Analytics at Wharton Research
An Integrated Model for Dynamic Brand Equity
This paper presents a unified statistical model designed to measure brand equity as it changes over time; and gauge the impact of increased brand equity on consumer’s product choices. Our model extends traditional models of brand equity which posit that strong brands are simply “more preferred” after controlling for the marketing mix (i.e. an intercept) to a more general model that allows for (1) brand equity to evolve (modeled as a Bayesian DLM model); (2) perceived product attributes to vary with brand strength ― an “X-perception effect”; and (3) perceived coefficients for product attributes to be a function of equity ― a “beta effect”. This extended model provides firms and researchers with a more comprehensive view of brand equity and how it manifests itself in consumers‘ product choices.
We apply the proposed model to a panel dataset on purchases in the pretzel category, demonstrating that brand equity had a measurable effect on price sensitivity and product attribute perceptions. Using the model we optimize the timing of marketing and derive optimal weekly prices. Comparing those results to simpler benchmark models of brand equity, we find that managers can potentially lose profit by using the reduced-form model due to mispricing.
Keywords: Brand Equity, Brand Management, Dynamic Multivariate Probit Model, Bayesian Computation