Tom Warden, SVP, Chief Data and Analytics Officer, EMPLOYERS
I was asked to write on how analytics has transformed the insurance industry. But the more I thought about the premise, the more it seemed backwards to me. You see, insurance IS analytics. It is better if we view the needs of insurance as the engine that drives the evolution in analytics. The first serious use of analytics in it dates to 1693 when Edmund Halley created the first statistical mortality tables. Spoiler alert: he didn’t use machine learning in the cloud to do it.
Why is this important? Because I believe we tend to get caught up in the heat of the moment when advances in technology and data grow at a fervent rate. We are certainly in one of those moments right now…and it’s mostly good. I increasingly hear and read about examples of data and analytics creating value for insurance companies and their customers. What cannot be so good at times like this, based on my experience of having lived through similar eras, is that the analytics “tail” starts to wag the insurance “dog”. When this happens, enthusiasm becomes pervasive, investment in analytics expands rapidly, C-suite interest and expectations rise beyond attainable levels and the cycle collapses…but hopefully only for a short time.
I started my insurance career in 1987. Mainframes and magnetic tape ruled the day. Oh, yeah…and IT departments controlled access to both. Data, the DNA of analytics, was pretty much locked up in policy and claims admin systems. I worked at an insurance research center and there was no lack of ideas for new applications of analytics there. But the computers and data were not widely available to execute them. Fortunately, the PC revolution and the advent of the minicomputer were soon at hand. The analytically inclined no longer had to fight for computer time, all we needed was data. Data began to flow, both from company data centers and external providers and we were off to the races.
Pricing is the insurance function where analytics has had the greatest impact. It’s purely quantitative, is abundant with complex data relationships and is the primary driver of a company’s profitability. It took a while, but once computational power and data management were made available to actuaries and underwriters Halley’s tables became more sophisticated and ultimately somewhat irrelevant.
It is better if we view the needs of insurance as the engine that drives the evolution in analytics
Over the last twenty or so years we have witnessed a handful of companies across each insurance line earn excess returns from becoming leaders in the sophistication of their pricing strategies and algorithms. Much of the rest of the pack has been racing to catch up while some have yet to join the race. The weak members of the herd are likely headed toward extinction. Eventually though, pricing will become a more even playing field as the sophistication of algorithms and creative use of data bump up against the tolerances regulators have for innovation in this space.
Risk Management is a close second to pricing with respect to where analytics has had the greatest impact. It is a bit harder to see the direct impact modern risk management has had on companies’ bottom lines, but perhaps the best indicator of the power of analytics is that fewer insurance companies are going bankrupt than in the past. The availability of vast amounts of data characterizing risk attributes has helped companies understand the risks lying below the surface of the numbers on the balance sheet. Massive computing power, advanced algorithms and complex simulation models allow companies to get the best look into the future they have ever had. But is it enough? Warren Buffett once warned, “Beware of geeks bearing formulas.”Will, once again, management’s increasing confidence in their risk management models encourage them to assume risks up to, and beyond, the tipping point of prudence? I hope not…but history is littered with corpses of companies who didn’t learn from the past.
Claims is increasingly seeing the benefits of analytics. Efficiencies in how claims are processed, more accurate determination of benefits necessary to properly settle a claim and more sophisticated identification of fraud all are driving loss and settlement costs lower. Business rules, predictive models and the mechanisms to standardize how they are applied across repeated processes are generally in the early stages of evolution in much of the industry. If your company is not investing heavily here, you should champion the cause. Good claims adjusters operate to a large degree as “human machines”, applying well-vetted processes and decision logic every transaction they process. Computers equipped with proper algorithms can do the same thing. Managing the transition while keeping the adjusters engaged in change is the greatest challenge in this arena.
I would be remiss not to mention Marketing in an article about insurance and analytics. Progressive has increased their advertising spend from $100 million in 2011 to $1.2 billion in 2018. GEICO has grown their ad spend equally aggressively. Both companies are analytics leaders among leaders and are on record for saying they will continue to increase their spend as long as they see positive incremental returns on their investment. So the next time you feel a bit high-browed about how intellectual and geeky insurance has become, just remember that not only has all of this advanced analytics brought us sophisticated pricing, it is also partly responsible for Flo, 24x7.
In summary, analytics is a key ingredient to insurance innovation. If you are on the “buy side” of the industry be careful of getting caught up in the hype. Focus on investments that have a clear path to realized-value and your praises will be sung from the board room. If you are on the “sell side”, please don’t tell CEOs that any of this is magic and that to realize it all they have to do is buy what you are selling. Okay?