More data, but do you really need it?

If you had something that didn’t just unlock your organisation’s potential, but helped it grow the more you used it, you’d want more of that something, wouldn’t you? That’s what data means to insurers. It is the “secret sauce” that helps underwriters access new and better sources of revenue.

But you can have too much of a good thing. An obsession with gathering data at the expense of considering its origin or quality can have insurers chasing their tails—building in more inefficiency, not less. At Underwriting Innovation USA on November 10, 2021, Dan Tatro, director of insurance practice at Precisely, will reveal to attendees how they can unearth a gold mine of undervalued internal data and supercharge their underwriting function.

Why do insurers keep searching for yet more data?

Insurance is the ultimate numbers game. It’s everywhere you look—in policies, claims, reserve, reinsurance and so on. Margins are minuscule, so it’s essential that insurers wrangle every one-hundredth of a percent from revenue. Data-driven decision-making impacts so many facets of the industry, from marketing to binding to risk prediction, and it’s understandable that marketers would want to go after every source available. Put simply, the more contextual the comprehensive data, the smarter the models.

It’s not necessarily a mistake to go after more data. Data scientists are voracious creatures. They want to examine every element to find out if it’s statistically relevant to inform those models.

Where should they be focusing their attention?

The most important elements for any data-science effort are real-world observations. Without them, we wouldn’t know the outcomes to predict or the behaviour to mimic. For insurers, this might mean campaigns, web traffic or converted quotes. It certainly includes the book of business—and not just the current policies-in-force, but all policies and claims.

Third-party information on risk is completely in play but, at some point, you must go back to what is most important. Historical data is crucial but there is a bit of nuance to that term.

We’re seeing the time frame getting shorter and shorter. It can mean “just before now”—whether that’s one second or one decade. But human behaviour—geofencing your phone if you are experiencing a rainstorm or flooding, for example—can mean that we should be looking again at coverage.

Why have insurers typically not mined their existing data?

Sometimes you don’t look in the spare room or junk drawer because you don’t want to acknowledge the chaos. Unfortunately, growth can lead to chaos. Data is siloed or forgotten over time, whether organically or through acquisition. Integrating data across dozens of systems or data stores can sometimes seem impossible.

What can underwriters do with this data that they couldn’t before?

With determination and investment, insurers can be confident in the integrity of their data. From iron to cloud, solutions for data integration, data quality, single-view master data management and enrichment are better than ever before. Additionally, modern cloud platforms such as Databricks provide the scale and speed to turn what were historically analytical calculations, such as pricing, risk and accumulation, into real-time determinations.

What challenges might attendees bring to this session, and what do you hope they will take away?

They may potentially come to the table with lots of different questions. How do we leverage the data we have in-house? Do we go with siloed information, whether it’s good or bad, in legacy systems and learn something from them? How do we gain knowledge and reach that single pane of glass that is often in dozens of systems, from customer relationship management to enterprise resource planning, marketing, and campaigns?

The conversation is going to be about breaking down and solving those siloes and data integration issues. At Precisely, we have the concept of data integrity—gathering it from the mainframe to cloud data warehouses, making appropriate integrations and mastering the single view.

That effort might seem very daunting but, as they say, you don’t have to eat the elephant all in one go. Take the first step where it’s the safest. That might just be data integration—simply getting it into a place where it can move the needle a little. Then, profiling that data to check it’s fit for use or, if it’s close, moving on to managing data quality and ending up with a single view.

I can’t think of a single insurer who isn’t going through this process now. It’s table stakes.

At Underwriting Innovation USA on November 10 at 11:45am, Dan Tatro, director of insurance practice at Precisely, will reveal to attendees how they can unearth that gold mine of undervalued internal data and gain the knowledge they need to succeed.

This is the biggest online event tailored specifically to underwriting—don’t miss out. Register now!