Interpretation of brand penetration figures that are reported by sub-groups

Managers are often required to interpret key brand performance measures such as penetration (the proportion of people who bought at least once) and average purchase frequency (number of times the brand is bought in a set period) that are presented according to certain categorisations of the population, such as by demographic groups. Such categorisations may also include psychographics, for example, but this analysis chooses to focus on demographics for the purposes of illustration. Brand performance measures may vary across these demographic groups, which presents a challenge for interpretation. This is because it is not easy to distinguish between the variance due to brand segmentation, and the variance that is due to variations in category buying among those demographic groups. This paper shows how to incorporate category level information in order to create 'theoretical' or expected brand performance measures for each demographic. The actual figures can then be compared with the expected figures, which makes exceptions or deviations that could signal brand-level segmentation easier to identify. The approach is tested on the breakfast foods category in Australia. The results show that over 95 per cent of the variance in brand performance across demographic groups is simply attributable to (a) the overall size of the brand; (b) the penetration of the category into each demographic; and (c) the average purchase frequency with which each demographic group buys the category. In this category, brand-level segmentation accounted for a very minor proportion of the difference in penetration levels for brands across various demographic groups. That said, the analysis method did identify two brands with markedly different performance levels compared with their expected levels, in specific demographics. One brand overperformed with younger families, and the other underperformed with larger families. These performance levels were checked against a subsequent year of data and found to be robust. The main managerial implication of this study is that a simple model can be used to identify 'expected' brand performance levels, and the method clearly shows that almost all the variation in brand penetrations into various demographic sub-groups is simply attributed to brand size and category level effects.