Related Links

A couple of years back, I prepared a report on behalf of Insurance Networking News that discussed the enormous cost-saving potential that comes out of applying predictive analytics to fraud detection. At that time, I noted, by integrating predictive analytic results into daily claims management activities, insurers can automate preventative measures as well as act immediately on fraudulent claims. Carriers only have a set amount of time in which they are mandated to cut a check, and applying predictive modeling technologies can help catch discrepancies within that short window. Such tools are not only run against the initial claim documents, but also against all transactions through the claim lifecycle—from initial filing to post-settlement.

Nowadays, the power of analytics--both predictive and historical--is sought for a range of business applications well-beyond fraud detection. And the benefits can add up to much more than cost savings. Analytics is now a way to find, engage and retain customers, understand new market opportunities, and understand product performance.

The way to get there is outlined by Dominic Barton and David Court, both from McKinsey and Company, in the latest issue of Harvard Business Review. Here are their guidelines on key steps needed to truly become an analytics-driven business:

1. Choose the right data. There is a lot of data flooding into organizations—the challenge is to identify and pick out those nuggets that are valuable to the business problem at hand.

2. Source data creatively. “Managers also need to get creative about the potential of external and new sources of data,” Barton and Court write. “One way to prompt broader thinking about potential data is to ask, 'What decisions could we make if we had all the information we need?'”

3. Get the necessary I.T. support. “Legacy I.T. structures may hinder new types of data sourcing, storage, and analysis,” Barton and Court caution. Business and I.T. need to work together to make analytics projects a priority.

4. Build models that predict and optimize business outcomes. Analytical models are essential to help managers predict and optimize outcomes, Barton and Court say. “More important, the most effective approach to building a model rarely starts with the data; instead it originates with identifying the business opportunity and determining how the model can improve performance.” Also, keep the models as simple as possible—“statistics experts sometimes design models that are too complex to be practical,” they caution.

5. Build trust. The path to an analytics-driven organization is through trust the Big Data-based models that are developed, the authors say. In one case, “front-line marketers who made key decisions on ad spending didn’t believe the model’s results and had little familiarity with how it worked.” New approaches need to “align with how companies actually arrive at decisions,” they add.

6. Embed analytics into simple tools for the front lines. “The key is to separate the statistics experts and software developers from the managers who use the data-driven insights,” Barton and Court state.

For example, one approach they recommend is to, rather than providing managers with reams of data and complex models, create a simple visual interface that highlights need-to-know information.

7. Develop staff skills to exploit Big Data. That means education and training for managers and employees, to help everyone on staff to upgrade their analytical skills and literacy.

Joe McKendrick is an author, consultant, blogger and frequent INN contributor specializing in information technology. He can be reached at joe@mckendrickresearch.com. This article first appeared on Insurance Networking News, a sister publication of Health Data Management.