Using Big Data To Improve Homeowners Insurance Products

There is no doubt that so-called Big Data is changing the way many companies do business. The volume of information being generated by web activity is growing exponentially and the technology has finally emerged to manage this astounding amount of data.

A few short years ago, the most coveted job description was Java engineer. Now that has been replaced by database engineers and math and statistics wonks. As always, the primary challenge with Big Data is separating the signal from the noise to find those nuggets of intelligence that provide competitive advantage.

After finding the value, the next challenge is making the intelligence actionable. Sometimes insight requires change that may be challenging to implement. Managers whose headcount or prestige is at stake can become roadblocks to change when their cheese is moved.

At fast growing entrepreneurial companies, the task is easier because continuous improvement is the mantra. But with larger, slower growth companies, that task can be more difficult. Leaders must be skilled in measuring ROI and steadfast in implementing the required change because there is probably a competitor who is seeing similar intelligence.

For decades, the typical Coverage C portion of a homeowners policy (personal property) has historically been a straight percentage (usually 50 percent) of the Coverage A (the re-building cost of your house). In other words, insurance carriers represent that the replacement value of your personal contents is related to the cost of lumber and concrete.

What this means is that identical homes in a given neighborhood, say one with a retired couple and one with a married couple and four children, require the same contents coverage. Similarly, a home in Santa Monica, Calif., will require two-to-three times more contents coverage than the identical home in Cleveland.

Our conversations with insurance carriers indicate that they are aware of this anomaly but there has never been data available to do anything different.

Our analysis of billions of dollars in settled claims in our data warehouse confirms that there is no correlation between Coverage A and C. In fact, our research demonstrates that the amount of contents coverage needed has a lot more to do with the demographics of who’s inside the house rather than the house itself.

This insight drove us to build an analytics product that enables carriers to accurately predict the required contents coverage with a limited amount of policyholder information such as insured marital status, age and gender of household members.

Yet even with all the benefits that come with improving the core mission of insuring to value (ITV), some carriers view the investment in their old way of doing business as too daunting to change. While some of the challenges of integrating new predictive analytics into their business models are not trivial, history shows that eventually the data-driven insight wins as more nimble competitors take advantage.

A case in point is how the use of credit data transformed automobile insurance underwriting. Instead of focusing on the repair and replacement cost of the car, 20 years ago, a few companies like Progressive and GEICO realized that who was driving the car was much more predictive of claim costs than the car itself.

That insight vaulted these companies to dominant market positions and changed the rules of the game for all other insurance carriers.

Theft Claim Data Suggests A New Approach Is Needed

Theft claims are usually the most common peril for homeowners.

More than a decade ago, consulting companies marketed data-driven strategies to help carriers reduce expenses and improve cycle-times for high frequency, low severity claims. The idea was to fast-track these losses in call centers staffed with lower skilled, usually junior adjusters. Carriers would realize cost savings while improving customer service with faster settlements.
As carriers have adopted this “fast-track” strategy for theft, the indemnity of theft claims has steadily risen to $6,200 on average, more than twice the average auto physical damage claim and bumping up against the average structural claim.

Digging further into our data, we also found that theft frequency was increasing while the number of reported burglary and larcenies reported to authorities was actually decreasing. Not surprisingly, theft frequency spikes also occur during holiday and tax seasons, correlating with the perceived need for additional cash.

All of this suggests that carriers need to think differently about theft and its relationship to homeowners products. Maybe certain types of theft claims should actually be inspected.

Most contents purchases are made with credit cards and all have online reporting going back several years. Potentially, carriers should require credit card statements as proof of ownership or otherwise be obligated to pay only actual cash value.

Of course, these types of actions require claim process and policy changes—and we have already pointed out that analytics may prove to be inconvenient truths to claims and underwriting organizations.

If homeowners insurance is in a similar position as auto was 20 years ago, the winners and losers will be defined by how well they manage Big Data and secondly, by how well they are able to make the process and product changes Big Data suggests.

History shows that there will be a few leaders emerging who change the rules of the game. These forward-looking companies will embrace the change, and position themselves to become the big winners of tomorrow.

Jim Fini is vice president, research and development and founder of Enservio, a provider of software and services for P&C carriers to price policies correctly and settle contents claims with more accuracy. Reach Jim at jfini@enservio.com or www.enservio.com.