Dumbing Down Behavioral Analytics
The biggest problem that Jason Rushin, director of marketing for behavioral analytics software firm Quantivo, runs into with companies is "they they are stuck in their traditional ways of looking at customers." That's one reason he and Quantivo joined with book publisher Wiley to produce "Behavioral Analytics for Dummies," part of the familiar Dummies series but aimed at marketers who are stuck segmenting their audiences by whether they are male or female, in what age range and living in what area of the country. Instead, these marketers need to change that equation and start with the end behavior the marketer wants to encourage. Then, the marketer should track back to what brings the customer to that point.
"What they should be looking at is more like, who has purchased a surfboard -- and then, within 10 days, who purchases a roof rack to carry that surfboard? Then we want to shoot them a campaign for a protective cover for their surfboard," Rushin says.
Marketers are astonished to discover that many of their old equations and assumptions become less important as they focus instead on what they want their customers to do by learning from what they have done in the past. Behavior is independent of age, region or any of the usual demographic lenses. "You care about what that end behavior is," Rushin says. "Whether it is a click or a video view on a Web site or a purchase or a call to a service rep -- what behaviors led up to that? Who has done 90% of those behaviors but hasn't crossed the finish line?"
While behavioral analysis may seem old-hat to those of you in the behavioral targeting ad game, it is still a relatively new way of thinking for many marketers. Hence, a "Dummies" book that makes the basic precepts accessible to marketers. In many ways behavioral analytics reverses the polarity of marketers' thinking and forces them to learn a new language with a new set of questions. "You start with the end question. I want to sell more X or I want more people to view sports content on my Web site or I want to get more ad views on videos. And then it is a matter of seeing people who have exhibited those behaviors and saying, let's work backwards," says Rushin. "What else have they done, and what propensities do they have to do X, Y, or Z that lead them to the final finish line?"
As Rushin and co-author Jennifer LeClaire say in the book, "These tools also allow you to uncover buried relationships that link customer characteristics to behavior -- even ones you never considered. For example, you can find attributes that are highly correlated with specific groups of online or offline transactions." The book has chapters on comparing various behavioral analytic approaches like Web analytics and traditional business intelligence and predictive analytics. It also covers specific application scenarios like using the analytics to optimize online marketing or B2B sales and support. While the book was done in partnership with Wiley, it is available free from Quantivo.
While the "Dummies" book is a promotional tool for Quantivo's software, Rushin says that the aim of both the book and the product is to address a real need in companies to dumb down the behavioral sciences so that they are more actionable.
In many cases companies already have the data they need to move their marketing in these directions. Part of the problem is that the data is not immediately accessible to the marketers. "The marketing guy has to ask a statistician or a Ph.D., 'I want to know X, Y, or Z,' and it takes a couple of days to get that answer," notes Rushin.
The marketers are not in close enough contact with the analytics that are even possible, so they fail to ask the right questions. "These marketers aren't used to asking the questions directly so they are forced to ask 'how many people came to our site yesterday?' because they know they can get that information," he adds. If the tools and the disciplines of behavioral analytics are made accessible to the marketers then they know better what follow-up questions to ask of the data.
"You look at a clothing retailer and someone who is in charge of men's pants," he says. "They're not looking at what their sales of men's clothing was yesterday. That tells them nothing. They want to know, who bought black pants? What have they purchased in the last three weeks leading up to that black pants purchase? And what should I promote to them next? If they haven't purchased a belt, then I will promote a belt to them. If they have purchased a belt, then I will promote shoes to them. It can get very granular quickly. But if you don't have access to that data, then you are making decisions based on very high level analysis and are starting to go back to gut feel."