Customers who binge-consume are more valuable, says study

"Predicting Customer Value Using Clumpiness: From RFM to RFMC," is by Yao Zhang of Credit Suisse, and Eric Bradlow and Dylan Small of The Wharton School: University of Pennsylvania. It appears in the Articles in Advance section of Marketing Science.

INFORMS is the leading international association for professionals in analytics and operations research.

Marketing managers traditionally segment customers by three summary measures (also known as the RFM model): Recency -- the period of time since their last visit, Frequency -- how often they visit, and Monetary value -- how much they spend on a visit.

Binge consumption is characterized by bursts of heavy buying interspersed by little or no buying. The authors call this pattern of consumption "clumpiness."

The authors develop a new measure of clumpiness that extends the "hot hand" literature found in statistics journals. That is, just as athletes have periods of hot and cold performance (e.g., shooting), customers also have hot and cold periods of visiting (binge-visiting) and buying (binge-purchasing). Even after controlling for frequency, visits and monetary value, clumpy customers provide more economic value to the firm than non-clumpy ones.

Some customers purchase very frequently (high F) for a period of time but haven't purchased in a while (low R). This behavior has two possible explanations. One is that these customers have quit or "churned." The other possibility is that these customers are clumpy or between "bursts." When they come back they will "burst" again. Firms can make money on customers who are between bursts if they successfully encourage them (possibly via target marketing) to return.

The authors also show two important results with implications for firms. First, by spending adequately on marketing, firms can increase the clumpiness of a customer with the hope of increasing their value. In the second result, the authors find that all customers are not equally clumpy. Women, young people, and customers in loyalty programs appear to be more clumpy than others. This focus on customer clumpiness shows new insights into buying and improves on the traditional model of recency, frequency, and monetary value.