Abstract
It is an everyday marketplace occurrence that brands lose and gain share. However, a brand's sales gain or loss can be attributable to very different factors, and thus understanding the sources of sales gain or loss would seem to be an important aspect of a brand manager's job.
The primary purpose of this research is to develop a model that can answer the following questions: (i) What are the sources of gain or loss of a brand's sales due to category volume and brand switching? (ii) What consumer and marketing characteristics affect consumers' purchase frequency of the product category and that of different brands? (iii) Do all consumers behave similarly, or are there distinct segments which respond to marketing actions differently? (iv) If such segments exist, what is the size and composition of each segment, and what is the appropriate strategy for each group of consumers?
Our modeling approach, which decomposes a brand's sales into category volume and brand choice components, has many similarities and also several differences with traditional approaches. Similar to the NBD and Dirichlet models, we assume that a consumer's category purchase rate follows a Poisson distribution, and the number of purchases per brand follows a multinomial distribution. Our model differs from traditional models by including marketing mix variables, by accounting for loyal or near-loyal consumers, by explicitly incorporating consideration sets, and by segmenting consumers on the basis of their brand perceptions and their responses to marketing mix variables.
We account for consumer hetereogeneity by identifying homogeneous latent segments that capture differences in consumers' response to marketing variables in both brand choice and category volume behavior. Calibration of the category volume and brand choice models and the separation of loyalty and switching segments are done simultaneously so that there is no need to assume any specific hierarchy and, in contrast to the usual assumption of independence between choice and volume which may be unreasonable at the aggregate level, the requirement is that choice and volume decisions be independent only at the segment level.
We investigate the properties of our model in the context of a national survey of supermarket purchases of jumbo paper towels. 2,500 households were surveyed who were asked to provide information on: rolls of paper towels bought in the last 4 weeks, rolls of each brand of paper towel bought in the last 4 weeks, brand usually bought, brands that a household bought or would consider buying, average price paid or expected for each brand bought or considered, ratings on 20 brand attributes (e.g., strength, absorbency, etc.), and demographic information (e.g., family size).
We found that not only did the modeling framework outperform a variety of competing models, it also provided insights into the competitive structure in this market. The loyal segments could be distinguished on the basis of price sensitivity. Less than 10% of the loyal households were price insensitive and, in general, households showed increasing price sensitivity in their category volume decisions if they had more children and were heavier users of paper towels. This was consistent across all brands. The model estimates that about 71% of the households are switchers and five switching segments are needed to characterize household purchase patterns. Two of the largest switching segments (labeled as Price Sensitive and Value Segments) are very price sensitive in both their brand choice and category volume components. In both these segments private label brands are dominant with about 30% share. Segments also emerged on the basis of Strength, Absorbency, and Tearing Ease of paper towels. Interestingly, we found that brand shares within these segments are quite consistent with the objective quality ratings of brands as given by Consumer Reports. Finally, price elasticity analysis for one of the brands, Bounty, reveals that a 5% drop in Bounty's price increases its sales by 13.6%. Almost half of this increase comes from brand switching, with the other half coming from increases in category volume (e.g., stockpiling). Bounty gains the most from the price sensitive segments (Price and Value Segments). Private labels are hurt least by Bounty's price cut.