Analytics

Leveraging a Big Data Approach

Insurers need an enterprisewide data and analytics approach to products and customer service that includes social media.
By: | August 14, 2014

Insurance companies will be able to capitalize new market opportunities and avoid costly exposures when they can better analyze and act on lessons from Big Data.

That’s one of the main findings from two recent industry reports highlighting the need for enterprisewide management of big data, including unstructured data from social media and mobile devices.

Very few insurers have an enterprisewide data and analytics approach. Most focus on targeted business functions such as pricing, underwriting and financial management, according to a survey of 72 P&C insurance professionals by Strategy Meets Action, an insurance strategic advisory firm in Boston.

“But [a siloed approach] is not going to be enough to differentiate and compete in this fast-changing marketplace,” said SMA partner Denise Garth. “Data analytics needs to take an enterprisewide approach that includes external telematics, [and] social and mobile data, so they can really leverage the power of analytics.”

Only about half of P&C insurers report that they have advanced reporting (12 percent enterprisewide and 36 percent in key areas).

Social Media and Mobile

Leveraging unstructured data from social and mobile is particularly important in designing products that customers want, according to a study by IBM’s Institute of Business Value.

Senior executives from 80 insurers surveyed by IBM said they are leveraging the cloud, big data, analytics and social technologies to “leapfrog the competition” in this way.

And nearly three-fourths (72 percent) of the insurers identified by IBM as market leaders in the study said they use social media to communicate with customers “to a considerable degree” — almost twice as much as non-leaders.

“Structures don’t make a lot of sense if insurers are building them in a vacuum — they need to reflect how insurers are targeting certain customer sets.” — Christian Bieck, global insurance leader, IBM Institute of Business Value

Big data analytics should incorporate four dimensions — customers, interactions, services and structure, said Christian Bieck, IBM Institute’s global insurance leader who is based in Stuttgart, Germany.

“The combination of those dimensions is very important, because insurers can only build new products and services in a sensible way if they have insight into what the customer actually wants,” he said.

“Structures don’t make a lot of sense if insurers are building them in a vacuum — they need to reflect how insurers are targeting certain customer sets,” Bieck said.

Strategies like these enable insurers to transition from the more traditional “organization-centric,” product-driven model to one that reflects the emerging “everyone-to-everyone (E2E) economy,” based on higher levels of collaboration between companies and their customers, Bieck said.

Insurers need to bring their historically strong analytical capabilities for predicting exposures to the marketing arena, particularly to cross-sell and up-sell to existing customers, said Sharad Sachdev, a managing director at Accenture in New York.

As part of this, insurers should follow the lead of the banking industry and analyze internal data of past marketing successes as well as competitor data and unstructured data from social and mobile to develop “propensity scores” — to determine which customers are more likely than others to accept certain offers.

“Consumers have many choices, so insurers can’t make offers in a vacuum, and that’s where social media comes into play,” Sachdev said.

Focus on Core Business

John Lucker a principal at Deloitte Consulting LLP in Hartford, Conn., who is the firm’s global advanced analytics and modeling market leader, said that most insurers are still struggling with how to best gain insights from past and current events, and are just beginning to adequately use predictive analytics for future events.

However, he said, “I think emerging technologies and analytics should be more R&D and exploratory, while companies should spend the bulk of their time getting good at the core of their business.”

If an insurer’s underlying organizational structure is not profitable, going after more customers isn’t going to make them more profitable; in fact, it might actually raise their expense ratio and make them less profitable, he said.

“They need to first be really good at pricing and understanding exposures, before they focus on getting more customers,” Lucker said. “I would suggest once an insurer has a combined ratio well below 100, maybe that’s something to talk about.”

The SMA report indicated that P/C insurers will spend more on predictive analytics, with nearly two in five (38 percent) planning budget increases of at least 6 percent per year over the next three years.

But, the bulk of the spend will be on claims recovery, and fraud prevention and detection, with almost half of the respondents piloting projects or planning future investments.

Insurers are beginning to evolve analytics for marketing and distributions, with survey respondents reporting new projects in customer segmentation, “single view” of the customer, and customer “lifetime value.”

Customer segmentation is the top area for new projects over the next three years, with 43 percent of insurers planning efforts in that area, and another 10 percent piloting or evaluating today, according to SMA.

Katie Kuehner-Hebert is a freelance writer based in California. She has more than two decades of journalism experience and expertise in financial writing. She can be reached at [email protected].

More from Risk & Insurance