Marketing decision analysis by TURF and Shapley Value

We consider a problem of marketing decisions for the choice of a product with maximum customer appeal. A widely used technique for this purpose is TURF, or Total Unduplicated Reach and Frequency, which evaluates a union of the events defined by the sample proportion of many products, or flavors of one product. However, when using TURF, it is often impossible to distinguish between subsets of different flavor combinations with practically the same level of coverage. An appropriate tool can be borrowed from cooperative game theory, namely, the Shapley Value, that permits the ordering of flavors by their strength in achieving maximum consumers' reach and provides more stable results than TURF. We describe marketing strategy reasons for using these techniques in the identification of the preferred combinations in media or product mix.