New horizons, new horizontals

A playbook for smart space EO ventures

Putting a device in space is a compelling, inspiring idea. What dream could be bigger than touching the future?

Photo by NASA on Unsplash

I was asked recently what I think of when I consider space. My answer, after some reflection, was 'fragility.' Our atmosphere is thin compared to the size of our planet, and our planet is not even a speck on the scale of space. That atmospheric skin is unique in our solar system and perhaps even our universe. A skin which has supported all known life. We can go to the Moon and to Mars, and we should, but we also need to be cognizant of what we have on Earth and its fragility. If we are to launch sensors, we must do so with intention and deep value in mind. This note is a briefing for those who feel they might have a good idea worthy of a ride to space.

This is not a discussion of new vs old space. I feel that argument is stale and largely filled with bravado and marketing. Instead, I will use the lexicon we have been developing around modern strategic thinking and unsettled markets. This language will hopefully inform those interested in building the future in a sustainable way.

That future is not new, but it might be smart.

We have been going to space for some time. But we have been using space since we looked up at a clear sky. I would argue that other than satcoms, navigation is still the single biggest use of space. The technology has advanced, but the product remains the same: getting to places.

That said, satellite communications have also burgeoned recently, and our low earth orbit is being decorated (or littered, depending on your perspective) with small communications satellites. Communicating is arguably the most critical human activity, and access to knowledge is a huge issue of equality. Satcoms, therefore, hold intrinsic value but at a cost.

So, where is Earth Observation (EO) on this spectrum of cost to value? I believe that EO has been central to the human understanding of landscape change. I also believe that our understanding of scale would be skewed without pictures from space. However, the barrier between pixels and people has been glossed over in most venture pitches for new satellite constellations.

The human value of the hundreds, even thousands of commercial EO satellites launched in the last decade has not yet been robustly demonstrated. And that's our fault.

Yet, our landscapes are changing in response to population pressures and climate change. Earth observation encompasses the best-suited technologies to provide meaningful measurement and monitoring capability. So, why is there a gap between demand and what seems to be a suitable offering? In classic strategy speak, I am going to summon the power of alliteration: Inchoate demand and Inadequate technology.

The demand is inchoate

I discussed this in more detail a few weeks ago. But the demand for EO is inchoate, which means there is no clear product. More so, those who want the product actually don't want the product, they want a derivative.

The technology is inadequate.

Yeah, no one wants your pixels: get over it. They might want the story your pixels tell. But, that story is not told in spectra it is told in numeric data and words.

There are some organizations that can consume pixels: large agricultural companies, scientific institutions, some resource companies and military/government organizations. These groups employ vast numbers of geospatial experts to look at images and discern facts and analyses about our planet. But this expertise is limited; frankly, there are not enough eyes to look at every pixel captured. Even today, I would argue that the vast majority of sensory data captured of our planet is not viewed or touched ever again. The data is captured, downlinked, stored and forgotten about.

Interestingly, this is a software problem, a geospatial problem. Around the geospatial sector, a variety of complementary assets have emerged. These assets now allow a series of new capabilities, yet these new opportunities have remained somewhat unexplored. And, because the geospatial sector has been wholly inadequate in communicating its own capabilities outside its small and somewhat insular echo chamber, the broader community simply does not know what is possible. As a result, there are no dominant designs.

---

So, what is a constellation to do?

The most common cliche in the space sector is that "space is hard." Apparently, though, finding a market is harder.

My experience is not as a hardware engineer or scientist but as a geospatial practitioner. I care about the use of data, and if you are involved in the development of new constellations of EO satellites, you probably should too.

In response to the apparent difficulty in finding robust commercial traction for EO data, this is my proposed playbook for a new, smart EO constellation:

1. Go open

2. Build a platform, but sell an API

3. GIFs are great, but they are not a product

4. Do one thing, and leverage

5. Don't sell, partner

6. Algorithms are your customer

7. Embrace custom (at least for now)

Ok, let’s dig in.

1. Go open

I talked about openness as a key complementary asset a few discussions ago. This is the case generally in geospatial, but in EO, openness is pivotal.

I don't care how big your ego is, you are not the only constellation in town.

Additionally, I assume you want customers to "use your stuff." So, you need to build a capability to interact with other sensory products. That means open standards. If you are not publishing your data in a broadly consumable manner, you should just go back to your artisanal teeshirt printing gig.

Firstly, adopt STAC. Secondly, determine a robust open data strategy. How will people use your data? The vast majority of data only gets used within one day of capture, are you still charging for data a year old? When we are trying to persuade organizations to adopt EO, charging the same price for old data is ridiculous at this point in our commercial EO market. Put your ego away and make archive access nominal. Not necessarily free, but make cost negligible. I'll touch on pricing again later.

2. Build a platform, but sell an API

Excel is the dashboard of the financial sector. How does your data look in excel? Exactly. Now we've cleared that up, you now know something needs to happen to those pixels before they become really handy. You will need an intermediary of some sort. That intermediary will be choosing which constellation to pull data from, your job is to make your data easy to use. the best way to achieve this is to adopt common API patterns. But you also need a way to share, publish and administer data. So, you do need a platform, but it will only be used to copy and paste API URLs.

3. How Gifs have built and broken EO

I love a good gif. Gifs are an amazing way to show landscape change. We are continually shocked by gifs of retreating glaciers and reservoirs. They are a wonderful media tool, so easy to share. Unfortunately, these are not at all quantitative. And sadly, Gifs are also misleading to the general public. It makes it look like we are actually robustly monitoring, not just building manicured snippets for social media consumption. Yes, use gifs to tell your story, but make them real and their production platform driven.

4. Do one thing, and leverage

I'll get in trouble for this. But I think constellations should be intentional. I believe that the cost of launching boxes of rare earth minerals into the sky should come with a promise of addressing at least one real human problem, not just a series of letters of interest from government bodies who will buy virtually any sensory data on orbit. Some feel that pure data is a reasonable play. Sure, but then have deeply committed partnerships.

The advantage of doing one thing is that you know what you are selling. If you just sell pixels but never use them, then you always rely on others for product quality feedback. This means you are always late to discover you are doing something wrong. If you are doing things wrong, being the last to find out is an awful place to be. So, eat your own dog food on at least one analytic, even just as an experimental way of testing your own data feeds.

This is when you say:

But, we are a satellite company, why should I be writing analytics?

and I respond:

No one cares about your data, they care about the story it tells, you need to be good at telling that story. So, be good at it.

Now, assuming you are doing that, you should deeply understand how your data is used. That knowledge is your fulcrum, your point of leverage. Put that to work.

5. Partners

Selling images sucks. Very few institutions can consume images. Additionally, when sold individually, they need to be wildly expensive to cover commission costs and to inflate the market. It's a crazy model.

Additionally, landscape change is a horizontal activity, like geospatial in general. So your product and sales teams will need to know everything about everything. The problem is no one does. No matter how big your company is, you cannot have a sales team which has a deep customer knowledge of every sector that geospatial and EO will touch, that’s idiocy. Instead, hire account managers, build partnerships, and treat those partnerships like a venture portfolio. Invest in promising domain-specific technologies or service companies and take royalties.

This partnership play is also a resilience play, I will let you noodle on that.

6. Sell to algorithms

If you don't have an algorithm as a customer profile, then you should rethink how you propose to do business in the future. There are too many pixels to look at, so all future monitoring applications will need an algorithm to touch pixels before any eyes do. What is the price for an algorithm to look at data in that case? Is it cheaper or more expensive for an algorithm to view pixels than eyes? Do you just have geographic all-you-can-eat subscriptions? One could argue that because eyes are only occasional customers, they would come for free with an algorithm subscription. Do pixels have to be downloaded or moved to be touched by an algorithm?

There are so many questions here, but you need to explore them. The EO products of today will not be the same as the products of even the near future.

7. Embrace custom

I'm biased, but I think we can all agree that EO-derived geospatial analytics is an unsettled, nascent market. You can hate me for saying this, and your venture board will be upset with this, but you should embrace custom algorithm development. Yes, services provide lumpy revenue and often do not have a repeatable component, but the market is presently in an explore mode. Services are an excellent tool with which to explore and create value/revenue. Do this right, and your services team will be independently profitable whilst providing your product team with critical market feedback.

Services seem poisonous presently, and there is a risk if managed poorly that they become distracting, but they can provide helpful revenue and the capability to go an extra step for customers. They turn a no into a yes.

However, if you avoid hijacking your product development team and keep product and service organizations separate, you can build a more resilient company. By allowing the services team to operate broadly, they can act as a sales tool whilst being arms-length from an account management or sales team.

---

I am sure I have missed numerous parts of the value chain here. My experience is around the use of EO data, not in any of the hardware development or manufacturing of a constellation. That said, I feel that this final step is often overlooked, and some planning up front encourages useful reflection.

Hopefully, these ideas provide a loose framework for the development of a resilient, new EO company. Over the next few weeks, I will dig into these seven ideas in Strategic Geospatial discussions. As always, feel free to leave a comment.