Advertising Effectiveness for Multiple Retailer-Brands in a Multimedia and Multichannel Environment

An important aspect of multimedia advertising effectiveness that remains unexplored is a customer-level analysis of the relative importance of each medium for multiple retailer-brands within a product category. The increasing availability of customer databases for parent companies containing multimedia ad exposures, sales transactions in several purchase channels, and information across multiple retailer-brands now allows for a broader examination of advertising effectiveness. In this research, the authors monitor 4,000 customers over two years, linking their exposure to three media (email, catalogs, and paid search) to their in-store and online purchases for three retailer-brands in the clothing category. They develop a Tobit model for sales response to multimedia advertising that captures within-brand and within-channel correlations and accommodates individual-level advertising response parameters. Due to the very large number of observations (2.4 million) and random effects (60), the authors employ an emerging machine learning technique, variational Bayes, to estimate the model parameters. They find that email and sometimes catalogs from a focal retailer-brand have a negative influence on other retailer-brands in the category, whereas paid search influences only the focal retailer-brand. However, competitor catalogs often positively influence focal retailer-brand sales, but only among omnichannel customers. They segment customers by retailer-brand and channel usage, revealing a sizeable group of customers who shop across multiple retailer-brands and both purchase channels. Moreover, this segment is the most responsive to multimedia advertising.