Increasing a brand’s competitive clout: the role of market share, consumer preference, and price sensitivity

ABSTRACT

The market power (or ‘competitive clout’) of a brand is an increasingly important component of modern marketing strategies. However, the factors that enhance a brand’s competitive clout (BCC) are poorly understood. This study therefore suggests an integrated model of BCC and three factors that are proposed to play a role in its formation: (i) consumer price sensitivity; (ii) brand market share; and (iii) consumer brand preferences. These variables are examined both individually and simultaneously to demonstrate the direct effect of each on BCC and how their inter-relationships contribute to BCC. In doing so, a two-step empirical analysis is conducted. First, two multinomial logit models provide an own- and a cross-price response matrix for a chosen set of competitive brands. Secondly, BCC is regressed against the variables of market share, intrinsic preferences, and price sensitivity using an interaction effects regression model. The results of the analysis show that market share is not the only way to increase BCC; in particular, consumer preferences, and especially pricing decisions, are shown to play a key role in developing a strong brand.

KEYWORDS:

• Market share
• intrinsic preferences
• competitive clout
• pricing decisions

ORCID

Juan Carlos Gázquez-Abad http://orcid.org/0000-0002-9181-2850

Notes

¹ We thank the two anonymous reviewers for this suggestion

² See Appendix 1 for original syntax and estimations in NLogit 4.0.

³ They can be obtained by exponentiating the multinomial logit coefficients (e^coefficient). Odd ratios are relative measures of effect size and, therefore, provide information on the strength of relationship between a single brand relative to the baseline brand. They are also known as ‘relative risk ratios’.

⁴ Logistic regression does not have an equivalent to the R-squared that is found in OLS regression. However, to evaluate the goodness-of-fit of logistic models, several pseudo R-squareds have been developed. These are ‘pseudo’ R-squareds because they look like R-squared in the sense that they are on a similar scale, ranging from 0 to 1 with higher values indicating better model fit, but they cannot be interpreted as one would interpret an OLS R-squared and different pseudo R-squareds can arrive at very different values. Thus, there are a wide variety of pseudo R-square statistics (e.g., Efron, McFadden, Cox & Snell, Nagelkerke, etc.; check Louviere and colleagues (2000) for a greater description of these statistics). In this paper, the McFadden’s (adjusted with the number of parameters) pseudo R-squared statistic has been reported. McFadden’s pseudo R-squared is defined as 1 minus the ratio of the log likelihood of the full model (i.e., the model under analysis) and the log likelihood of the intercept model (with no predictor variables). The adjusted version of this statistic penalises the model for including too many predictors

⁵ See Appendix 2 for original syntax and estimations in Stata 12.0.

⁶ Nevertheless, Cronbach (1987) supported that the major threat of multi-collinearity in interactive models is not substantive but rather practical. In this respect, multi-collinearity does not affect the properties of the OLS. Nonetheless, high correlations between predictors can cause computational errors on standard computer programs.

⁷ Most commonly, a value of 10 has been recommended as the maximum level of VIF (e.g., Hair, Anderson, Tatham, & Black, 1998; Kennedy, 1992). Nevertheless, a recommended maximum VIF value of 5 can be also found in the literature (e.g., Rogerson, 2001). In this paper, the latter recommendation will be followed.

⁸ See Appendix 2.

⁹ In addition, the F-incremental test also suggests that model 3 does not provide a significantly better fit than model 2 (F_1,4 = 0.1142).

¹⁰ We thank the two anonymous reviewers for this suggestion.

¹¹ Hausman test results and more details are available upon request to the corresponding author.

¹² Variables were mean-centred before estimating regression models.

Additional information

Notes on contributors

Juan Carlos Gázquez-Abad

Juan Carlos Gázquez-Abad is currently an associate professor of marketing at the Economics and Business School, University of Almería (Spain). He obtained his PhD in Marketing at the University of Almería. He was a visiting professor at the University of Ghent (Belgium) in 2005. His research interests cover several marketing topics, especially those related to retailing and consumer behaviour. He is Associate Editor of the International Journal of Business Environment.

Francisco J. Martínez-López

Francisco J. Martínez-López, MSc in Marketing and European PhD in Business Administration, with Extraordinary Doctoral Prize (University of Granada, Spain), is Professor of Business Administration at the University of Granada and the Open University of Catalonia (Barcelona) in Spain. He has been visiting scholar at the Zicklin School of Business (CUNY, USA), Aston Business School (Aston University, UK), the University of Chicago Booth School of Business (USA), the Michael Smurfit School of Business (University College Dublin, Ireland), and the Complutense University Business School (Madrid, Spain). He is Editor-in-Chief of the International Journal of Business Environment (Inderscience Publishers), Associate Editor of European Journal of Marketing (Emerald) and belongs to the Editorial Board of Industrial Marketing Management (Elsevier).