The market response to product line extensions
Bruce G. S. Hardie · Scholarly Commons (University of Pennsylvania) · 1994
The objectives of this research are (i) to gain insight into the determinants of the performance (specifically market share and cannibalization) of the most common type of new product, the line extension, and (ii) to develop tools for forecasting its performance at the pre-test-market stage of the new product development process. First, we propose a conceptual model of line extension performance which relates the performance of a line extension to characteristics of the line extension itself, characteristics of the brand launching the line extension, characteristics of the market into which the line extension is launched, the marketing support it receives, and the nature of any competitive reaction. The resulting propositions are tested using data for the set of line extensions launched in the toothpaste product category over the 1988-1992 period. Results indicate that parent brand strength, line extension characteristics (including its similarity to other items in the product line), and the marketing support received, have the greatest influence on the performance of line extensions. However, if the line extension is too similar to other products under the same brand, cannibalization becomes extreme. The pre-test-market forecasting of line extension performance focuses on methods which rely solely on historical data. In particular, we develop a scanner data-based choice-model that parsimoniously describes consumer choice among SKUs. The distinguishing feature of this model is that we estimate consumer preferences over the attributes that define each SKU, rather than directly estimating the preferences for each SKU. As a result, we are able to use this model to estimate the performance of certain classes of line extensions. This model is calibrated using scanner panel data for the fabric softener product category, and used to predict the market share and expected cannibalization of two line extensions launched into the fabric softener product category. The initial results are very encouraging, and demonstrate the potential of such modeling frameworks for forecasting the performance of line extensions.