PHILADELPHIA—Insurance companies have never been short on data. They have always collected great quantities of it, ranging from home owner addresses, to the crash histories of drivers, to the slip-and-fall claims and property losses of commercial policyholders down to their individual stores, factories and offices. Actuaries, the life-blood profession (and butt of many jokes) in the insurance world, were the “original data scientists,” as Sunil Soares, founder and managing partner of consultancy Information Asset LLC, described them during the 2013 Insurance Data Management Association (IDMA) Conference here on April 15.
“We have been taking in volumes and volumes of data for years. We just never called it big data before,” said Lisa O’Rourke, data management section manager at Lincoln, R.I.-based Amica Mutual Insurance Co.
What insurers are in need of, however, are data standards and governance structures to leverage their data, according to Soares and other panelists at the event. They realize that data governance—ensuring the integrity of the data and that data management best practices are met—is essential; their data must be applied strategically across the enterprise. The old siloes between data producers and data consumers must be broken down.
Explained Soares, a focused data governance initiative can lead to benefits across an insurer’s operations, from enhanced data from agent networks and claims adjusters, to better-informed actuarial decisions and more certain compliance with capital adequacy regulations such as the European Union’s Solvency II. Data governance programs can lead to meaningful cost savings, including more efficient customer communications and reduced IT costs, Soares said.
Clarity for Customer Profiles
One issue that sound data governance can solve for insurance companies is defining their customers.
Amica senior management tasked Susan Haney, assistant vice president of information systems, with devising an “enterprise customer profile” in 2010. She told IDMA attendees she knew it would not be easy. Data was stored in disparate places. Multiple sources existed for the same data. Which source was valid? Amica’s various data storage systems, across different business units, needed to be standardized, work which included devising a common terminology across business divisions.
The initiative held implications for Amica’s core strategic mission: to provide peace of mind for policyholders and create long-lasting relationships with customers, said O’Rourke.
“Poor data quality can impact your customers’ experience,” she explained. Improved data, on the other hand, could lead to decreased need to ask customers for contact information that already exists, personalized customer service by improving customer preferences, and reduced bounced emails and duplicate mailings.
Haney said that support from senior management was essential. “Without their support, there was no way we were going to succeed,” she said.
Haney said that her group is now emerging from the “long and winding road” that was the early years of Amica’s data governance initiative, involving weekly meetings with representatives from business units and education to get everyone on the same page. She kids that her staff played the role of “Switzerland” — with her as “Geneva” — in that they were neutral and allowed each business unit to find its way, define their customers and name which data points were important for customer profiles.
By the fall of 2011, Haney and her staff had produced a list of short-term and longer-term goals. “We actually got their approval right on the spot,” she says.
Amica has since established a data governance council, with all major business units represented, and data governance working groups consisting of lower-level data stewards. They also hired a data governance leader, an in-house recruit who currently is working in this position part time. They have begun analysis of email addresses, decided on data quality tool and will be starting data profiling soon.
O’Rourke said she looks ahead at Amica to even-greater amounts of data that can be associated with the true (and not duplicate) customer file: telematic data from auto insurance policyholders’ vehicles, unstructured loss data such as claims adjusters’ written notes, social media and customer surveys – more fodder for the original data scientists.
Path to Process Improvement
For Maria Snow, data operations executive at the National Council on Compensation Insurance (NCCI) and a conference speaker, it was essential to implement data governance through the Boca Raton, Fla.-based insurance research firm’s accepted quality principles for process improvements.
Similar to Six Sigma strategies for process improvement, Snow said these quality principles can be encapsulated in four steps: create, deploy, learn and integrate. With data governance, they took the form of: (1) devising a framework of governance and piloting it in 2008; (2) implementing it for all data management processes in 2010; (3) full staff review in 2011, followed by full management review in 2012; (3) and integration across business units in 2013.
Matthew Brodsky is a freelance writer based in Philadelphia. Follow him on Twitter: @MatthewLBrodsky.
Correction, May 1, 2013: The is article has been updated to correct the spelling of Susan Haney’s last name.