How to Make Big Data Smarter
Insurance Experts' Forum, December 26, 2013
A few months back, I heard that since the phrase “big data” was getting shopworn (and a bad reputation on top of that), some analysts were upgrading the term to “smart data.” It does make sense that all big data is actually pretty dumb, and all it really amounts to is a big pile of data that is mainly useless to decision makers. But how does introducing something called smart data suggest we're getting more than just another marketing term?
Jeanne Ross, director and principal research scientist at MIT’s Center for Information Systems Research, proposes a set of actions companies can take to turn big data into smart data. In an interview posted at MIT Sloan Management Review by Michael Fitzgerald (contributing editor to SMR, not to be confused with Celent's own Michael Fitzgerald), she also points to two leading insurance companies as examples of how to make this transition.
After all, even with all the nice open-source tools and platforms that come with big data, it's still expensive to maintain. Plus, Ross says that there are vastly multiplied security risks.
Ross has the following recommendations for adding more smarts to big data. This is actually tried-and-true advice advocated for years in the IT and data management space, but it takes on even greater urgency as petabytes of data flow through organizations:
Establish a single source of truth. At Aetna, leaders of business units produced spreadsheets that showed their divisions were profitable. By bringing this data into a central repository, the business was better able to understand which divisions were leaders, and which were lagging.
Use scorecards. Ross relates how Aetna's IT department developed a scorecard that showed daily performance. While the data was substandard at first, the act of measuring prompted department leaders to start developing better data.
Create ownership of business rules. Ross cites the work of Allstate Insurance Co., which originally “had a rule that it must wait 30 days to pay out a policy when a car was stolen.” However, this was causing ill will among customers, particularly those is some parts of the country in which stolen cars tend to not get retrieved. The company cut the check waiting time to 24 hours in some parts of the United States, which required ownership of the associated business rules.
Cultivate talent. An important piece of developing smart data out of big data is having a workforce that knows how to work with the data. In order to do this, coaching and training is key.
Joe McKendrick is an author, consultant, blogger and frequent INN contributor specializing in information technology.
Readers are encouraged to respond to Joe using the “Add Your Comments” box below. He can also be reached at email@example.com.
This blog was exclusively written for Insurance Networking News. It may not be reposted or reused without permission from Insurance Networking News.
The opinions of bloggers on www.insurancenetworking.com do not necessarily reflect those of Insurance Networking News.
Add Your Comments...
If you have already registered to Insurance Networking News, please use the form below to login. When completed you will immeditely be directed to post a comment.
You must be registered to post a comment. Click here to register.