Analytics is the new frontier for insurance – exciting but dangerous

  • 14 July 2016

When investment returns look uncertain and market conditions are challenging, executives often listen that bit more keenly to new ideas about identifying and exploiting market opportunities. This is just how a competitive market should behave, isn’t it? And that’s right, but perhaps not in such simple terms.

A recent report by one of the big four audit and advisory firms noted that “93% of insurance CEOs … see data mining and analysis as more strategically important for their business than any other digital technology”. You can therefore be pretty sure where those CEOs will be looking to reduce their combined ratios by those precious percentage points.

That same report presented a neat graph about ‘the new frontier of analytics’. It showed four levels of analytics:

  • descriptive analytics – what happened?
  • diagnostic analytics – why did it happen?
  • predictive analytics – what will happen?
  • prescriptive analytics – judging what should happen and making it happen.

It is in such a nexus – the problem, the insight, the solution – that the fibre of a firm’s ethical leadership is measured. Is it skin deep or nearer the heart? Fine words or a decision making lens? As you’ve heard me say on several occasions, most misconduct in corporate settings is not down to bad apples, but down to good people making bad decisions. And then rationalising them away, through phrases such as ‘everyone else is doing it’, or ‘the business needed this from me’.

There’s nothing wrong with analytics of course. It’s what you decide to do with it that is the measure of your integrity, of your firm’s integrity. And as you drop down that list from descriptive analytics to prescriptive analytics, those judgements become more pertinent.  Take prescriptive analytics for example. If your analytics team shows you how they can identify the ‘what, when and why’ of how people make certain decisions about buying insurance, and then how they can nudge people towards those circumstances, then those precious percentage points of combined ratio might seem that bit nearer to reality. After all, analytics is often presented as ‘understanding our needs, sometimes even before we do’.

And it is in such circumstances where some critical thinking is most needed. Prescriptive analytics may add up in one direction (the business case, as presented by big consultancies), but not everyone sees it that way. Turn it around and it can look like what I’ve previously called ‘manufactured vulnerability’, a practice about which even the most hardened of marketers can have doubts.

Those 93% of insurance CEOs were right to see data mining and analysis as more strategically important than any other digital technology. What seems to be missing is aligning the strategic importance of digital technology alongside other strategic aims such as trust and integrity. The answer of course is that they’re both important, but at the same time, that they’re not separate initiatives. Digital can build trust through better engagement; trust can build digital by bringing customers closer.

Treat prescriptive analytics with caution. In some insurance scenarios (such as commercial claims), it could work wonders. In other scenarios, it could wreck reputations. Double edged swords are meant to be handled with care.