The Future of Earth Observation: In the Foreground, Middle ground or Background?
Thoughts from an Outsider
Earth Observation in March 2021: Setting the Context
We have just entered March of 2021. But, it seems like so much has happened in the space sector in the past couple of months - from a number of SPACs to a few Starship ‘belly flops’. I don’t know about everyone who works in space tech or is just a space enthusiast, but it’s been a tad overwhelming for me to keep up with everything, with some FOMO (So, I decided to curate news for myself, but then I also decided to publish it for those interested). It’s only been a couple of months since I put out my last post about the iPhone moment for the space industry, but the developments of the past few months have got me thinking about the future of Earth observation (EO). But first, let’s take a step back and first take a look at those interesting developments that got me thinking (as quick hits):
We have yet another radar data (SAR) constellation, called Alpha Insights, and another one - Umbra, which we hadn’t heard from in a while, has raised some funding (I had one of their employees, Joe Morrison, on my podcast and it seems like they are working on making sure accessing & buying satellite imagery is as easy as shopping for a house with transparent imagery pricing, easy-to-access interface etc., something Joe is extremely passionate about).
I learnt that there are more EO satellite constellations (yay!), but at least they are planning to acquire different types of data - Aurora Insights (radio frequency data), Orbital Sidekick (hyperspectral data), Hydrosat and ConstellR (thermal infrared data).
Iceye, the radar data provider with its SAR constellation, has entered into a strategic partnership with Swiss Re for reinsurance solutions for flood risk management and disaster management. I am delighted to have some news in EO where the paying customer is not just a governmental institution or the military, but personally, I was expecting a major insurance provider to announce plans for their own constellation (I called it first!).
There was some consensus from Earth observation industry representatives that “ companies are collecting more Earth imagery from satellites than ever before, but for some customers, the data remains too expensive and too difficult to consume.” I am glad because the first step in solving a problem is accepting that there is one.
The National Geospatial-Intelligence Agency (NGA) might want to acquire more satellite data and is “stepping up research to better understand the emerging commercial geospatial intelligence market.” Good news, because some of these EO satellite constellations, might at least get their favourite customer, after all.
There were a few more developments that didn’t make the news, even within the EO community, as much as they should have:
Farmers Edge, a Canadian agtech company went public through an IPO and seems to be doing well (so far). Oh, and it uses insights extracted from satellite data within its products - Planet until 2020 and then Airbus for the next few years (apparently, Airbus is 60% cheaper than Planet, but that’s another story!).
EOS Data Analytics, a company that offers information products based on satellite imagery for predominantly agriculture, among other sectors, wants to build its own satellite constellation (optical data), and become vertically integrated - data collection, data processing and product delivery. In a similar vein, SatSure, an Indian startup that offers geospatial analytics products and platforms for agriculture, banking, insurance and infrastructure sectors, is also planning to start launching its satellite constellation in 2022 (that’s 10 constellations so far in this piece if you have been keeping track).
And finally, bear with me, my two favourite (and potentially most interesting) nuggets of news, that actually led me to write this article.
Three Australian startups will work together to build and launch (indigenously) a bushfire detection satellite, led by Fireball.International, a company that specialises in wildfire monitoring and prediction systems (that are already in place and used for the Calfornia wildfires).
ClimaCell, a Boston-based company that offers weather software (mobile app and APIs) announced its plans to builds its own satellite constellation, producing radar data that will be used to improve its own weather forecasts, as opposed to selling data to others (entirely different from companies with similar constellations - GeoOptics and the recently SPAC-ed Spire, who have the NOAA as their anchor customer).
Now that we are all on the same page with respect to the developments …
If you have read my previous posts, you might know I like to simplify (or oversimplify) things, by introducing some random terminology. I have done that again, and in this case, for defining different approaches that the Earth observation industry is taking to become that multi-billion dollar market, which is to be in the foreground, in the background or find and settle on a middle ground.
The Foreground Approach:
Continuing with the current approach of selling EO data through APIs and interfaces, with hopefully transparent, scalable pricing options (as envisioned by Umbra). The key value proposition that would push growth with this approach is either the ability to collect better and more EO data (infrared data / high-resolution imagery / emission data etc.) or the capability to facilitate the processing of EO data through some sort of value-added services (object detection / change detection / analytics). More significantly, this growth is dependent on the uptake from those who have an understanding of working with satellite data and extract the needed insights for them. And most likely, the uptake is contingent on the EO data providers acquiring anchor customers for selling the data (large contracts), along with a few institutional partnerships and ad-hoc users.
This approach is still business-as-usual, as value is still leveraged from selling EO data be it to the NGA (defence & intelligence), Iceye’s recent partnership with Swiss Re (insurance) or Planet’s super recent partnership with Solinftec (agriculture). Hopefully, with an expected increase in the supply of data (we are at 12 EO satellite constellations now), EO data can actually become democratised and drive the growth of the industry. Platforms and marketplaces such as UP42 and Skywatch might help facilitate the adoption, although they are still heavily dependent on the strategic roadmap of the EO satellite companies / data providers. This foreground approach can potentially pay off in the long-term, assuming that the challenges of pricing and accessibility are solved.
However, there is a caveat - not many folks within the larger tech industry (i.e. outside the EO community) actually understand remote sensing or have used geospatial data before. Convincing them to buy more EO data is probably the path to the multi-billion dollar market. But, apart from the above-mentioned anchor customers who already understand the value of EO data, have strong competences in processing it and know how to integrate it into their workflow, I am not particularly very bullish on this approach in the short-term. Perhaps, that is why the recently SPAC-ed Blacksky and Spire position themselves as both data and analytics providers thus settling on a middle ground. In the long-term, more companies across different sectors could perhaps start to realise that they need to build in-house EO data teams in order to take advantage of all that value from satellite data (probably the best-case scenario of all EO companies today, who are hell-bent on a foreground approach to growth).
The Middle Ground Approach:
Insights are probably more valuable than data (pixels). Some EO companies are starting to realise that although the foreground approach could fetch them some anchor customers and ad-hoc partners (with a sexy value proposition of “solving problems with data collected from space”), they have much more of a chance of getting to a bigger share of the market by providing information products to the vast majority of the users (who do not have the expertise in acquiring and processing EO data). As such, a segment of the EO market has obviously settled on a middle ground approach, wherein they take on the onus of acquiring whatever type of EO data they want from potentially different data providers, process them using proprietary machine learning algorithms to extract relevant insights, sometimes combine and calibrate those insights with other types of data (e.g. IoT, drones, financial & economic data) in order to offer an information product to end-users.
Some of these companies possess significant domain expertise in the verticals they operate in (e.g. Kayrros, Cape Analytics, RS Metrics), knowing their respective end-user needs, buying habits and required go-to-market strategies. However, even though they might have bridged the knowledge gap (by hiring experience personnel from the specific sectors), their market share is probably limited by the major players operating within the respective verticals. Because, as a customer, you ask yourself, I already pay for enterprise software X (depending on the vertical you are operating in), is this new, cool, EO-based product worth the extra cash? It is not an absolute given yet. Perhaps, that is why strategic go-to-market partnerships seem to be an underlying theme in this segment - SatSure + AXA, Orbital Insight + Bloomberg Terminal, Cape Analytics + Duck Creek Technologies, LiveEO + SAP and so on. A couple of these companies have decided to go all-in (EOS, SatSure) and build their own constellations to become vertically integrated, in the hope that they can strengthen their competitive offering and become a one-stop solution to their customers, and perhaps even steal some market share from the big enterprise corporations and market leaders. And apart from these names, there are always going to be some fringe companies who will continue to operate and offer their products, derived from EO data, independently to ad-hoc customers simply because of the niche sector they are operating in.
Is there a caveat to growth through a middle ground approach? Perhaps. The larger corporations and names within specific verticals (insurance, agriculture, financial services etc.) might decide to form their own internal EO data processing teams in order to build their own set of tools that perfectly respond to their needs, thereby eliminating the need for a middle ground approach altogether. I don’t expect this to happen in the short-term as they are probably still in the discovery phase of understanding what EO can do to help grow their market share. Once the product-market fit is established, we might see more consolidation here, with the successful EO companies (both data and analytics providers) being acquired by the major corporations of our time - the SAPs, Oracles, Salesforces, AXAs, Swiss Res and Munich Res (in fact, it has already happened with IBM with the Weather Company that offers a Vegetation Management service, powered by satellite data). They could go via the acquisition route or, they might just decide to push towards a space strategy (because, why not!) …
The Background Approach:
I may have changed my mind about quite a few things in EO and the larger space sector, but ever since I started working in space tech, I have always stuck to this fact: EO data is just another type of data. End-users, be it a consumer using a mobile application or an employee in a major corporation using an enterprise software application, do not care how they get their answers, they just want to get their “job done”. Just like how I didn’t know which communications satellite was used to broadcast the football game on Sky Sports last weekend. Just like how I didn’t know the data from which weather satellite was used to predict my weather this morning. Talking about weather, allow me to explain why I am very excited about ClimaCell and its space strategy, and what I call the background approach.
Unlike a company following a foreground approach to growth in a similar vertical (Spire and GeoOptics), ClimaCell is not a space tech company that plans to offer its data and products for use in the weather industry - it is rather a private weather software company that plans to use satellite data to improve its products.
Unlike a company following a foreground approach to growth, ClimaCell does not need to look for anchor customers (like NOAA or the US DoD) and figure out a specific go-to-market strategy for each customer and vertical - ClimaCell already has some well known paying customers: Amazon, Uber, Ford, Google, Intel, Porsche and so on. In fact, I doubt their customers will even care if satellite data is used (maybe they might just see an upgrade for better weather forecasts through satellite data option in their software suite).
Unlike a company following a foreground approach, space will still not be part of ClimaCell’s core value proposition (i.e. their pitch decks will not drastically change from offering weather software services) - in fact, I would imagine that the insights from their planned radar data satellites are either completely integrated to their core product or available as an optional feature for their users.
Now, am I suggesting that their space strategy is going to go ahead with no difficulties whatsoever? Not at all. In fact, I agree with some experts in the space industry who have questioned their budget ($150 million) as well their competencies (although they have hired John Springmann from Spaceflight Industries involved in Blacksky’s satellite constellation and plan to collaborate extensively with a team at MIT). To be honest, even the tech giants - the Big Five (Amazon, Apple, Facebook, Google, and Microsoft) - have not fully succeeded in at least some of their extravagant hardware plans (remember Google Loon or Amazon Fire Phone?). So, by historical standards, the probability that a relatively small startup, ClimaCell, would pull this off still remains quite low. However, I will leave the feasibility analysis to the experts within the space industry and weather domain. What I am excited about is the background approach, that I have been writing about since I started blogging last year - to consider EO data as just another source of data and build it onto the product, making the product the selling point rather than the underlying technology itself.
The other news about Fireball.International planning to launch a small satellite reinforces my viewpoint. This company seems to already have a working wildfire monitoring system that is already being used by institutions across the United States and Australia. The insights from the EO satellite that they plan to launch is going to be integrated into their existing wildfire monitoring & prediction system. The focus is on the product, once again, rather than EO itself. Sure, the product is completely dependent on insights from EO, but I would imagine that if I am in the middle of a bushfire season and get an evacuation notification based on predictions, I would not care about the algorithm it uses or the satellite that was used in the system.
I was also very excited to find that IBM uses Airbus Pleiades imagery within the IBM Weather Operations Center, to provide insights from EO data, combined with other sources of data, for use by institutions, utility companies, insurance corporations and in agriculture. What is more important is once again, EO sits in the background, powering this software and providing insights that only satellites can provide.
Long Story Short…
As you might have realised, I am betting on the background approach in the long term. Perhaps, it is quite obvious to a lot of you, and this entire piece just deserves a “duh!” but because I have written all this, I am just going to continue and finish up. I believe that the biggest value of EO is when it is in the background, quietly enabling all the big enterprise software applications of our age and solving the biggest challenges of our time (from food security to economic development, from climate change to disaster management). Irrespective of the vertical of interest, the final customer makes a decision based on two important factors - the quality of service (is it adding value to me?) and price (is it worth it?). Whether the software uses EO data or not, the focus should be on solving the problem for the customer and more often than not they do not care how it is being solved. Most customers still do not have a space strategy to solve their problem, when they do, EO can finally be back in the foreground.