Predicting transitions in alcohol buying behaviour
This article uses the Negative Binomial Distribution (NBD) to model alcohol purchase frequency and predict the transitions among non-buyers, light (the bottom 50% of buyers), medium (the middle 30%) and heavy (the top 20%) from one year to another. Using purchasing data over two years in the United States (49,915 households) and United Kingdom (9,316 households) including beer, wine and spirits, we find that the NBD accurately predicts these transitions. The model as such provides a useful baseline for evaluating the effects of interventions to reduce heavy alcohol buying. For both ethical and commercial reasons, instead of targeting heavy buyers, companies selling alcohol still can more responsibly strive to increase their sales by acquiring new alcohol buyers as most of them will become light buyers.
CitationTrinh, G., Dawes, J., Sharp, B. (2025). "Predicting transitions in alcohol buying behaviour". Forthcoming in the Journal of Marketing Management.
Valuing the contribution of ultra-light buyers
Ultra-light buyers, those who, on average, buy a brand once a year or less, are important by number and their contribution to brand purchase occasions. By examining over 850 brands in almost 60 categories, we identify that ultra-lights make up 62% of a brand’s buyer base over five years and contribute 32% of purchase occasions and 31% of sales volume and value. In line with previous findings, we find the prevalence of ultra-light buyers is higher for national brands than private labels, likely reflecting the higher loyalty often observed for private labels. We also find no difference in ultra-light buyers’ contribution between high and low-price brands, indicating that price is not a significant factor affecting the contribution of ultra-light buyers. Finally, we find that both growing and declining brands have fewer ultra-lights than stable brands. As growth occurs via increasing penetration and loyalty, for growing brands, new buyers may be more ‘light’ than ‘ultra-light’ in terms of their loyalty; while for declines, these findings imply that there are not enough ultra-light buyers to maintain brand sales. A failure to recruit ultra-light buyers may be very dangerous for brand health.
CitationHossain, A., Anesbury, Z., Driesener, C., Trinh, G. (2023). “Valuing the contribution of ultra-light buyers”. Forthcoming in the Journal of Consumer Behaviour.
Where is the brand growth potential? An examination of buyer groups
Practitioners and academics have long discussed strategies for brand sales growth. A recent example is an industry debate in which different brand growth strategies were argued: https://www.mmaglobal.com/thegreatdebate (MMA Global & Neustarr, 2021). A central question in this arena is whether a brand should focus on its heavy, light, or non-buyers in its efforts to grow its sales. This study contributes to our knowledge about how sales growth can occur by investigating the potential contribution these three buyer groups can make to any sales gain. Using both a simulation study and an empirical study of purchases of approximately 12,400 households in the UK, across different brands and categories, we show that almost any brand’s headroom growth potential lies mostly in light or non-buyers of that brand. Even for large brands with high penetration the growth potential of light brand buyers eclipses heavy brand buyers.
CitationTrinh, G., Dawes, J., Sharp, B. (2023). "Where is the brand growth potential? An examination of buyer
groups". Forthcoming in the Marketing Letters.
Using the Dirichlet model to predict how fashion brands compete and grow on eBay
E-fashion brand competition has historically been studied from an attitudinal lens, through surveys and theory-based approaches. These studies generally examine consumer attitudes, satisfaction and loyalty; highlighting that trust, satisfaction and reputation are key e-commerce success elements. However, the empirical consumer behaviour literature rebuts the use of attitudes to explain brand performance, criticising their subjectivity and the overall ineffectiveness of loyalty as a brand growth tool. The current paper presents Gerald Goodhardt and colleagues’ Dirichlet model as an alternative approach to understanding e-fashion buyer behaviour. The Dirichlet model is a robust, stochastic model that has reliably predicted well-established law-like patterns of buyer behaviour and brand competition. We apply Dirichlet modelling to a new, non Fast Moving Consumer Goods (FMCG) category to extend consumer behaviour research in the online environment. The study uses consumer panel data from the United Kingdom, across two consecutive years. We conclude that the Dirichlet model has an excellent fit within the eBay fashion brand market. The study also identifies that the well-known double jeopardy pattern exists within this market, demonstrating that e-fashion brands grow through acquisition rather than retention of customers. This provides a different viewpoint than most e-fashion brand growth literature. In addition to this, we examine the predictive capability of the Dirichlet model. We use a holdout sample to show that the model can predict future brand performance metrics, which is an exciting new development in consumer behaviour research.
CitationChowdhury, S.,Barker, A., Trinh, G. and Lockshin, L. (2021) "Using the Dirichlet model to predict how fashion brands compete and grow on eBay". Forthcoming in the Journal of Consumer Behaviour.
Quantifying the target market for advertisers
Marketers who want to protect their brand’s share or grow it need to know who to reach and nudge with advertising. This paper uses continuous household panel data for 55 leading, advertised brands in 12 cpg categories to quantify their target market over different time frames and conditions (market type, brand size and dynamism). Results demonstrate that the customer base (brand penetration) must swell dramatically over time to maintain, let alone grow, market share. For stable brands, penetration typically doubles from its level in one quarter to a year, then again from one year to five years as brands continue to attract lighter buyers who underpin long run sales. Over five years, over a third of brand buyers are so light that they buy the brand just once, but such buyers are vital to sales and critical to growth. As well as quantifying the five-year target audience for brands across these conditions, we test the predictive accuracy of the NBD-Dirichlet as a benchmark. The implications for advertising and media strategy are detailed. The long-term lessons for targeting become clear – unless brands “target the market”, they have adopted a counter growth strategy.
CitationGraham, C., Kennedy, R. (2021). "Quantifying the target market for advertisers". Forthcoming in the Journal of Consumer Behaviour.
The Unbearable Lightness of Buying
Although marketers are increasingly asked to manage brands for the long-term, it is difficult to do so when no clear picture exists of long-term brand buying. This study reports cumulative behavioural loyalty outcomes for 200 UK consumer-goods brands when observed in a five-year household panel of continuous reporters. We examine these brands in intervals from one to five years against NBD (Negative Binomial Distribution)-model projections. Stationary brands attract over twice as many buyers in five years as they do in one. Of these buyers, 80% purchase the brand at a rate of once a year or less, yet contribute 40% to total sales, a Pareto ratio of just 60:20. For managers, this light buying is broadly predictable from NBD fittings to annual data, and implies a renewed emphasis on nudging the brand buying propensities of the whole market.
CitationDawes, J., Graham, C., Trinh, G. and Sharp, B. (2021) "The Unbearable Lightness of Buying". Forthcoming in the Journal of Marketing Management.
The Dirichlet Model in Marketing
We present a review of Gerald Goodhardt’s most famous contribution to marketing science – the NBD-Dirichlet model. This provides a powerful illustration of the complex pathway and useful associated discoveries that over 25 years lead to the specification and application of a key marketing model. We identify the process that started with the negative binomial distribution (NBD) applied to purchase incidence, examined alternatives such as the logarithmic series distribution (LSD) and the beta binomial distribution (BBD), and along the way developed conditional trend analysis (CTA). We then explore the development of brand choice beginning with duplication of purchase as a method for understanding the underlying patterns, before moving into various approaches to modelling choice including a multi-variate NBD, and eventually a multi-variate BBD – the Dirichlet multinomial distribution (DMD). We discuss key events in modelling consumer behaviour and outline the model’s implications for how marketers should think about consumers, or more specifically the Dirichlet consumer. Finally, we provide a survey of the model’s applications uncovering a rich recent history and bright future for Goodhardt’s legacy.
CitationDriesener, C. and Rungie, C. (2021) "The Dirichlet Model in Marketing". Forthcoming in the Journal of Consumer Behaviour.
The Long-Term Erosion of Repeat-Purchase Loyalty
The study investigates the long-term erosion of repeat-purchase loyalty among consumers who purchase brands in a one-year base period. We find pronounced erosion in repeat-buying over the long-term. The proportion of buyers from a base year that fail to buy the brand in a later year increases steadily over time, from 57% in year 2 to 71.5% by year 5. Loyalty erosion has been reported before (Ehrenberg, 1988; East & Hammond 1996) but only over short periods. This study examines the phenomenon over five years, confirms that the rate of erosion does diminish over time, and that it is related to category and brand characteristics, as well as marketing mix decisions.
CitationDawes, J., Graham, C., Trinh, G. (2020) "The Long-Term Erosion of Repeat-Purchase Loyalty". Forthcoming in the European Journal of Marketing.
The Natural Monopoly effect in brand purchasing: do big brands really appeal to lighter category buyers?
This paper investigates the Natural Monopoly effect, which is that large brands have buyers who are on average less frequent or ‘lighter’ purchasers of the product category. The study analyzes the NM effect for brands in 28 consumer goods categories in The Netherlands. The analysis employs a multiple regression with category purchase rate as the dependent variable; and brand penetration, together with brand price, brand type, average pack size and promotion incidence as independent variables. The study finds that higher brand penetration is indeed associated with a lower rate of category purchase, controlling for the other variables in the model. The NM effect is reasonably large: the largest two brands in a category tend to have a buyer base that on average purchase the category about 25% less frequently than those of the smallest two. The study also derives an explanation for how large brands are generally purchased more frequently, even when their buyer base on average buys the category less frequently. The findings imply that a focus on heavy category buyers is inconsistent with the goal of growing a brand.
CitationDawes, J. (2020). "The Natural Monopoly effect in brand purchasing: do big brands really appeal to lighter category buyers?". Forthcoming in the Australian Marketing Journal.
Expanding marketing empirical generalisations to health behaviours: physical activity is not so different from buying behaviour, after-all
The Negative Binomial Distribution (NBD) is a model that describes consumer purchase frequency over time. This paper tests the applicability of this model to a novel context: physical activity behaviours (using data obtained from Australia, the United States, and Singapore). The fit of the NBD to the data demonstrates that physical activity behaviour is consistent with other consumer behaviour patterns. Within a one-week period, the majority of people are either non- or light engagers of the different intensities of leisure-time physical activity. Yet, people are not ‘active’ or ‘inactive’, rather, degree of engagement varies. Infrequency of reported levels and variety of physical activity might be due to health promotion having a strong focus on rational persuasion and less focus on mass communication that builds mental availability. Our contribution broadens the applicability of the NBD showing it can be helpful for those seeking to promote health behaviours, not just purchases.
CitationWilson, A., Sharp, B., Nguyen, C., Bogomolova, B. (2017). "Expanding marketing empirical generalisations to health behaviours: physical activity is not so different from buying behaviour, after-all." Australasian Marketing Journal Accepted November 3, 2017. .