Line extensions and cannibalisation: an evaluation of measures and methods based on longitudinal consumer panel data

Lomax, Wendy (1997) Line extensions and cannibalisation: an evaluation of measures and methods based on longitudinal consumer panel data. Doctoral thesis, University of London: London Business School.

Abstract

Practitioners, throughout the last two decades, have favoured line extension as the branding strategy for their new products. The use of an existing brand name for a new product has been fuelled by the appreciation that brands are assets, leading naturally to a desire to gain maximum benefit from them. This coupled with an increasing focus on effective resource allocation, has led to line extension developing into the marketing phenomenon of the 1980s and 1990s. But attractive though the use of an existing brand name may be, it is not without its potential costs. Dominant among these is the risk of cannibalisation. Sharing a common brand name may benefit the extension but makes the parent brand vulnerable to share loss. The new line may succeed only, or at least partly, at the expense of parent brand sales. The aim of the study is to enhance our understanding of the important effect of cannibalisation. Since most new product introductions are line extensions, a proper comprehension of this risk is critical in evaluating the success or potential success of much new product activity. This study examines the fortunes of 17 line extensions in fmcg categoriesin Germany and the UK through the late 1980s and early 1990s. Consumer panel data is used to generate five measures of cannibalisation, not previously used in academic research. These measures are market share and gain loss analysis (both in common use by practitioners), and deviations from expected share shifts, duplication of purchase tables and buyer analysis(derived from academic theory). The findings show that 1. The market shares of line extensions tend to be small 2. They almost always take share from the parent brand, i.e. canniballse. But the range is considerable. Zero percent to 100% of the line extension's volume can come from the parent brand. 3. Despite this, in most cases the total brand benefits (in volume terms) from the launch of the line extensions 4. Buyers of the line extension buy the parent brand significantly more often than other category buyers before the launch of the line extension 5. They continue to buy the parent brand after the launch of the fine extension 6. Predictably the Independence of Irrelevant Alternatives(IIA) axiom is not valid in the context of line extensions. This puts line extensions in a category distinct from other types of new products. 7. There is some evidence for product form loyalty which over-rides that for the common brand name. The UK detergent market shows powdered detergent parent brands resisting cannibalisation by their liquid line extensions. 8. Cannibalisation is a complex phenomenon to measure The five measures used in this study show varying degrees of correlation, but none display total agreement and most have limited levels of association. The timing and the duration of the period of measurement has a significant effect on the result. The implications of these results are that cannibalisation is a real risk of using an existing brand name. Although there are circumstances qhere share loss by the parent is not unwelcome, e.g. for contribution or defensive reasons, it is still important to be able to quantify it. This research examines the measures currently available, and concludes there is a need for improvement. A new definition of cannibalisation is proposed, and a corresponding new measure of cannibalisation is outlined. This measure takes into account trends in the parental market share, and the importance of the timing and duration of this measure. For many line extensions, the immediate post-launch period may be of interest but perhaps even more important is the longer term effect on the parent brand. As a result of this study, we are in a position to develop a definitive measure of this critically important effect.