Tom Link is founder of SpatialKey: A data enrichment and geospatial insurance solution that had been incubated within Tom’s professional services company, Universal Mind. Having recently sold SpatialKey to Insurity, a Data Analytics company for insurers, Tom stepped down from being CEO to pursue his next adventure. Before entering the entrepreneurial arena, Tom was in the business of technology consulting at Allaire that was ultimately acquired by Adobe. With a keen interest in insurance and building company culture, Tom had a virtual sit-down with our CEO/CTO David to talk about the benefits of a map-based approach to understanding data, the importance of insurance, and how to wield data company to become a service.
Wikipedia’s Definition: A conceptualized framework that provides the ability to capture and analyse spatial and geographic data.
Tom’s Definition: Essentially it’s a fancy term for map-making, or anything that involves using geography.
It came from my time at Allaire Corporation. Around years 1999-2000 they were looking to create a job that required someone who could jump to different companies and perform “firefighting operations”, such as server stability issues. This required a very niche set of skills and 100% travel, so we created a remote division within the company for this task that worked extremely well.
Consequently, when we started Universal Mind, we hired within that same network and followed what had worked for us previously, regarding remote working, instead with a full start-up company instead of one division within an established company. Having no idea how large the company was going to be when we started meant that remote working happened very naturally for us.
There are always challenges with running a remote company and there were times we thought about converting it to office-based. However, each time we approached another growth phase, we found that there were more benefits than disadvantages in running a remote company.
Back in 2007 we saw a lot of GIS companies trying to figure out how to build richer maps using some of the flash-based technologies. As a professional services company, it got us thinking more about the idea of using just maps to visualise information which is how SpatialKey was born.
But we came up with this idea knowing nothing about the traditional GIS landscape. Once we started to explore the landscape we found that there was a real opportunity to start from scratch for two reasons.
Firstly, the current landscape was very desktop-based. The products were then web-enabled in some way. What we found was that people wanted to throw large amounts of data at these things and even as desktop systems, the upper limits for data handling was often in the tens of thousands of records which becomes even less reliable when trying to web enable it.
Secondly, both desktop-based and web enabled products lacked the user experience that from our perspective was extremely valuable because there are GIS questions people want answered who do not have a GIS degree.
We learnt that most data has a geographic context therefore can be linked with a little location data. Whether its real estate transactions, crime records, or insurance policies, they almost always occur in a place and we saw a opportunity to see that “place” as a canvas to be able to understand information. It was clear that there should be a tool where someone dealing with tens of thousands or millions of records should be able to easily import those records and be able to start sifting through them and gaining insights.
Absolutely. Most of the time we think of maps as an output that distributes information. But very early on we saw the value in using it as an input as well, as once you have looked at a map you want to dig in further and those explorations often lead to more questions. Usually in the more traditional GIS model that would require some sophisticated expert to go back to the GIS system who knows how to operate that, and often that round trip time of getting that response could be days. With our approach of a more self-service environment, that could be minutes or seconds instead.
Yes, although we have found that getting the essentials in place first is really important before inundating a project with a whole range of data sources.
Very often companies couldn’t see the plain raw data that they were accessing on a daily basis in more regulated formats, let alone the additional universe of data that could be available for them. Rather than going in all sophisticated and boiling the ocean to bring the universe of data together, we found it was better to go in and start with unlocking their core data first so they can see it in a new light whether that be generating different kinds of reports or viewing trends in a more geographical way.
It depends on the project. We started out being very focused on building a product company and not driving revenue from services.
The core issue we had with consultancy was that the work was not always repeatable. Most of our consultancy projects came from people who had the most unusual business cases and were therefore very interested in keeping them proprietary. However consultancy made it easier for us to solve the harder problems such as importing millions of records, geo-coding and other issues, as they could be dealt with as a one-off consultancy operation.
However, we really wanted to solve those core problems that made way for you to interrogate data geographically and do it in a systematic way that was not a consulting offering, which is what makes a more service-based approach appealing.
Back in 2007 it was an enormous challenge, but we were aware that we were in this line of work earlier than most people and that over the following decade it would get easier.
One of biggest challenges was that the only people who were interested in this type of service were the ones who were highly specialized in it. That meant they had their own internal teams that either wanted to do the work themselves, or those organisations were very proprietary and wanted to keep everything as private as possible, which meant it became hard to repeat.
However when you compare the data landscape now to what it was in 2007, the appreciation for needing to work with larger datasets has definitely increased. You now have insurance companies with innovation departments who are realising that you need to be making these kinds of investments, and you need to turn to innovative software firms with a track record of success rather than figuring implementation out on your own.
Most sectors understand the benefits, the challenge lies in knowing how to sell the service to them and addressing the challenges of bureaucratically what keeps them from being successful with data. Otherwise they remained generally averse to technology and data.
Law enforcement has always been the most challenging as we never really learned how to sell to state and local government and build solid partnerships. What is more, they tend to be afraid of technology in all directions despite the clear benefit a cop could get from this kind of data service. But this is changing over time and we have seen that through our work with insurance companies as, whilst they are slow to adopt technology, they are slow to change services once they obtain them which allows you to build a solid relationship.
Insurance became the most interesting space for us and it became really important to me that I was delivering a clear purpose to our employees that they understand how insurance gives us a more sustainable and resolute world. When we look around, whether it be in a city centre or a place, that none of it is possible without insurance. Communicating this complex ecosystem and understanding the associated risks was why we developed such as sweet spot for insurance.
By illuminating the inefficiencies you find in insurance, whether it be improperly priced policies or insuring locations which were higher risk than first thought, you are providing a service that brings better data to a space and allows better decisions to be made quicker, which is always exciting for us.
Its always worth being aware that some things you must learn the hard way and starting in 2007, we had to spend a lot of time telling people the value of working with data. When we started working with insurance companies, which wasn’t until about 2010-2012, each company believed that everyone was doing the service poorly and that their company was the status quo which was good enough. I think if we had known this attitude, we would have invested in more marketing that drove home how bad the status quo was and what it was costing everybody, a lot earlier.
The fact that as a result we are building a safer, more resilient society and driving down costs. On top of that I am convinced it is possible to do all these things together, and understanding the data we have is a crucial and promising part of that.
In the insurance space you see this first-hand as this kind of data collection allows us to make better decisions about the investments we make that to reduce costs and risk for all of us. For insurance that means saved lives and reduced time frames to get things done. All of this comes from understanding the data that is right in front of us. We are actively seeing this today and knowing that this is just the tip of the iceberg is really exciting.