Cam Rungie

Is there more information in best‐worst choice data?

Best Worst Scaling (BWS) has been shown to be a powerful method for preference and attribute importance measurement and is already widely established in wine marketing to analyse what drives wine consumers’ purchase decisions. Most prior BWS studies utilised Best Worst scores on an aggregated level only to measure relative attribute importance for the total sample or used sociodemographic or wine behaviour related variables to describe apriori segments. Not many studies considered individual differences in the Best Worst scores to find post hoc segments based on revealed differences of attribute importance. Is there more information contained in Best-Worst data than has been considered to approach this problem? We exemplify for data of British on-premise wine purchase behaviour how considering the attribute variance-covariance matrix allows valuable insights into what drives consumer heterogeneity. Attributes with high variance signal respondents’ disagreement on their importance and indicate the existence of distinctive consumer segments. Attributes jointly driving those segments can be identified by a high covariance. Based on the variancecovariance matrix we identify five dimensions of utility with a principal component analysis which show to be a very efficient tool for an effective interpretation of behavioural drivers of clusters derived by latent clustering. Our analysis opens new doors for marketing research to a more insightful interpretation of BWS data. This information gives marketing managers powerful advice on which attributes they have to focus to target different consumer segments.

Revealed preference analysis of red wine attributes using polarisation

Purpose – The usual method of analysis of product attributes in marketing is to fit a multinomial logit model within a stated choice experiment, to determine the impact of attributes on the choice probability, which is equivalent to market share. The market share is intuitive and is based on each single choice in the study. However, revealed preference allows for a study into repeat purchase and loyalty, which can also be rich constructs for determining consumer preference.Design/methodology/approach – The authors introduce a loyalty measure, polarisation, and show results based on a wine data set of revealed preference. Polarisation is a function of the beta binomial distribution (BBD) and can also be a function of the Dirichlet multinomial distribution (DMD). The DMD provides a standardised or average loyalty effect for each attribute (such as wine variety), and the BBD an individual effect for each attribute level (such as cabernet) within the attribute. While the DMD results provide a rich “first‐pass...

Using polarisation to identify variations in behavioural loyalty to price tiers

Purpose – This paper has two purposes: to use polarisation to identify variations in loyalty and to apply polarisation to an important non‐brand attribute, price.Design/methodology/approach – A comprehensive revealed preference data set of wine purchases is used to apply polarisation. Polarisation was defined in two ways: as a function of the beta binomial distribution (BBD) to give a measure of loyalty for an alternative; and as a function of the Dirichlet multinomial distribution (DMD) to give a baseline level of loyalty. Variations were identified by analysing the differences between the BBD and DMD.Findings – Polarisation was shown to be one way of identifying variation across price tiers. In the empirical example used, the DMD model is violated with the price tiers not being directly substitutable with one another. Buyers show excess loyalty towards the lowest and highest price tier levels. One tier shows “change‐of‐pace” loyalty. Small brands do better when they focus on high loyalty tiers, middle b...

How Loyal are Italian Consumers to Wine Attributes

The paper focuses on the study of loyalty towards three product attributes for wines sold in the Italian retail sector, through the polarisation index. A product attribute is a characteristic of a product (for example, price), made up of various levels (at least two for each attribute) of that characteristic (for example, <€3; €3–€5; €5–€7; >€7). The main findings highlight that the format proves to be the attribute which generates the highest loyalty level for Italian consumers. It is followed by quality designation and price. In particular, table wines, wines sold below €3 and wines purchased in >1.5 litres formats achieve the highest loyalty values, while foreign wines, wines sold at more than €7 and in <0.75 litres bottles are those that discourage behavioural loyalty the most.

Is the polarization index a valid measure of loyalty for evaluating changes over time

Purpose – This paper aims to argue that the polarization index (φ) represents a valid loyalty measure for evaluating changes over time.Design/methodology/approach – The brand performance measures (BPM) are a valid and useful tool for marketing managers in measuring the loyalty consumers attach, in a single time period, to a product or brand. However, the BPM reflect other attributes and not only loyalty. Over time, what might appear to be a change in loyalty may actually be a change in market size or market share. The polarization index (φ) is not biased in this manner and is more appropriate for evaluating changes over time. The study compares the results obtained with three well known BPM utilised for the analysis of loyalty – the purchase frequency, the share of category requirements and the repeat rate – with those obtained with the φ on the purchases of wine made by Italian consumers in the retail sector over two three‐year periods (2003‐2005 and 2006‐2008).Findings – The study shows that the BPM are...

Integrating consumer characteristics into the stochastic modelling of purchase loyalty

Purpose – This paper aims to extend a widely used stochastic model of purchase loyalty to include covariates such as demographics, psychographics and geodemographics. Potentially, this allows covariates to explain variations in brand performance measures (BPMs) such as penetration/reach, average purchase frequency, sole buying, share of category requirements, repeat purchase and so forth. The result is to integrate consumer-based segmentation into previously unsegmented stochastic models of brand performance. Design/methodology/approach – This paper describes a model for predicting BPMs. Covariates are then introduced into the model, with discussion of model specification, model estimation, overall model assessment, and the derivation of generalised theoretical BPMs. The outcome is a practical procedure for behavioural loyalty segmentation. Findings – The implications for strategy and management in applying covariates to the BPMs are considerable. Where there are concentrations of consumers with high repe...

Regularities in the consumption of a subscription service

Purpose – The purchase distribution of consumer packaged goods has been extensively modelled by the negative binomial distribution (NBD). As the characteristics of packaged goods differ from those of subscription services, the latter may be a boundary condition for NBD. In this study, the authors aim to test whether NBD extends to a subscription service. Also, NBD assumes non‐heterogeneous purchase behaviour. The authors determine whether augmenting the model with a covariate (customer tenure) to account for consumer heterogeneity will produce a better fitting model.Design/methodology/approach – Phase 1 fitted a base NBD model using a random set (n=1,546) of mobile‐telecommunication consumption (monthly billed amount). Phase 2 extended the base model by incorporating customer tenure as a covariate, and re‐fitting the model.Findings – NBD does apply to subscription markets. Also, accounting for heterogeneity in customer tenure produces a better model fit, and shows that with increasing tenure individual co...

Brand Loyalty vs. Loyalty to Product Attributes

Typically, in a specific category, a product or service can be defined not only by the brand it bears but also by multiple other attributes (e.g., pack size, price level, product formulation): a total of ten attributes for our example, detergents. While many papers have been devoted to brand loyalty, very few have been devoted to loyalty to other attributes: Is a household loyal to a certain pack size? To a certain price level? To a certain product formulation? These questions have important managerial implications in terms of the design and management of a brand’s product line. In this paper, we propose a systematic comparison of brand loyalty against loyalty to other attributes. We show that the two common measures of behavioural loyalty, share of category requirements and repeat rate, have problematic validity, due to the confounding influence of market share and purchase rate. In contrast, we argue that the Polarization index avoids these confounds and is therefore a better measure of loyalty. On the example of detergents, we show how to use Polarization to compare behavioural loyalty across attributes (e.g., are households more loyal to brands or to price levels?) and across different levels of one attribute (e.g., are consumers more loyal to high price levels than to low price levels?).

Wine and Tourism: A Good Blend Goes a Long Way

Wine and tourism are seen as good bedfellows, and rightly so. Globally the wine industry is increasingly dependent on leisure and tourist visitors to sustain the growth of local and export sales, both in terms of immediate turnover from “cellar door” transactions and as a long-term image generating mechanism (Lockshin, 2001). Growing interest, both applied and academic, in wine tourism as an industry is indicative of the fact both practitioners and researchers consider it an important contributor to continued success for wine producers (Bruwer & Lesschaeve, 2012). However, only a relatively small percentage of the total international visitors to a country actually visit a winery when on holiday (Department of Industry Tourism and Resources, 2009).

L’impatto degli attributi del prodotto vino sulla fedeltà comportamentale dei consumatori italiani

The paper focuses on the study of loyalty to 5 wine attributes – price, format, denomination of origin, production areas and brand positioning – for wines sold in the Italian modern distribution. The observation and evaluation of loyalty levels for these 5 attributes is undertaken using the polarisation index. This index is obtained from the Beta Binomial Distribution (BBD) model – already applied in literature for the analysis of the loyalty toward brands and product attributes, and for the positioning of a brand as a niche or a change-of-pace brand. The main findings highlight that format results the attribute able to generate the highest level of loyalty. Denomination of origin and production areas affects loyalty at the same extent and they do it more than price. Loyalty to brand market shares showed two different situations with the general value being the second most significant attribute, but when the analysis concentrated on the 0.75 litre format, consumers demonstrated to care less about the positioning of the brand.