It's hard to predict the future without a crystal ball. In terms of what insurance carriers do, when predicting how things will turn out the data is their crystal ball.
Those with good data can turn it into actionable decisions and those without good data are doomed to poor decisions, according to Merit Smith, vice president of Robert E. Nolan Co.
"I think divisional-level operations and IT staffs probably have a more realistic sense of data quality issues and how well their data would facilitate advanced analytics," he says. "Generally, before you can do anything you really need to tend to your data, clean it up, and get it in properly structured models. So there is better understanding of that need closer to the operation level than at the corporate level."
(Smith is featured on a new podcast with Tech Decisions editor-in-chief Bob Hyle as they discuss data and predictive analytics. To hear their conversation, click here.)
The process for data issues begins with those insurers who don't have much awareness of the quality of their data, according to Smith. The next step is to become aware of the problems associated with a lack of data quality.
This is followed by the period where companies focus on cleaning their data, continues Smith. "Then they evolve to what I think we are all interested in–being able to use high quality data as part of their strategic solution," he says.
Some data collected by insurers is naturally cleaner than other data, points out Smith.
"For example, demographic data–who, where when, how much–are probably cleaner and will yield higher initial benefits in analytics," he says. "The more complicated uses, such as integrating data across several product lines for cross-selling purposes are much more sensitive to data quality problems and frequently have a lower impact because of their sensitivity to it."
© Arc, All Rights Reserved. Request academic re-use from www.copyright.com. All other uses, submit a request to TMSalesOperations@arc-network.com. For more information visit Asset & Logo Licensing.