Marketers must be careful not to let their models become malleable to suit their theories
A quote from Kurt Lewin, the father of modern social psychology, is indelibly written into my brain: “There is nothing more practical than a good theory.” Lewin saw theories as immensely helpful not just in understanding behavior but in generating new ideas to solve problems. I fear, however, that Lewin’s message is often lost in this age of Big Data and analytics.
Sophisticated software tools search mounds of customer data for patterns of regularity, leading to thousands of poorly controlled micro “field experiments.” Actions (e.g., targeted offers) that seem to work, for some reason, are repeated until they stop working, for some reason. Empirical prediction trumps explanation. It is a process that creates little learning, limits creativity and ultimately produces waste.
People say that I am a bit “model happy.” Yes, I’m one of those people who draws circles, boxes and arrows on the white board to explain why customers are acting the way they are and what can be done to improve results. Experience has taught me that marketing phenomena are complex, that data can be misleading, and things are not always what they seem.