Once More and I'll Scream

16 May 08 - Mark Klein

This week, when I asked a new client what kinds of insights her company was expecting from using Longbow, I had to restrain myself from an offensive, too sharp reply. She said she wanted our analytics to identify her best customers. We get that request often, and sometimes it can drive me up the wall.

There were several reasons for my frustration. First, she should already know her best customers, because they are spending the most and buying often. How well could her company be managing if they don’t know their best customers? Second, if her company had been doing any segmentations at all, those top customers would stand out. Even unsatisfactory linear methods like RFM will identify best customers. Third, she was asking the wrong question.

I appreciate that many companies are so busy keeping more new customers pouring into the top of their funnel than leak out the bottom that they don’t have the time or smarts to do any meaningful analysis. ‘Spray and pray’ marketing is all they know how to do. They have no “analytic eyes” in our parlance. When they do begin to look critically at their customer population, their knee-jerk reaction is to locate their best customers. But today that knee-jerk reaction is inadequate.

Mathematical marketing is capable of answering much more interesting and useful questions. Here are some that she should have been asking:

• Which customers are going to buy next, and what are they likely to purchase?
• Which of my customers, especially which of my productive customers, are in danger of defecting to a competitor? How much revenue is at risk?
• Which of my customers are good cross-sell prospects, likely to by something from my product set they have not previously purchased? What products should I offer them to increase our share of wallet?
• Which of my inactive customers are good candidates for a win-back campaign that will bring them back into our active customer list?
• Which of my middle tier customers are good prospects to make one additional purchase that could kick them up into a higher segmentation tier? How much new revenue would that bring into our company?

The difference between these questions and the one that almost caused me to lose it is that these questions are actionable. The answers lead directly to marketing campaigns that will improve top line revenue. That’s a lot more satisfying than screaming, though I have to admit that screaming (or venting onto a blog) can be satisfying too.