ll take a dump in your front yard while your lawyer stands over your shoulder whispering, “nothing can be done.”

Since Ben wrote about AWS launching a service directly competing with MongoDB and Redis’s paid offerings, AWS hasn’t exactly crushed the competition…in fact, MongoDB’s stock price has risen over 275% since then, and Redis just raised $100M to officially cross the magic $1B valuation threshold a couple months ago.

The bigger story is not simply that Mongo and Redis thrived in response to getting shanked by AWS, but how they did it. Both fought back the way companies always fight back: with an army of IP lawyers. Redis adopted a strategy of adding a severely restrictive “commons clause” to updated versions of their existing open source tools, prompting some prominent open source commentators to feel that they’d thrown the baby out with the bathwater:

— @webmink

Redis later shifted to an entirely novel and idiosyncratic license, which can’t be worse than the commons clause fiasco, but almost certainly isn’t much better. MongoDB also took the “poison pill” legal strategy by shifting to a new and equally made-up license for their software, “Server-Side Public License (SSPL)” shortly after AWS launched its competitive “DocumentDB” service. In both cases, the moves were specifically aimed at making it harder for public cloud providers to screw them over.⁵

Perhaps more accurately, they made their open source offerings radioactive to any company large enough to have a department called “compliance.”⁶

These changes made Redis and Mongo fundamentally less open. The companies have since done quite well, although perhaps at the expense of their original ideals and some of the goodwill of their early adopters. The open source community of evangelists and contributors felt that they’d been sold a lie and treated as a stepping stone in the process. And I tend to agree with them, although I don’t fault either Redis or Mongo for being rational economic actors.

Back to Mapbox, Again

There is at least one cloud provider that has publicly copy-and-pasted Mapbox code into services they charge for: Azure, Microsoft’s cloud service. Last year, Azure announced map styling powered by Mapbox GL JS and it remains a key feature of their “Azure Maps” service. Mapbox even wrote an announcement about it on their company blog.

While Mapbox made the Azure announcement out to be a good thing, I suspect it was the beginning of the end for Mapbox GL JS remaining open source. After all —in the incomprehensibly competitive world of public cloud computing, once one cloud provider starts offering a service, the rest are surely soon to follow. Mapbox found themselves in a similar position to Mongo and Redis: they were subsidizing R&D for a handful of trillion-dollar tech giants.

Unlike Mongo and Redis, Mapbox resisted the urge to try to have their cake and eat it, too. Rather than change the fundamental license for the “open” part of their offering, they made a clean break. The old version, which is still a mature and extremely useful library, will retain its original, permissive license. Efforts are already underway to fork it and gather community members to maintain it indefinitely, which I’m hopeful will be successful.

The new version will remain public (the code is all published on Github here). It’s just not permissively licensed anymore. To me, this feels like a more honest approach than trying to thread the needle with a creative and totally unproven license or some “gotcha” clause.⁷

Some might feel that this is a tragedy because it means community contributions will probably fall off. It’s true that, despite a long list of contributors, current and former Mapbox employees contributed the lion’s share of investment in the library. But the project had attracted a massive, global community of engineers who built things with it, gave tech talks on it, and forked/remixed it countless times. Yesterday was a sad day for those enthusiastic users, there is no question. They’ll survive, but I can’t blame them for feeling a sense of loss.

As for the question of whether or not this choice strays from Mapbox’s mission or culture, I think this tweet from a current employee who has been with the company since its founding about sums it up:

— @samanbb

Remembering the Good Times

Once upon a time, I really thought you could give away your trade secrets and still be successful. I thought the scale of the internet had enabled a new genre of company that could become massive despite only capturing an infinitesimally small fraction of the value they created. I believed the act of building a company around open source software was virtuous and ethical. I saw it as an end in itself.

These days, I’m not so sure.⁸

I still believe open source software is a powerful force for good in the world. I still think companies can strategically and earnestly contribute to open source software in service of not only their own mission but also the collective good.

But, I no longer believe venture-backed companies can responsibly pursue a strategy of giving away the software at the core of their value proposition. I no longer think it’s a feasible model for companies with ambitions of becoming very large or those actively avoiding consulting work. Eventually, if they’re successful, they will be forced to choose between betraying their loyal early adopters and dying a long, slow death by rubber chicken bludgeoning.

Yesterday was a sad day, indeed. But not because Mapbox did something wrong or disappointing. Because yesterday, I finally conceded to myself:

Cloud killed open core.

¹ For newcomers: Mapbox is the preeminent software company focused on mapping. They produce an outsized share of innovation in the field — from Mapbox GL JS (their renderer for 2D maps), to autonomous driving and navigation libraries, to augmented reality, 3D visualization, and even video game tech. The reason this news is shocking is that their commitment to creating open source software, even the “secret sauce” at the very core of their work, is fundamental to the way many of their users perceive them and a huge part of why they are so commonly admired by software engineers.

² If you’ve already read that blog post, congratulations. You are one of the 106 people that I consider to be my closest friends.

³ My mother liked to use this expression when I was growing up, and I am still quite fond of it.

⁴ If you’re reading this blog, you’re desperate enough for answers about the geospatial software industry that you’ve turned to a guy that writes on Medium because he’s too technically illiterate to host his own website and too cheap to pay someone else to host it for him.

⁵ I am not a lawyer, but I play one on the internet. If you want to understand these licensing strategies in more depth, I suggest you start by getting a law degree, specializing in intellectual property and copyright, and then spending your entire life studying it.

⁶ A very helpful commenter has pointed out that I have heavily implied Redis itself is no longer open source, when in fact it remains open under the BSD license. Just the “modules” that Redis Labs makes carry proprietary “source available” licenses.

⁷ If you haven’t caught on by now, let me be clear: I think inventing your own license is almost always a bad idea. Contracts are enforceable in court. Either you plan to go to court a lot until the credibility of your license is well established, or you should pick one that is already credible. Personally, I don’t like to get sued. And if I must get sued, I want to feel confident I’m gonna win.

⁸ I’m reminded of the lyrics to one of my favorite Joe Pug songs:

Stealin’ was so easy then

I wish that it still were

Now as I pick my own pocket

I know that these days

I’m not so sure

The first time I spoke with Jennings Anderson, I couldn’t believe what he was telling me. I mean that genuinely — I did not believe him. He was a little incredulous about it himself. I felt like he was sharing an important secret with me that the world didn’t yet know.

The open secret Jennings filled me in on is that OpenStreetMap (OSM) is now at the center of an unholy alliance of the world’s largest and wealthiest technology companies. The most valuable companies in the world are treating OSM as critical infrastructure for some of the most-used software ever written.

The four companies in the inner circle— Facebook, Apple, Amazon, and Microsoft— have a combined market capitalization of over six trillion dollars.¹ In almost every other setting, they are mortal enemies fighting expensive digital wars of attrition. Yet they now find themselves eagerly investing in and collaborating on OSM at an unprecedented scale (more on the scale later).

What likely started as a conversation in a British pub between grad students in 2004 has spiraled out of control into an invaluable, strategic, voluntarily-maintained data asset the wealthiest companies in the world can’t afford to replicate.

For the Uninitiated: What is OpenStreetMap?

I will admit that I used to think of OSM as little more than a virtuous hobby for over-educated Europeans living abroad — a cutesy internet collectivist experiment somewhere on the spectrum between eBird and Linux. It’s most commonly summarized with a variant of this analogy:

OSM is to an atlas as Wikipedia is to an encyclopedia.

OSM acolytes hate this comparison in the much same way baseball players resent when people describe the sport as “cricket for fat people.” While vaguely truthful, it doesn’t quite get to the spirit of the thing.

OSM is incomparable. Over 1.5M individuals have contributed data to it. It averages 4.5M changes per day. The stats page on the OSM Wiki is a collection of hockey sticks that look like this:

You can think of OSM in several ways:

• A distributed community of mappers contributing information about the geography of the world to a common repository

• A free web map hosted at https://www.openstreetmap.org/

• A loosely affiliated collection of free and open source tools for mapping the world

• A real-time stream of instructions representing how to add, change, or remove cartographically projected geometries and associated metadata based on a prior state

• …Google Maps, but openly licensed

It’s hard to get people to agree on what exactly OSM is, but almost everyone agrees on one thing: it’s extraordinarily valuable and important.

What Jennings Told Me about OSM

For those paying attention, none of what I outline below will be news. However — outside of a relatively small cluster of weirdos who pay attention to trends in geospatial technology— almost no one seems to be paying attention.

That’s mostly because so few people have even heard of OpenStreetMap, despite the fact that hundreds of millions of people rely on it during any given month. If you’ve ever opened Snap Maps or Apple Maps or Bing Maps or even just peeked at the dash of your obnoxious neighbor’s new Tesla…you’ve used OSM.³

In May of 2019, Jennings co-authored a paper with Dipto Sarkar and Leysia Palen titled, Corporate Editors in the Evolving Landscape of OpenStreetMap. If you prefer the research in presentation form, this talk is a fabulous summary of their findings:

In that talk, Jennings outlines the findings presented in his research. Not only was there already significant corporate investment happening in OSM in 2018, but in many cases corporate editors were responsible for the majority of edits in the specific geographies they were focused on. For instance:

For areas where corporate teams are active, on average, the non-corporate editors are now responsible for less than 25% of total road editing activity, which is down from closer to 70% in 2017.²

Jennings noted, importantly, that as of 2018 non-corporate editors were still responsible for the majority of activity on OSM (about 70% of all edits) and were significantly more active on edits to buildings, places of interest, and amenities.⁴

In a more recent talk from State of the Map in July 2020, Jennings presented updated figures showing that the torrent of corporate contributions only increased from 2018 to 2019 and beyond with Amazon and Apple trending along the steepest slopes:

Also interesting to note is Mapbox’s apparent decision to stop investing significantly into direct OSM edits and contributions. Apple was responsible for more edits in 2019 than Mapbox accounted for in its entire corporate history…I don’t have a good explanation for that. I wonder if they decided their effort could be more highly leveraged on core web mapping technology rather than manual digitization.

The Clash of Cultures Happening in OSM

I’m in no position to comment on most of the things I write about. But in this instance, I’m particularly unqualified — OSM has amassed a long-lived, fantastically diverse, and inherently fragmented community. I’ve never even commented in one of the forums.

But one thing that is clear even to a casual observer like me: one of the consequences of increased corporate involvement in OSM is a significant backlash from members within the OSM community that feel the community (and data) is being irreversibly adulterated by these profiteering intruders.

At the last OSM annual conference Frederik Ramm, a prominent and quite thoughtful OSM community member, summarized the attitude toward corporate contributors this way:

“[…] none of these companies is essential to OpenStreetMap. They are contributors, but OpenStreetMap could work perfectly well without them […] the mainstay of OpenStreetMap is the millions of hobbyists, individuals that contribute to OpenStreetMap.

A vocal minority of voluntary contributors to OSM seem to have a bit of a chip on their shoulder when it comes to the suits. A consistent undercurrent that I’ve noticed is skepticism about the motivations and incentives of for-profit firms. Here’s a typical sentiment excerpted from Serge Wroclawski’s magnificently controversial blog post, Why OpenStreetMap is in Serious Trouble (published in February of 2018).

Many of the founders of the project, as well as others, have launched commercial services around OSM. Unfortunately, this creates an incentive to keep the project small and limited in scope to map up the gap with commercial services which they can sell.

I think the playing field has changed significantly since Serge wrote those words — he was likely referring to projects like CloudMade (now defunct) and Mapbox ,which sought to offer generic map services on top of OSM’s dynamic map database (rather than enhance in-house products where mapping is ancillary to their core value proposition like it is for FAAM). He makes an interesting argument that OSM itself should be offering these services rather than letting companies piggyback on the efforts of countless volunteers while capturing all of the economic value.

What’s Motivating These Companies?

I wrote earlier this year about the concept of “Commoditizing Your Complement,” in my explanation of why Facebook acquired Mapillary and then gave away all the data they had just purchased for free.

The concept is simple: undermine your competitors’ intellectual property advantage by collaborating with aligned entities to cheapen it with a free and openly licensed alternative.

I would wager that corporate participation in OSM is less about directly monetizing souped-up versions of OSM data provided as modern web services and more about desperately avoiding the existential conflict of having to pay Google for the privilege of accessing their proprietary map data.⁵

Whatever the motivations of these mega-corporations, they’ve succeeded in carving out a niche for themselves within the OSM community whether the hobbyists like it or not. I’d like to highlight a nuance often lost in this discussion — just exactly who are these companies hiring to add data to the map? They are often already-active, enthusiastic contributors to OSM. These are people living the open data fanatic’s dream: getting paid to do a job they find so fulfilling they would otherwise do it for free in their spare time.

There’s obviously a lot more to it than just sticking it to Google. Facebook, for instance, has ambitions of building new types of digital experiences that interplay with the real world (as evidenced by their focus on augmented reality and acquisition of novel user interface technology like CTRL-labs). Apple has added LiDAR to its new line of iPhones and iPods allowing customers to scan the 3D world in high fidelity among other exciting uses:

— @Rengle820

These firms have outgrown your office and your living room. They want to be with you literally every where you go, and constantly seduce you with entertaining and immersive experiences. The more of your attention they can monopolize, the more money they can make from selling chunks of it to advertisers and people developing software on their platforms.

Whether you like their motivations or not, the result is a desire to map the world in higher fidelity and at larger scale than even they can afford to accomplish independently. And that has, for better or worse, brought their interests into alignment with the grassroots OSM community.

Why Does it Matter?

Well, anytime the wealthiest institutions in history are quietly collaborating on something, I think it’s worth noting. I’m not sure there is a precedent for such a collaboration — if you know of a case where otherwise embittered mega-corporations worked with a global community of volunteers on a public dataset…let me know. I’d love to learn about it.

The question on my mind is how idiosyncratic this situation really is. Does OSM represent a model for strategic corporate sponsorship of public goods moving forward? Or is it tragically inimitable?

For instance: I work for a company called Azavea that, among many noble efforts, maintains Cicero. It’s a database of elected officials and legislative districts in several countries around the world that gets updated daily. You can imagine that this should be a public good — like, doesn’t the government already have this information? Turns out…nah. Cicero requires ceaseless, grueling work to keep updated, and that means serious investment of time and money.

One of the key differences between Cicero and OSM is a community of contributors. Community is what makes OSM special. Without it, the project is “default dead,” as they say in Silicon Valley. Much like elected official information, map data goes stale fairly quickly and therefore requires constant life support.

OSM’s community seems conflicted about whether or not corporate participation is ok (let alone good) for the future of the project. And yet the community is precisely what attracts corporate contributors. OSM provides two advantages over just buying privately collected data:

• Existing data is free and growing apace

• Proprietary data contributed to OSM may be expanded upon and/or maintained at no additional expense by the community

Some may squirm at the idea that their contributions to OSM help FAAM…after all, do they really need the help? But what’s beautiful is that FAAM is contributing (rather than passively mooching) because of the compounding value of having any/all data make it into the community’s growing number of hands.

I’m kind of shocked to be saying it, but somehow — almost inexplicably — the goals of the OSM community and corporate contributors seem to be largely aligned. They all want an accurate, ubiquitous map of the world that can be maintained in perpetuity as sustainably as possible.

It’s the opposite of the Tragedy of the Commons — all of the private property holders, acting in their own self interest, are enriching the common resource rather than depleting it.

¹ True as of November, 2020.

² https://youtu.be/gAIoDFVwMJA?t=442

³ As an astute commenter on Hacker News pointed out, Tesla uses Google Maps for its dashboard. They only use OpenStreetMap for the auto-summon feature, which incidentally has resulted in lots of parking lots being added to OSM by the Teslarati as covered in Jennings’ most recent SOTMUS talk!

⁴ The 2019 paper is full of amazing tidbits like this: ~1.4% of contributors account for ~90% of the edits to OSM. This roughly adheres to something called the 1% rule of online communities which states that, “1% of Internet users are responsible for creating content, while 99% are merely consumers of that content.”

⁵ …but it’s definitely also about enriching their own applications and cloud services.

A Word of Warning

Anything worth doing will be publicly criticized. In the words of Taylor Swift, the poetic conscience of my generation,“Haters gonna hate.”

Today, I’m the hater.

That said, I deeply admire the entrepreneurs starting satellite imagery platforms and marketplaces. I share their passion for fixing an industry that is obviously, fundamentally broken. If they succeed, we’ll all be better off for it.

But…I do not think they will succeed. I hope I’m wrong.

Alright, enough hedging.

The Commercial Earth Observation Ecosystem

Like any good Thought Leader™, I have my own little taxonomy of earth observation startups:

• Suppliers (e.g. Airbus, Maxar, Planet, Satellogic)

• Marketplaces (e.g. Arlula, Shadowbreak International, Skywatch, UP42)

• Platforms (e.g. Astreaea, Descartes Labs, Orbital Insights, SpaceKnow)

• Consultants (e.g. Azavea, Development Seed, Element 84, Sparkgeo)

• Solutions (e.g. Arturo, Cloud to Street, Farmer’s Edge, SilviaTerra)

In my (inflated) opinion, only the first and last categories present venture-scale opportunities. Consultants are also a great category of business, but they simply aren’t conducive to venture capital; they tend to be small, profitable businesses that sell for modest multiples of annual profit (as opposed to outsized multiples of forward-looking revenue). Don’t get me wrong; I like consultants. I work for one!

Marketplaces and platforms, on the other hand, seem like venture-scale businesses. And they’ve most certainly soaked up lots of speculative money— on the order of hundreds of millions of dollars in equity financing over the last five years.

Yet, to survive, they almost always become consultants (although you’ll never catch them admitting to it). Fundamentally, I don’t think they work. Allow me to explain.

Defining Marketplaces and Platforms

Marketplaces are businesses that go around to as many imagery suppliers as possible and sign reseller agreements in order to distribute a bunch of disparate sources of imagery through a single web storefront.

Platforms, on the other hand, don’t typically emphasize transactional imagery sales in their value proposition (like marketplaces do). Rather, they focus on simplifying the challenge of imagery analysis either through a suite of hosted analytical tools or through their own proprietary information feeds (or both).

Both share a lot in common, including the hallmark characteristic of aggregating various sources of imagery behind a single API.¹

Why They’re Such Seductive Businesses to Start

Commercial earth observation is a complex, international, fragmented, highly-regulated, opaque industry. It’s ripe for disruption (as a fervent capitalist might put it).

I liken earth observation to the telecommunications industry before Twilio. Or the online payments industry before Stripe. Or the mid-19th century banjo parts industry before eBay.²

Surely, satellite imagery can’t be more complicated than legacy banking infrastructure or international telephony? Distribution of satellite imagery will universally catch up to other digital goods…at some point. It’s a trend so obvious that it feels inevitable.

The thought might arise in your mind…someone will eventually make the eBay (marketplace) or the Stripe (platform) of earth observation.

And when they do, it will be an enormous business that cannibalizes all of its competitors as network effects kick in. As more users flood to the app, negotiating power with the imagery companies will increase, which in turn will push prices further down, which in turn will attract even more users, which in turn will push prices further down, and on, and on.

As this theoretical company grows, new imagery suppliers will clamor to integrate their data rather than build their own software infrastructure at great cost. A massive untapped market of semi-technical and non-technical users will suddenly have access to a trove of data representing one of the largest stores of latent value in human history.

Satellite imagery will finally have become…democratized.

Why It Doesn’t Actually Work

One word: demand.

Here’s the thing…Google Earth Engine (GEE) already exists. Imagine dedicating years of your life and millions of dollars to build an incredibly ambitious software platform for processing planetary-scale satellite imagery…and then Google just comes along with their own version and decides to offer it for free to anyone with a pulse…

Dear reader, I don’t have to imagine that, because I personally experienced it. I worked on such a platform, Raster Foundry. I got to enjoy the prolonged and horrifying realization that I was working on a project competing directly with a team at Google that had a bigger budget, more engineers, practically infinite cloud computing resources, and, best of all, no mandate to return any of that ongoing investment.

Since capitulating, I’ve noticed something perhaps even scarier — there appears to be so little commercial demand for GEE that they can barely give it away to companies.

It’s true that they have managed to completely revolutionize remote sensing research (right at a moment in history when we desperately need it). But as someone constantly squinting in their general direction, I can’t quite make out a long line of commercial users wrapping around the block to use this generationally transformative technology on anything more substantive than pilot projects. If there were even paltry commercial demand for such a thing, you’d expect GEE’s website would be teaming with examples. As it stands, every single example on their site is a research or non-profit organization.

Therein lies the main difference between Twilio/Stripe/eBay and the growing pile of satellite imagery platforms and marketplaces. The former tapped into immediate, pent-up, massive demand. The latter is bluffing.

What People Really Want

“No one wants satellite imagery, they want insights,” is a platitude I hear (and say) a lot. People tend to misunderstand the wisdom of it, though. They think, “Of course! People don’t want imagery, they want information like road networks, farm field boundaries, and building footprints!”

NO!!!

No!

No one wants satellite imagery, sure, but also almost no one wants information derived from satellite imagery.

They want insights. Meaning: they want tools to help them make better and more timely decisions. Insights almost always happen at the intersection of multiple, disparate datasets — arriving at them requires a deep, contextual understanding of the other dataset(s) you’re combining the satellite imagery and related derivatives.

Realizing that information is simply not enough is why platforms and marketplaces almost always become consultants to survive. In order to arrive at the valuable thing that people will actually part with money to attain, they must build custom, detailed, one-off software that is mostly about the other data being combined with satellite imagery (rather than the imagery itself).

But Wait, There’s More

Yeah, it gets worse.

Not only is there not demand for platforms and marketplaces, but there are significant structural disincentives working against them. In order to exist, they must partner with imagery suppliers. So, it’s worth looking at what motivates the suppliers — after all, without their participation, there are no platforms or marketplaces.

Satellite imagery suppliers, historically, have made their money through large, ad hoc, defense-and-intelligence deals. This has been a very lucrative business, resulting in several multi-billion-dollar incumbents.

Platforms and marketplaces pitch suppliers with a value proposition like, “work with us, and we’ll drive more business your way.” But even if that were true (which, as I established in the section above, it is not), how much additional revenue would you have to bring to the suppliers to sufficiently motivate them to relinquish control of their distribution?

It’s probably a lot more than you might expect — after all, this is essentially “free” incremental revenue. The platforms and marketplaces are taking on all of the customer acquisition costs.

But, these companies relish having ultimate control over who can access their data and on what terms. The draconian licensing strategies suppliers have historically pursued reflect a culture of conservatism that does not suggest an appetite for the significant additional legal exposure created by some upstart reselling their data in tiny dribbles to end-customers they don’t know.

Plus, they’re worried existing customers that they enjoy direct relationships with will migrate to these other services if their data is more favorably priced or licensed there. They’ll never give a third party better blanket terms than they can offer customers directly. It wold cannibalize their existing business.

In case you have any remaining doubts about how spectacularly this can blow up at a moment’s notice, let’s look at the case of TerraServer.³ Dating back all the way to the 90’s, TerraServer was the first company ever to offer an open marketplace of satellite imagery on the web. In 2015, it was acquired by PrecisionHawk (a very well-funded drone startup) for an undisclosed amount.

Then one day earlier this year, TerraServer suddenly vanished. No one will say what happened — PrecisionHawk hasn’t commented on it publicly, as far as I can tell. It just…shut down overnight.

I can only speculate, but I suspect the terms TerraServer enjoyed for decades came due for reassessment and the licensing gods did not look fondly upon their new stewards, PrecisionHawk. This is a theme that comes up over and over in this industry — if someone at one of the major imagery providers decides you’re done…you’re done. It doesn’t matter how long your track record is.

Even the two largest platform companies in existence today, Orbital Insight and Descartes Labs, appear vulnerable to the precariousness of building relationships with suppliers. For instance, I can’t find any reference to Maxar anywhere on either of their websites. That strikes me as surprising and unlikely — Maxar is the first place I’d partner with if my goal were to make access to the best high-res imagery easier for my customers.

Why I Could Be Wrong

Before even getting into the rationale, I think one strong signal indicating I could be wrong is the large number of very smart people making very big bets on the other side of this trade. More than a handful of startups in this space have raised tens of millions of dollars from VCs, including some of the best institutional investors in VC history (Sequoia, GV, In-Q-Tel, Techstars, etc.).

If I were pitching VCs for funding, my main argument would be that the stagnant oligarchy of suppliers today will morph into a truly thriving, competitive landscape over the next 5–10 years. Just look at all of the suppliers rushing in right now, even just in optical alone:

These challengers are chucking shrapnel into space much more inexpensively than their predecessors, sometimes by an order of magnitude. The premise of building out world-class distribution software themselves is daunting — it still takes several years and millions of dollars to do right. What if they could just plug into an existing set of platforms/marketplaces and monetize their data instantly with relatively little upfront investment?

Additionally, open standards like the SpatioTemporal Asset Catalog specification, Cloud Optimized GeoTiffs, and OGC Features API are reaching the early stages of maturity right now — these are the software-building-block-equivalents of a Rosetta Stone that makes translating between systems easier than ever. As data becomes standardized (read: commoditized) the learning curve for exploiting imagery or building applications on top of it becomes much less intimidating. Reduced technical barriers to entry for non-traditional remote sensing technologists will spur growth for the industry overall as the prospect of getting started with satellite imagery gets less and less intimidating over time.

Lastly, there are some businesses that have managed to make it work at small scale. Apollo Mapping, for instance, has run profitably for almost a decade matching the right imagery to the right customers. They do it with a very consultative, hands-on approach and with a small staff (two folks according to LinkedIn).

The best industries to invest in are the small ones that become large very quickly when no one is expecting it — and the fact that there’s sufficient demand to keep Apollo in business that long is a sign that there’s a real problem being solved repeatably. The question, then, is one of timing — will the influx of new supply over the next five years bring with it a proportional influx of demand?

What You Should Build Instead

Let the existing marketplaces and platforms fight it out — I’ve met a lot of their leadership teams, and I can say without a doubt that very smart, skilled, motivated people are working on those problems. I would rather not compete with them, if I were you.

Instead, focus on a specific problem that earth observation can uniquely address. Become a “Solution” business focused on a particular vertical and compete with the status quo in that industry rather than a sea of other technology startups trying to do the same thing as you.

I believe our industry desperately needs more entrepreneurs willing to commit themselves to obsessively tackling one problem and carrying it all the way through past imagery, past information, to insight. Climate change, humanitarian aid, natural resource management, urban planning, disaster preparedness — the list is nearly endless. Find a topic you care passionately about and go deep.

If you enjoyed this post, or if you didn’t and you’re a masochist seeking to continue torturing yourself, you can hear me share more ideas about this topic on this recent Scene Above Podcast episode. Or you can always argue with me on Twitter at @mouthofmorrison.

¹ Application Programming Interface (API) is fancy software jargon for an organized way to expose data or services online. It makes it easier for software built by different people to exchange information reliably.

² My dad restores civil war era banjos (as one does), and was recently explaining how he can find original parts for 150-year-old banjos on eBay in a few days now (as opposed to searching for years or even decades in the past).

³ TerraServer’s corporate history is probably a long, fascinating blog post in itself. Before it was an independent entity, it was a project embedded in Microsoft and had a licensing deal with GeoEye (before it merged with DigitalGlobe). They were truly innovators.

I believe we are on the cusp of a second golden age of geography. During the first wave, we were on offense. Cartographers during the Renaissance held the keys to unlocking unprecedented understanding about the scope and nature of the world.

But the “second coming of geospatial” that we’re about to live through is all about defense. Climate change is a clear and present threat to humanity’s cumulative wellbeing. The race is underway.

Time is Against Us

There is a dire need to take immediate action to reduce atmospheric carbon (even assuming we manage to reduce greenhouse gas emissions globally through cleaner energy sources).¹

The good news is that the global transition to renewable energy is in full swing. Wind and solar are on pace to be a cheaper source of energy than coal in all major markets by 2030.² When I look at this chart, I’m struck by how incredible it is to be living through such a dramatic inflection point in real time.

The bad news is that we’re collectively realizing that simply reducing emissions is very unlikely to stave off a disastrous mass extinction along with a host of other horrifying consequences. At the infamous 2º Celsius warming threshold, about 18% of insects, 16% of plants, and 8% of vertebrates will see their natural habitat shrink by half or more.³ Drought, flooding, wildfires, sea level rise, vector-borne diseases, and other biblical plagues await us at unnerving levels.

We must therefore move beyond emissions reduction to also reduce the existing pool of atmospheric carbon, which is going to require radically improved land use management at global scale.

Understanding and managing land is where I believe geospatial technology will play its most highly leveraged role over the coming decades. We’ve been cataloging the Earth’s contents in increasingly fine-grained detail for millennia — now we’re going to be forced to make dramatic choices about its composition over the next decade or face entirely avoidable and horrifying consequences.

The thing is…we don’t need a technology breakthrough. We need a resource allocation breakthrough.

Competition vs. Collaboration at Species Scale

I sometimes find myself fighting for scraps with our competitors at Azavea. For better or worse, I tend to adore them. I don’t know what it is about geospatial software that attracts such incredibly charming people. Maybe it’s because there’s not much money in it.

The trouble is that while we’re all busy spending discretionary R&D dollars on open source projects and standards to try to make collaboration easier, we almost never actually get paid by clients to collaborate. When you only get to spend 10% of your time on massive problems that require industry-wide coordination…you don’t tend to solve them.

When I say “we have the technology,” I’m being a little facetious. But only a little. What would you need in order to make detailed plans for saving the world from cataclysmic climate change in the next few centuries? I’d start with a map of every corner of earth, updated in real time, with logic built on top of it to model the effects of any decision in the near- and long-term.

I know a handful of firms that could build that tool today. I just don’t know who would pay them to do it.

What a Solution Looks Like

In order to manage land use with the efficacy required to slow and eventually stop climate change, the first thing you need is high resolution data of the earth’s land cover. “Resolution,” in this instance, refers to four complementary characteristics:

• Spatial (what shapes are visible?)

• Spectral (what physical properties are parsable?)

• Temporal (what frequency does the data update at?)

• Dimensional (what depth do features on the surface have?)

How could we possibly get information like that for every point on earth on a regular basis? We kinda already do.

Last month, in Partnership with Norway’s International Climate and Forests Initiative (NICFI), Planet Labs announced that they’ll be releasing a monthly basemap for most of the central swath of the world. You can scroll around at look at it here. It’s pretty incredible.

That’s a start. More exciting developments are underway. For starters, there’s the Synthetic Aperture Radar revolution currently underway that I’ve written about previously — SAR will inch us further toward affordable, global-scale monitoring of land cover change in perpetuity.

And just at the horizon, work like that of Greg Asner’s Center for Global Discovery and Conservation Science at Arizona State University hints at the kind of detail we can expect to derive from hyperspectral sensors capable of breaking down the electromagnetic spectrum into minute, individuated channels of light that allow for incredibly detailed characterization of ecosystems and landscapes. His recently published work, the Global Safety Net, revealed a shocking return on informed land use management:

We estimate that an increase of just 2.3% more land in the right places could save our planet’s rarest plant and animal species within five years.

New organizations are popping up constantly to carve out a niche within this massive puzzle — recently, National Geographic spun out a new firm called Impact Observatory that aims to provide policy makers with some of the land planning tools I have been hinting at throughout the piece.

Meanwhile, quiet but important initiatives like the Spatio Temporal Asset Catalog (STAC) standard, Cloud Optimized GeoTiffs (COGs), and the Open Geospatial Consortium API Features effort are crafting the common languages that most of these various efforts will eventually speak.

Prototypes are being built. Standards are being written. But mostly, we’re all working in isolation on our own branded repackaging of the same basic idea.

A Call to (Collective) Action

I want to work with our competitors on huge projects trying to make a small dent in the fight against climate change. I think many of them feel the same. Unfortunately, we simply can’t afford to collaborate the way we’d like to right now. As it turns out, it’s extremely hard to get individual entities to pay for a collective good.

What can be done today to expedite progress on understanding and mitigating climate change using geospatial technology? I implore major funders of climate-related work to consider the following four suggestions:

1. Design projects that encourage collaboration between potential bidders. I’m not suggesting that you should eliminate competition for the work. On the contrary — make vendors continue to earn their spot by regularly bidding new work to a shortlist of vetted firms on “Indefinite Delivery/Indefinite Quantity” (IDIQ) contracts. Refresh the list every few years. For truly massive, long-term projects the last thing you want is to grant a monopoly to a firm that will naturally drift toward building idiosyncratic and bloated software.

2. Invest in open standards and open source tools. The Radiant Earth Foundation has consistently hired one fellow at a time to work on critical open standards like STAC and COGs. The OpenStreetMap Foundation is just now finally able to hire its first engineers to dedicate themselves full-time to OSM tooling. These are foundational technologies used by the largest multinational corporations in the world, but they’re being funded in drips and drabs by scrappy charitable groups and informal networks of volunteer software engineers. If you use open tools that are directly applicable to climate-related work, and you are weighing the choice between buying back billions in your own stock or making a few strategic acquisitions, maybe consider setting a little aside to start pulling your weight.

3. Subsidize private data that has become a public good. The data that will underpin climate-focused geospatial applications will not be cheap to produce. While incredible initiatives from governmental groups like the European Space Agency and NASA have resulted in troves of valuable, global-scale data, commercial entities collecting similarly valuable data can’t afford to just give it all away for free. Choosing to partner with private sector data providers to create essential public goods is, in my view, the job of governments and intergovernmental groups. NICFI’s investment in the recent data dump from Planet is a great early example. I hope to see more of that in the coming years, especially with some of the new SAR providers.

4. Make government work more accessible to non-traditional contractors, including unclassified DoD projects. I work for a company that does not work on technology that supports war fighting. But that doesn’t mean we won’t work with the military — in the United States, funding for disaster response, resilience and preparedness, critical infrastructure management, and even foreign aid is often concentrated in the DoD (for better or worse). You need a PhD in bureaucracy to know how to find, let alone navigate, the average military contracting process. A great example of a sane entry point for non-traditional contractors is the Defense Innovation Unit which uses a combination of tightly scoped RFPs and rapid procurement processes to make DoD projects more attractive to tech companies. More of that for non-warfighting projects, please!

¹ For a nice summary, I recommend Nat Geo’s accessible primer from last year, To curb climate change, we have to suck carbon from the sky. But how? For a deeper explanation of the various paths to halting global warming at 2ºC, the The Intergovernmental Panel on Climate Change’s Special Report is worth a look.

² Wind and solar plants will soon be cheaper than coal in all big markets around world, analysis finds, The Guardian, March 2020.

³ A Degree of Concern: Why Global Temperatures Matter, NASA Global Climate Change Website, June 2019

TL;DR

1. Release a small percentage of your data under a truly open license, preferably CC0 1.0 Universal (CC0 1.0) Public Domain Dedication. Ideally, partner with a cloud provider so that data is hosted for free and there are more stakeholders invested in its upkeep than just you (e.g. Earth on AWS, Google Earth Engine, or Microsoft AI for Earth). Let’s call it…1% of your capacity. That has a nice ring to it. “1% for the world…” lol.

2. Decide on a price for tasking and for archival data, and share it publicly. This will allow your customers to plan and to prioritize use cases based on what is economically viable. This information will be shared whether you post it publicly or not, but you’ll engender a lot more loyalty if you just commit. You can always change it.

3. Gate access to paid data however you see fit, but once “in,” allow customers to disintermediate sales people and simply purchase data directly.

4. Invest in education. Marketing is not the same as education — a truly great piece of educational content will help your competitors as much as it will help you, save for the goodwill it will generate.

First, an Acknowledgement of Your Predicament

You face two fundamental dilemmas:

1. The proprietary nature of your data is the basis for most of the perceived value of your company. Giving away your secret sauce is counterintuitive at best, self-sabotaging at worst.

2. You have two privileged classes of customers. The Federal gov’t is one and all foreign gov’ts are, in aggregate, the other. These two customers have vast resources and existential threats you will be helping to mitigate. They are willing and able to pay more than any other classes of customer. Everyone else is, proportionally, noise.

I recognize you have to survive as a company before you have the luxury of addressing the long tail of commercial uses for your data. I don’t begrudge anyone that focuses on government customers. What I cannot stand is half-assed efforts to address commercial customers’ needs. It’s worse than just never opening the door — not only does it distract you, but it creates animosity toward you where none is necessary. And the animosity soon grows to become a mutual disgust — after all, at least you’re making an effort, right? Maxar, Airbus, and Planet catch a lot of shit despite doing more than anyone else to work with commercial entities. On the other hand, have you ever heard of a commercial entity complaining about Hawkeye 360?

How to Effectively Invest in the Long Tail

There is currently next-to-no commercial demand for satellite imagery (of all sorts). But, at least intellectually, there is demand for satellite imagery derivatives. We face, as an industry, a chicken-and-egg problem. There cannot be demand for imagery if none is easily accessible to experiment with. And there will not be sufficient investment to make imagery more accessible without evidence of demand (or visionary leadership willing to take heat in the short term in order to grow the industry in the long term).

The four primary impediments to imagery accessibility are as follows:

• Capacity

• Licensing

• Transparency

• Education

I’m not going to waste time on explaining capacity to the people fundamentally increasing it. But I do want to spend some time on the other three.

Licensing

The cardinal sin of publishing open data is licensing it restrictively enough that it’s not useful for the thing you’re trying to encourage. Over and over again, earth observation companies have released “open” data in support of humanitarian initiatives, or for researchers, or simply for fun. I put “open” in quotes, because if it’s not licensed for commercial reuse, it’s not technically open, at least according to the Open Knowledge Foundation and other similar bodies.

Your mandate is to maximize value for shareholders, not to save the world. So, I’ll be making my argument through that lens: in my opinion, restrictively licensing data that you give away hurts value for shareholders. The nefarious and ignorant will use it for commercial purposes anyway (thus diluting the value of your remaining data in exactly the way you feared by licensing it restrictively). The lawful and diligent will not, but these are precisely the category of companies you want to work with. By truly giving away your data with no strings attached, you’ll unlock in-kind donations as responsible companies allocate their own research and development dollars to projects involving exploitation of your data. And if you truly believe that real-time monitoring is what drives value for commercial customers, then your archive is not that valuable anyway, except as a reservoir of half-baked training data. Treat your archival data like pickaxes and shovels and your tasking capacity like gold.

Transparency

The most common complaint among commercial customers of earth observation data is the sales experience. If you force me to inject a 2-month imagery negotiation in the middle of a 2-month pilot project negotiation…I no longer have a pilot project, and we both lose out. The sales infrastructure you build to support selling into the government will stand in direct opposition to your commercial partners’ needs. In one context, they will seem consultative, organized, and “value-added.” But in the other context, they will appear territorial, bureaucratic, and disenfranchised. Simply get out of your own way. Make the difficult choice of what to charge, and live with it. You can always change it — the stakes are extremely low in the early days. Like I said, there is no significant commercial demand for satellite imagery today. If you raise prices 10x, you might piss all three of us off for a while. But if you never publish prices, you’ll always be stuck with a small group of highly motivated niche partners rather than a massive market.

Transparency also brings with it the benefit of enabling self-service transactions, meaning the cost to serve your commercial partners is lower and they are happier. You’re both outsourcing the consultative sales process for free and you get to call your customer a “partner” so they feel better about the whole thing. No commercial EO providers allow for this currently, to my knowledge.¹ Everyone negotiates on an ad hoc basis and gates access (even for repeat partners) behind an army of sales people. It’s infuriating, and a fairly low bar to clear.

Education

This requires little further explanation. I think it serves your best interests to be seen as a “thought leader” in the satellite imagery space, both because it engenders trust with potential customers… and because it optimizes organic search to your lead generation forms. I think the company that provides the most sincere, accessible, and valuable satellite imagery education will have a compounding advantage over their peers. Investing in education is a way to both grow the overall market of people equipped to exploit your data while incrementally growing your share of it. But none of your marketing dollars spent on education will return an investment if you haven’t solved the problems of capacity, licensing, and transparency — it will be a big “top of the funnel” leading to a pinhole.

Risk

Of course, this is all easy advice to give. I do think this strategy is not without risk. The first and biggest risk is that it’s all a waste of time and focus — you could have been doubling and tripling down on making the U.S. government happy with your product offering and that might have generated a higher return without all of the headache. The second is that your commercial strategy directly undermines your negotiating power with gov’t customers — what if they’re willing to pay more per unit than the transparent pricing you publish on your site? In many cases, the U.S. will contractually obligate you to give them the best pricing you give anyone (although companies skirt this all the time by defining their products in such narrow ways that it effectively gets them out of this bind). Lastly, you might find yourself enabling competitors. Because you are choosing to generate analytics yourself, some of the markets that internal team will serve will also have external partners that come to you for imagery. Navigating a situation where you are competing with your own partner is tricky, and usually results in bad blood. Too much of that, and your reputation will suffer (after all, it’s a small, small geospatial world).

It would be easier not to entertain any of this at all. But ignoring non-gov’t uses for your data will cap your valuation and your impact.

In chess, there’s a concept of a “poisoned pawn,” a pawn seemingly free for the taking that requires making a positionally fatal risk to capture. I think dabbling in commercial use cases is like grasping at the poisoned pawn — it’s doomed to fail. If you reach for it, you have to fully commit to the complexity that will arise as a result.

¹ I am not referring to the experience of trying to sign on as a reseller for a satellite imagery provider. I am talking about self-service access to raw data that doesn’t involve a complicated legal agreement and sales volume commitments. That’s an open letter for a different day.

I recently wrote about the commercial satellite imagery industry and how fundamentally backwards it is. In that piece, I was mainly addressing companies serving up data derived from the narrow swath of electromagnetic spectrum we humans are most comfortable with — visible light (and its nearest neighbor, infrared light). As it turns out, there’s a whole world of spacefaring companies capturing data about the earth in the oft-overlooked expanses of the electromagnetic spectrum that we humans can’t naturally perceive.

The emerging subset of the industry that I am most excited by is the cluster of companies focused on synthetic aperture radar (SAR), a very complicated-sounding name for an equally complicated technology.

I’ll explain more about what SAR is and why I find it exciting below, but I’d first like to acknowledge companies like Spire Global and Hawkeye 360 that use satellite constellations to measure radio frequencies in order to produce interesting analytic products (a blog post for another day). All too often the “earth observation industry” is reduced to its most familiar use case: pretty pictures of earth. But there’s a growing number of less intuitive, equally compelling types of earth observation that get little attention outside of the occasional profile in space industry trade publications.

A Brief History of SAR

SAR is nothing new. Early airborne experiments were completed in the 1950s. The first spaceborne SAR sensor, Seasat, was launched by the U.S. government all the way back in 1978 — remarkably, much of the data collected by that mission has been made available for free today by NASA. Contemporary SAR constellations like the European Space Agency’s Sentinel-1 mission have enabled a much broader range of researchers and companies to start dabbling in SAR analysis at no cost. And private investment in the commercial side of the industry has ballooned in the last year promising a new wave of high resolution constellations (more on that later).

SAR has some unique properties — it is famously a “24-hour, all weather” sensor, because it can “see” through clouds and at night just as easily as it can “see” during daylight with clear skies. SAR data is anything but human-friendly; after all, it’s measuring microwaves rather than visible light. But it can be processed into derivative products that are interpretable by human eyes, although doing so often means having to eliminate useful information present in the raw data (not altogether too different from color-correcting an optical image for human eyes rather than statistical inference).

Unlike optical imagery, SAR can also be used to derive changes in relative elevation over time, literally measuring which parts of the earth are sinking and which parts are rising. USGS has used SAR to measure land displacement from space going back over 20 years with sensitivity to changes on the order of 5–10 millimeters…from space!

For a full summary of SAR’s unique characteristics, and an explanation of the physics behind how it works, I can’t recommend Daniela Moody’s SAR 101 video enough.

The Commercial SAR Industry: A Cresting Wave of Innovation

The vast majority of SAR innovation over the last forty years has happened behind classified doors. It makes sense — if you’re interested in tracking enemy troop movements at night or mapping post-hurricane destruction under cloudy skies, you can’t use optical imagery even if you can task a satellite at the drop of a hat. Only SAR can “see” in those conditions with global reach, so it has a unique strategic role in defense applications.

In 2008, the existence of the US military’s Lacrosse constellation was declassified, revealing a total of five SAR sensors that were launched between 1988 and 2005 (three of which remain on orbit). Germany, Russia, Italy, Japan, India, and other countries have all launched SAR missions over the years.

Only a handful of commercial companies have successfully launched and serviced SAR satellites to date. MacDonald Dettwiler and Associates (the star-crossed acquirers of DigitalGlobe who late last year were piteously spun back out by their own acquisition target) has been operating a commercial SAR constellation since the launch of RADARSAT-1 in 1995, and is still operating its successor, RADARSAT-2.

But the commercial SAR floodgates are only just now, this year, slamming open.

Why SAR Promises to Revolutionize Commercial Earth Observation

When I say SAR is about to “revolutionize commercial earth observation,” I don’t mean we’re on the precipice of an overnight transformation. What I mean is that aspiring SAR purveyors are collectively ushering the rest of the industry toward a crossroads faster than we would arrive there without them . For years, startup after startup has been pitching the ability to construct an oracle-like prescience about the future and the past — better crop yield prediction, flood modeling, oil inventorying, construction activity monitoring, deforestation alerting, you name it.

Underneath the shiny patina of the average remote sensing startup’s fancy marketing is the implicit promise of contextual, meaningful, automated change detection. Since I started paying attention to the steady stream of remote papers, I think change detection is the single theme that has dominated all others. Sure, technically, it’s possible. But in the practical sense that people imagine when they first hear the term, I think we have a long way to go.

Today’s production-grade change detection algorithms are extremely brittle and narrowly scoped, laborious to create, and incessantly in need of ongoing maintenance. Companies that create them, one client and project at a time, are the opposite of scalable (a Silicon Valley euphemism for fundable).

What is holding the industry back from achieving ubiquitous, coherent, generalized change detection? Four things, in my estimation:

• Currently there isn’t an overabundance of high resolution, task-able satellite imagery capacity (especially after the premature death of WorldView 4 last year). Planned constellations like Maxar’s Worldview Legion aim to ameliorate this problem, but the reality today is that you’re going to be fighting over scraps for high resolution imagery, especially if your use case requires lots of revisits.

• Even if there were more capacity, the price of imagery has remained stubbornly high despite increasing capacity. Anchored by the price governments are willing pay for imagery to support national defense efforts, satellite operators have struggled to serve most commercial use cases (which often start making economic sense at just a fraction of the current retail prices).

• Clouds are a way bigger deal than I like to admit (even to myself). Most of the world is obscured by clouds most of the time — only about 30 percent of the landmass of the earth is completely cloud-free at any given moment. Add to that constraint the vexing fact that you can only “see” during daylight hours, and your window of opportunity for collecting information starts to look like a rigged roulette wheel.

• Even if you somehow had access to ample high resolution imagery and you could afford it and you were focused on an area that was rarely ever cloudy…automating the analysis you’d like to perform would likely still be limited by one extremely scarce resource: training data. While the advent of deep learning has enlivened the remote sensing community with far-reaching possibilities, the sad truth is that amassing the requisite volume of annotated imagery remains prohibitively expensive and time-consuming for all but the most highly motivated commercial entities.

Change detection may be the talk of startup town, but in reality most imagery sold to commercial customers is used for a much less exciting purpose: static basemaps. Companies like Apple, Google, Microsoft, Mapbox, and a laundry list of lesser-known web mapping providers purchase high resolution imagery by the boatload in order to make those pristine, cloudless mosaics we’ve all grown accustomed to.

I’m skeptical of the rate at which passive-sensor-derived change detection will become feasible in any context outside of national defense given the challenges I laid out above. Companies like Nearmap and Near Space Labs are aiming to grow the commercial market by providing low-cost aerial imagery over densely populated areas, which is a model I’m hopeful for. Overall, I anticipate a long, slow slog. However…

In my opinion, the most exciting aspect of the impending SAR revolution is that it will accelerate the advent of global, timely change detection.

SAR does not easily address the single largest commercial use case for satellite imagery: static basemaps. It’s not an inherently visual product — it’s an inherently analytic one. SAR’s properties make it ideal for high-revisit, high-resolution activity monitoring and feature extraction. You wouldn’t use SAR to map buildings and roads over a large area — you would use it to figure out which roads are inundated with water during a hurricane, or where new houses are popping up at the outskirts of a fast-growing city.

And unlike the optical portion of the industry (which has experienced merger, after merger, after merger, after merger), the competitive field for SAR providers is experiencing historic expansion. ICEYE, the Finnish startup which made headlines recently after raising $87M in a Series C financing round, has a head start — it currently manages a constellation of 3 satellites with 12 more planned in the next two years. Capella Space’s first satellite reached orbit just over a month ago with plans for 30+ more to come. The Japanese startup Synspective raised $100M and announced the launch of its first satellite will happen this year. PredaSAR announced $25M in funding earlier this year, while Umbra Lab came out of stealth in April as their patent application for a new SAR antennae design became public.

Whew! And that was just a sampling. Seriously.

The Impending SAR Crunch

With so much new SAR capacity coming online over the next year, the question is…what will happen to all of that imagery? If the optical side of the industry is any indicator…a whole lot of nothing, at least for a few years.

I recently spoke to some graduate students at the University of Pennsylvania, and I described to them a dilemma that the SAR explosion is only going to exacerbate. As an industry, we can’t keep up. We’re like the dog still chained to our traditional analytical tools and proven use cases, speeding away in a convertible fueled by raw satellite imagery…

And that’s a problem in the optical imagery realm, where I can actually explain to my parents what they’re looking at. SAR is far less intuitive, far more fragmented in terms of file type standardization and metadata specification, and enjoys far fewer open source tools for lowering barriers to working with it. It’s a steep learning curve (SAR idiosyncrasies) stacked on top of a steep learning curve (geospatial raster data) stacked on top of a steep learning curve (petabyte-scale datasets in general). Education will take time, and product-market-fit in the diverse verticals that could benefit from SAR-derived analysis will take even more time.

The princely sums that ICEYE, Capella, Synspective, PredaSAR and others have raised suggest a strategy similar to Planet Labs. First: launch a constellation and operate it. Second: build out an analytic product suite and sell derivative information. Third: ???. Fourth: Profit.

The reason for this strategy is that venture capitalists don’t actually want to fund the next Maxar. Personally, I think Maxar is just about the the perfect business: technically interesting, undeniably meaningful, legitimate competitive moat, highly regulated monopoly, etc… but it’s valued at less than 3x revenue on the public market. Satellite imagery companies sell for multiples closer to microchip manufacturers than SaaS companies. Multiple expansion one-time liquidity events are how you make your money back in venture capital, not dividends.

VCs want to fund high-margin software with zero marginal cost at scale. They want the 10–170x revenue valuations (seriously) that SaaS companies enjoy on the public markets. And to make that argument, they encourage their portfolio companies to focus on repeatable subscription revenue. Thus, look for analytics to be sold directly by SAR providers early and often — that will be the narrative hook for pitching future investors (or acquirers) on how they’re really a software play, not a satellite imagery play.

But in-house analytics cannot build a thriving commercial segment on its own. The beauty of the geospatial analysis world is that it cuts across almost every vertical…and in this regard, it is more-or-less impervious to productization. I think of Esri as an analogy: its in-house services are prodigious, but its legendary partner network is what helped grow the overall market at a faster rate than Esri could capture it. For SAR to be successful, a similar ecosystem will need to emerge.

Ursa Space Systems is the most advanced SAR-focused, third-party software and analytics company I can find. They’re extremely well positioned to benefit from the influx of new providers. Cloud to Street is another SAR-savvy company focused on mapping and modeling flood inundation and doing an incredible job at it. But the list of companies like Ursa and Cloud to Street is surprisingly short. I see this gap as a huge opportunity for new company formation — especially folks willing to focus on a single vertical and build a great user experience that masks the complexity of the SAR processing happening behind the scenes.

Well-regarded firms like Orbital Insight and Descartes Labs will undoubtedly add high resolution SAR processing to their ever-expanding list of services — Descartes even recently released a paper describing a SAR-derived deforestation calculation. And Indigo Ag is leading the agricultural charge into SAR — they recently showed off their SAR-derived, country-wide crop monitoring system and added a 45TB of freely accessible processed imagery to AWS. SkyWatch contributed to an open source SAR-processing toolset under ESA’s direction and provides access to Kompsat-5 SAR data via their API. At Azavea, where I work, we’re currently using SAR data on both commercial and government projects, as well as incorporating it into our satellite imagery labeling tool, GroundWork.

The industry has taken note. The most innovative companies are positioning themselves to move quickly. But the revolution won’t happen overnight. It’s going to play out in painstaking slow motion over the next five to ten years. But, one day soon enough, we’ll come to realize that commercial SAR helped crack the code on affordable change detection and, along with it, uncorked a tidal wave of economic value.

The bottom line is this: as an industry, we’ve been promising global-scale monitoring and change detection for years. As SAR goes from “tradecraft” to “trade show,” it’s going to be time to put up… or shut up.