There are a lot of buzzwords circling around lately like “The Internet of Things” and “Big Data”. The trouble is that many don’t really understand how it all works. The insurance industry seems jazzed up about big data and the need for data scientists. The question that is rarely spoken however, where is all that data going to come from?
In an earlier post we spoke of the coming tech disruption of insurance and gave the example of Metromile, a company offering pay-per-mile driving insurance. We hinted at the idea that “given the right data, one could readily track your driving style to understand your risk profile in a much more intimate way”. If you can readily identify the risk profile of your drivers, you can more accurately adapt your rates and your costs to it. This is where The Internet of Things and Big Data come in. How do you collect that data, parse it and analyse it to be able to make decisions based upon it?
Let’s start with a sample business case. You decide you want to write insurance for logistics firms. As a means to undercut your competition you decide to leverage technology in the form of in car analytics. You have the grand idea that if you could readily track details about drivers and their trucks you could analyze them to understand their risk profiles and have a better idea of the risks of the business. This allows you to cut your premiums as well as your costs because you can readily identify deal with problem drivers or problem vehicles. The idea part is easy, the question is, how do you do it?
Metromile is using a phone app. That’s a great place to start. Going further would be to tap into the standardized on-board diagnostics (OBD) port available in most every car and truck these days. It is what mechanics use to diagnose problems with your car. The technology around this is maturing as there are a growing range of products and apps which leverage this port to allow you to track many data points like car health and the way the car operates. Combined with the right electronics and the right applications one could readily build a very comprehensive data profile of both the state of the driver and that of the car. The implications are profound as suddenly you can have keen insights into the risk of insuring our sample case of a fleet of logistics drivers. The problem is collecting and analyzing that amount of data isn’t exactly trivial.
The idea seems simple, the diagnostics app and device is the “Thing” and somehow you need to connect it to the internet. This “connect it to the internet” bit of plumbing is what few people talk about. How do you collect information from thousands or even millions of devices that could all communicate at random times, possibly all at once? This is the big missing gap that we’re getting closer to crossing. On-demand or “Cloud” computing is making it easier to build cost effective, scalable, cloud based, software architectures that are able to process incredible amounts of data. This is why people in the tech community are jazzed up about the internet of things, we’re getting very close to the tipping point where we can actually build these sorts of solutions.