The Wharton Customer Analytics Initiative (WCAI) matches global companies overflowing with untapped customer data to the academic world’s top analytics researchers. These collaborations have given WCAI Codirector Peter Fader, who is the Frances and Pei-Yuan Chia Professor of Marketing at the Wharton School of the University of Pennsylvania and a member of the analytics association INFORMS, a unique perspective on the evolution, and current state, of customer analytics.
Q: What does the Wharton Customer Analytics Initiative do?
Fader: The Initiative acts as an incubator for new ideas in customer analytics, stretching across a wide variety of industries. Our corporate partners—often awash with the wealth of customer data arising from multiple platforms and technologies—provide samples of such data to the center. We then make this data available to carefully selected faculty from top universities around the world who use the sample as a catalyst for analytics R&D activities, creating advanced statistical methods to glean insights, and improving decision-making. We welcome the participation of consulting firms, syndicated data providers, and other vendors who not only deepen a client relationship by working with us, but can also learn new methods that can be useful in other settings.
Q: What strikes you about the current state of customer analytics?
Fader: It’s the best of times and it’s the worst of times. All of us analytics people are like kids in a candy store right now. There’s so much to play with and to think about; and there is an intense interest on the part of so many companies to actually talk to us geeks and nerds. That’s really, really great.
Q: What’s not so great?
Fader: There are a couple of disturbing trends. One is that there are a lot of people jumping on the bandwagon who don’t really have a full appreciation of the history of this field or of many well-established customer-behavioral patterns that were painstakingly proven years ago. A lot of newbies are basically reinventing the wheel because they think that the old rules don’t apply and that old data sources are of no relevance. We’re seeing a lot of people who are doing things that are no better than the work that was done decades ago and sometimes worse. That’s problematic.
Q: What’s an example of a commonly overlooked but time-tested approach?
Fader: One would be a depth-of-repeat model. The model assesses how quickly customers return to buy a new product after purchasing it for the first time; there’s a pattern from trial to first repeat purchase, from first repeat to second, second to third, third to fourth, and so on. Back in the 1960s some researchers—they were actually practitioners, not academics—found striking patterns in how regularly customers return to buy a product after they’ve purchased it previously. I can predict with alarming accuracy how quickly people who have purchased a product five times will buy that same product a sixth time, and that’s the case whether we’re talking a brick-and-mortar retailer or Amazon. Today’s “data scientists” know nothing of this stuff. In fact, many of today’s data scientists aren’t even scientists, they’re technicians. They’re very good at manipulating data, and searching and storing and all that. But they don’t know about the science behind the data and that’s a troubling sign.
Q: What’s another troubling customer analytics trend?
Fader: Many companies are quick to overuse or even abuse the data. A lot of the behavioral targeting you as a consumer are subjected to—browsing for a pair of shoes that then follow you around the Internet for weeks, for example—are not terribly effective. Worse, these types of analytics practices will eventually lead to public backlash or legislation that could squash a lot of the wonderful stuff these companies should be doing with analytics.
