Get Smarter with Marketing Data, Says Stanford's Lattin

Jim Lattin of Stanford's Graduate School of Business (GSB) gears his marketing research and course offerings around making sure managers know the effects and returns of their marketing investments. Last week, we heard how he educates his students on the difficulty in structuring and running sales forces. Now, he'll tell us about how marketers underuse and misuse customer data as they structure their marketing programs.

BNET: How has your teaching of the subject of marketing changed recently?

Lattin: I spend a lot more time than in the past teaching about measuring market performance and figuring out the ROI on marketing strategies. Students are great at telling you about their plans--"I think we should do this and this and this, and here's our really creative idea." I say great, but what you really need to tell me is how are you going to know when you've succeeded and how much is it going to cost you to get there? How do you know it's going to work? You need to invest as a marketer in tracking and measuring, so you know where the good dollars are going and where to turn off the spigot. More and more we have companies that have great data, but they are lagging in terms of their sophistication and their ability to leverage the information--many of the programs they would consider just won't pay out.

BNET: What misperceptions do marketers commonly have about the effects of certain types of programs and promotions?

Lattin: Loyalty programs give marketers the opportunity to identify their quote "best customers" unquote. They give them recognition or special deals to keep them happy and keep them coming back--. What marketers haven't done is assess the extent to which that spend makes a difference. You want to be allocating resources to customers who will respond the most. After I spent money on them, the "best customers" would still be good customers, but I would not have moved the needle at all with respect to satisfaction, longevity as a customer, or the overall margin they contribute to me. I think that marketers are trying to get really fancy, creating tiers and tiers of customers, describing how good they are and what they are spending, but they don't test to see what's the return on redirecting those resources to these various customers. When people create a coupon program, for example, they might view a 40 percent redemption rate as a huge success. But it's not a success at all if 100 percent of those redeemers would have bought the product anyway in the absence of a coupon.

The goal is to increase the productivity of your spend. Some feel better about themselves because their spend is going toward "better" customers versus "worse" customers, but they are not maximizing the bang for their marketing buck.

BNET: So, if you want to be proactive in your strategies, how can you use information to structure a great marketing plan?

Lattin: One of the things about marketing that has been shown over and over again is that in most categories, demographics are only weakly predictive. What marketers want to do instead is to set up a transaction history [about their customers]. This is why paid search advertising on the web is relatively effective. Search is behavior; people are doing something that reveals information about their interests. Marketers find that much more powerful than knowing someone's gender and age and education and so on. Yes, gender will help you predict the sales of something like feminine hygiene products. But for the sales of Coke and Pepsi and certain brands, the discriminating power of those sorts of addressable demographic characteristics is really quite weak. With a supermarket, they probably know the ZIP code the person is coming in from--and they may know some of the characteristics about their household. But they also have a lot of data about the deals the customer has responded to in the past. With a little bit of analytical expertise, those pieces can really be put together compellingly.