Striking a Balance Between Big Data and Underwriting Integrity

By Richard McCathron | March 4, 2019

From accountants to advertisers, everyone is talking about the power of big data. Underwriters and insurance agents are no exception.

While big data is undoubtedly an extremely valuable tool when it comes to risk assessment, policy pricing and customer service, it doesn’t need to upend the entire industry. The insurance experience can improve with the use of big data, but it’s important to remember that insurance does effectively protect millions of Americans.

Consider this scenario: your insured is getting ready to close on a home and needs to show a mortgage lender evidence of insurance to finalize the loan. To take out a homeowners insurance policy, they’ll have to answer all sorts of questions about the home. Of course, they didn’t build the house and have never lived in it. Chances are, they’ve only seen it a few times. Yet, almost every insurance carrier will ask how old the home’s roof is before providing a quote.

Even though the material and age of the home’s roof are some of the biggest factors in designing a policy and determining rates, most people provide an approximation. In some cases, they just guess. The same scenario can apply to a range of carrier questions, meaning coverage is often based on inaccurate answers to pertinent questions.

It's a balance between possibility and reality that keeps the average homeowner's premiums affordable while providing a financial safety net in case disaster strikes.

Big Data and Homes

Big data is helping solve that problem by revealing the age of a home’s roof without asking the customer directly. In addition to getting a more reliable answer, this also reduces friction, and that means fewer awkward phone calls between a buyer, their real estate agent and the home’s seller to get information about the home.

Sounds incredible, right? A streamlined application process and a more precise pricing model has helped insurers create innovative products and attract a large base of policyholders. Unfortunately, some companies ignore industry fundamentals in their technology-centric process. They are transparent that their goal is exactly matching a policyholder with their propensity for trouble.

Identifying the likelihood of risk for each customer and charging them accordingly is not how insurance works, though. Insurance is about spreading risk over a large group of people. The idea has long been that everyone pays a little so that no one has to pay too much. It’s a balance between possibility and reality that keeps the average homeowner’s premiums affordable while providing a financial safety net in case disaster strikes. If insurers calculate a personal risk profile with an exact match for what each policyholder will face, the risk is not spread.

Exact match doesn’t work for the customer, either. If someone will have a $10,000 claim each year and their insurer can predict that exposure perfectly, why would they pay the $11,000 premium it would take to cover their claim and the insurance provider’s overhead?

Exact exposure pricing is actually the opposite of insurance. This is part of the reason that several insurtech companies that rely on big data and automated underwriting have experienced unsustainable loss ratios. Worse yet, their customers are treated like policyholders rather than people; they get ushered between bots until their claim is settled, often with a lower payout than they feel their circumstances require. Still, these providers have lost money at the end of the year.

In a spreadsheet, it may be easy to assign a person to a number. In reality, risks are dynamic — they evolve. Data is just one component in the insurance equation. It must be combined with human inputs and insurance expertise to achieve both proper coverage and proper pricing. It’s also important to recognize that data is not infallible. It varies from one source to another. Some sources are more reliable than others. And data needs to be monitored well beyond the new business stage. Policies are not a “one and done” calculation.

At Hippo, we use multiple, verifiable data sources to identify risk throughout the life of the policy. We also blend the data with AI and machine learning to determine which data components are most trustworthy and how to segment certain risks. By embracing new technologies and big data, we have lowered premiums and created a better experience for customers without sacrificing the integrity of our policies.

We’re not just insuring homes, we are insuring families. For many, a home is where children grow and memories form. It’s important for providers to remember that insurtech is made from two words: insurance and technology. Data definitely plays a key role in the process, but it must be integrated properly into an insurance model that spreads risk. Even though there are touchpoints where we can remove friction, accurate underwriting tends to require some human intervention.

Human Touch Still Needed

On top of deciding which sources are trustworthy and ensuring that inputs are accurate, there will be times when a provider needs to ask the customer a question to ensure they are properly covered and the coverage is appropriately priced. Along the same lines, sometimes a seasoned underwriter will see a policy and realize that it needs to be adjusted. Without recognizing that data is one of several inputs, it’s likely that a provider will under- or over-insure their customers at either inadequate or excessive prices.

As I said, big data does present big opportunities for the insurance industry, but too many disruptors are too keen to find an absolute solution. While data is transforming the way we evaluate and underwrite risks, it is not the be all and end all solution. Considering the nature of insurance, it’s unlikely that such a solution exists — or needs to.

In addition to various data and human inputs, we must not forget the input of experience. Since the establishment of the first fire insurance companies more than 300 years ago in London, the industry has learned a few things about risk, data and pricing.

Although there is room to improve and innovate, let’s not forget that there are a few things that we can learn from three centuries of experience.

The insurtech revolution must be based on a balance of technology and underwriting integrity in order to modernize the industry in an effective way.

About Richard McCathron

McCathron, CPCU, CIC, is the chief insurance officer at Hippo Insurance.

Was this article valuable?

Here are more articles you may enjoy.

From This Issue

Insurance Journal West March 4, 2019
March 4, 2019
Insurance Journal West Magazine

Small Business Market; Hospitality Risks Directory; Markets: Homeowners & Auto, Business Interruption; Special Supplement: The Florida Issue