Importance-performance analysis for internet stores: a system based on publicly available panel data

We demonstrate how publicly available ratings of Internet stores can be used to perform actionable importance-performance analysis (IPA). We use a panel of 422 non-food online merchants, which allows us to account for store-specific effects, and therefore obtain better estimates of each attribute’s influence on purchase intention. Our approach is economical, as merchants do not have to spend money on surveys that firms typically have to conduct for IPA. We have enhanced traditional importance-performance analysis by deriving attribute importance indirectly using the Shapley Value decomposition of a first-difference regression’s R2. We have also modified the standard “importance-performance” graph by adding the information about the directions of changes in attribute ratings between the two periods, which made it easier to point out strategic mistakes. We demonstrate the diagnostic power of our approach by doing the actual importance-performance analysis for two of the stores. The potential of enhancing the expert system using other methods and other types of publicly available data is discussed.