Wise Practitioner – Predictive Analytics Interview Series:  David Schey of Digitas

By: Eric Siegel, Founder, Predictive Analytics World

In anticipation of his upcoming conference presentation, Uplift Modeling Versus Traditional Response Modeling – Which One is Right for You?, at Predictive Analytics World San Francisco, March 29-April 2, 2015, we asked David Schey, Senior Director, Analytic Consulting Group, Digitas, a few questions about his work in predictive analytics.

Q:  In your work with predictive analytics, what behavior do your models predict?

A:  We model a wide variety of behaviors including response, attribution, spending, attrition, ‘clone’, customer value, market basket, fraudulent activity, next most logical product/product lifecycle, insurance claims, and a variety of credit evaluation dimensions.

Q:  How does predictive analytics deliver value at your organization? What is one specific way in which it actively drives decisions?

A:  Predictive analytics delivers a superior experience to our client’s customers. By more actively engaging with customers, we optimize the customer experience, increase retention rates, and enhance spending patterns. Our models help managers deliver the right message to the right customers through the right medium. Contact decisions are frequently made based solely on some of the models referred to above.

Q:  Can you describe a successful result, such as the predictive lift of your model or the ROI of an analytics initiative?

A:  A major retailer found that control groups were performing in line with mailed groups. Although response models were used, they did not provide the incremental lift that the retailer was expecting. Out uplift model resulted in identifying 20% of the mailed population that provided 16.5 basis point lift, a result that the retailer had never experienced.

Q: What surprising discovery have you unearthed in your data?

A: The primary lesson is not to take data at face value! During one exercise for an auto manufacturer, we found that 39% of our clients’ customer base owned 9 cars! Seems outlandish, and it was. The value ‘9’ was used as missing value indicator. If it’s too good to be true-it isn’t.

Q:  Sneak preview: Please tell us a take-away that you will provide during your talk at Predictive Analytics World.

A:  There’s more than one way to skin a cat. While there are several approaches for developing uplift models, not all of them provide satisfactory results all the time.  Also, uplift models provide a completely new lens by which direct marketers will increase the sophistication of their contact strategy.  The concept of not "waking a sleeping dog" (i.e. customers that are negatively influenced by communications) is often foreign to traditional direct marketing principals…not only will you increase profit by suppressing certain customers, but those funds can be reallocated to new communications for even greater impact! 

Don't miss David Schey’s conference presentation, Uplift Modeling Versus Traditional Response Modeling – Which One is Right for You?, at Predictive Analytics World San Francisco, on Tuesday, March 31, 2015, 3:55-4:40 pm.  Click here to register for attendance. 

By: Eric Siegel, Founder, Predictive Analytics World

Eric Siegel, Ph.D., founder of Predictive Analytics World and Text Analytics World, and Executive Editor of the Predictive Analytics Times, makes the how and why of predictive analytics understandable and captivating. He is the author of the bestselling, award-winning Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die, a former Columbia University professor, and a renowned speaker, educator, and leader in the field.