Risky business

Spatial finance is emerging as a significant component of the geospatial landscape.

Geospatial is shrugging.

As I have suggested before, this change has been happening incrementally for the last decade*. But on a recent trip to London, I feel that change has now reached something of a tipping point towards geospatial technology being more generally consumed by the finance sector. From what I see, there are a number of drivers.

Caveat: As an entrepreneur, I am notoriously optimistic; perhaps I am early. So, feel free to disagree on timing. If I have a blind spot, that's it.

Also, Don't, for a second, think that I am diminishing the magnitude of activity in the Defence and Intelligence (D&I) sector (that of the US specifically, but the rest of the world more generally, too.) D&I will continue to be the dominant funder of geospatial data and software companies. But what I want to illustrate is the rise of the commercial sector, which modern Earth Observation (EO) companies have been so desperately seeking and upon which their 2020/21 Special Purpose Acquisition Company (SPAC) valuations were based.

Is this a story of rampant optimism or one of better late than never? Perhaps we are still in for a false start, and there are numerous market complications which could accelerate or stunt the growth of this delicate market flower. Nevertheless, some of the appropriate market conditions are now in place.

To understand the signals I am seeing, I will try to weave a broader market tapestry from a number of situational threads. This note is an effort to help me draw those threads together into what is hopefully a holistic and understandable story.

Part One: Schumpeter wins again

The modern insurance sector is massive, but much like the banking sector, it is built on a series of acquisitions, mergers and compliance fear. This means every major insurance company is presently babysitting an on-premise Jenga tower of teetering and malformed technology blocks which no individual wants to touch for fear of their job and reputation. Thus, the motivation to "change" in a major insurance company is low.

Wait, this doesn't sound promising for geospatial.

But where we see a problem, there is a cadre of mid-career insurance people who have been living in this situation. For twenty years, they have been told to deal with it because that's how it’s done. The real reason is that some internal process seems to be working, and no matter how square the wheels are, those wheels are the wheels we have and we understand how they work. So, through their careers, this cadre has seen no change and little willingness to risk trying. Rightfully, they are identifying this inertia as nonsense and are starting their own new breeds of insurance companies: mid-career entrepreneurism, born of frustration. Built on new technology stacks and new data products, these companies want to be innovative and actively intend to have a massive impact with a surprisingly small headcount. There is little more fearsome to the establishment than this group of domain experts with a clean slate and a willingness to build.

So, in the short term, this group are accessing legacy capital, while building new capabilities. In the slightly longer term, they will be able to access their own capital. And, the Schumpeterian snake will once again eat its own tail to ultimately renew itself. Shedding the skins of legacy workflows. This creative destruction is one signal of the tipping point for geospatial, and the key phrase here is "low headcount."

Part Two: All bets will be off.

Yesterday be damned, all bets will be off. The expectation of non-traditional GIS users is already higher than can be met by traditional planimetric GIS tools. I see this already with drone teams that use better geospatial tools than GIS teams. With companies like Google, Niantic and DJI introducing geospatial technology to the masses, the expectations of digital geography have been raised. This new breed of insurance and risk companies will not be hiring traditional GIS teams. Indeed, with a desire to keep costs and headcount managed I would expect a range of managed services with discrete and actionable tools will be much more attractive.

Systems which natively communicate with each other via sensible, published, but secure APIs will allow risk-management systems to support the needs of different small to mid-sized insurance companies and banks. This will allow them to move very fast and have access to post-event tools to inform the prompt movement of capital. These refined workflows can potentially save tens of millions of dollars/pounds and several stomach ulcers per event**. But remember, geospatial can inform users about risk in various phases of the asset management process, this is not just about post-event analysis.

With this new approach and the possibility of a cloud-native structure, suddenly many new opportunities around the use of data can be realized. This is where our friends in Low Earth Orbit (LEO) might again prick up their ears. One of the biggest barriers to the use of EO has been technology infrastructure. With a modern infrastructure and less need to support ancient Fortran workflows, suddenly, there is funding and technologies to support new data paradigms.

So, all bets are off, what we have been doing will be identified as superfluous, manual and slow.

Part Three: Some biggies are innovating.

While some companies are harbouring the teetering technology stacks, some have been more circumspect and careful with integrating. These organizations are also seeing the benefits of a changing landscape. For this reason, the Mergers and Acquisition (M&A) landscape for risk-oriented geospatial analytics is very fertile (Fathom -> Swiss Re, Geosite -> Descartes.) A modern pattern here, however, is to hold these acquisitions at arm’s length and pull in relevant data via API, thus avoiding the worst of the Jenga effect.

Part Four: Dominos of AI.

Before any AIs, there are people. We have always trained machines, and now these machines are supporting our decisions. This process is being borne out in real-time in the insurance sector. Parametric insurance depends on robust and agreed-upon data sources. However, the granularity of these data sources is becoming increasingly sensitive. The best data, and the best models are being developed and then disputed (Again, a Schumpeterian cycle we should celebrate.) In time, though, as these dominos fall, confidence in the process rises. In traditionally manual processes, a change in the model would be hugely problematic and might incur compliance issues. In the future, we will see more granular models providing more revenue opportunities. By this, I mean that without geographic sophistication, insurance companies had no data with which to parse major climatic landscape changes like increasing wildfires in California, without "good data," all that could be done was to decide to avoid entire political areas, like Counties or even States. With better data, more granularity can be applied to this decision-making, ultimately increasing revenue opportunities. This is particularly true with flood modelling, where higher resolution can indicate various houses that might be far more resilient than others to rising rivers or pouring ponds.

So, these are my central observations on risk. This note does not touch on the financial regulatory environment for ESG or Biodiversity Net Gain, both of which are additional and significant opportunities for the geospatial sector. In no uncertain terms, spatial finance is emerging from the mists of unstructured markets to become a reality.

* I was deeply disappointed this year at SatCamp when I asked if a speaker felt that the customers of the EO sector had changed in the last decade, only to be challenged with them not knowing who the customer was last decade, as they were still in school. I was disappointed not in them but in me, thinking that a decade seems short, when in fact I am just old!

** What is the noun of assemblage for a stomach ulcer? A blister of ulcers?