Customers, Data Stewards Still Wary of Insurers' Intentions

After Admiral Insurance quickly scuttled its attempt to leverage social media data to underwrite policies, the insurance industry appears farther away from being able to leverage certain big data sources and analytics in order to radically alter the customer onboarding process.

The UK insurer's firstcarquote initiative was aimed at younger drivers in the market, who traditionally have paid much higher rates for auto insurance. Previous attempts to get more accurate views of younger drivers have revolved around using telematics devices, and contributed to a small boom in usage-based insurance in the UK over the past few years.

Admiral developed an algorithm that could scan and analyze Facebook posts and refine the much higher quotes young drivers receive without the need to install third-party equipment. However, Facebook rejected the company's app after word spread of its pending launch, citing privacy concerns. Now, firstcarquote will relaunch using only voluntarily data offered by drivers, according to a company spokesperson.

“Following discussions with Facebook, the product is launching with reduced functionality,” the company said, in an e-mail. “Admiral does not have access to customers’ Facebook data and does not hold social media data to set prices for its customers.”

According to Donald Light, director of Celent’s North America property/casualty practice, the appetite simply isn't there to allow social network users to turn over broad swaths of data to insurance companies. Each data point being turned over has to come with an explicit opt-in in order to gain market acceptance, he says.

“Social media as a whole is going to be a fertile field for improving pricing,” he said. “But the opt-in method has to be the main path forward. There is an ethical and public perception that opt-in [social media discount programs] should be okay.”

Despite the headwinds, insurers are preparing for a future where social media data is more accessible, Light says, noting advancements in social fraud detection, particularly in workers’ compensation. But with companies now getting more creative at looking for new kinds of data to make better underwriting decisions, they still must confirm that the data collected is actuarially sound.

“Early movers will price smarter and subject their competition to adverse selection,” Light says. “Rating factors should be justified. Insurers cannot go crazy giving away discounts without having statistical justification.”

In fact, the future may not ever include using social data and other third-party information in the insurance underwriting process, says SMA principal Mark Breading. He says that navigating the regulatory and reputational minefield isn't what insurers should spend their time on. Instead, he suggests, “my advice to insurers has been don’t use the data for rating. Use smart home, telematics and social data for added services, such as location and risk management, which don’t necessarily need regulatory approval.”

Nathan Golia contributed reporting

For reprint and licensing requests for this article, click here.
Business intelligence Workforce management Telematics Analytics Digital distribution Data security
MORE FROM DIGITAL INSURANCE