Risks are tricky things to get to grips with. How you identify them, analyse them and then draw conclusions from the results are processes permeated with interpretation and subjectivity (a few decades ago, I spent a year studying just that for a Masters degree). This shouldn’t be a surprise to the insurance market, for if it were not the case, there wouldn’t be a market in the first place.
So when I hear the claims currently being made for telematics, I can’t help but apply some caution about what can be drawn from all that new risk data. The promises being made for telematics sound stupendous. The insight on offer seems awe inspiring. The impact on motor insurance will, to quote one presenter at a Post Events conference this week, be like a “sledgehammer”.
All these things will come to pass only if our ability to handle this veritable avalanche of data improves beyond measure. To achieve those new heights, insurers would need to overcome the pitfalls encountered by the legions of analysts employed by investment banks around the world to sift through financial data in pursuit of profits beyond measure. That would indeed be a challenge, especially given the painful lessons the economy has had to learn since 2008.
Drawing out an understanding from past data is difficult enough, as an astrophysicist in the family will testify after several years spent running unbelievable amounts of observatory data through one of the UK’s principle supercomputers. Drawing inferences from past data about future behaviour is ever more of a challenge, as any investment bank CEO will confirm. It takes more than a fine software algorythm to draw insight from data.
Greater amounts of data can certainly lead of greater levels of insight, but only up to a point – it’s not a relationship that survives being scaled up indefinitely. Eventually insight stalls and further data muddies the water. Consolidation in how data is used then takes place until another revolution comes along and jolts us into new ways of gaining insight. It’s been the way of scientific progress for many a century.
There’s also the risk that you begin to rely too much on what you think your data is telling you. Data will not tell you what happened. It will allow you to draw better inferences about what happened, whereupon you’re then able to hold a better conversation with those who actually experienced what happened. So while insurers have much to gain from telematics, they need to remember that it is less likely to provide them with an avalanche of answers and more likely to equip them to ask better questions.
