Section III: Marketing-Mix Analytics

This section presents multiple regression analysis. A fundamental premise of marketing-mix analytics is that the mix elements affect customer behavior simultaneously. Therefore, models that estimate elasticity should reflect this notion. It is therefore necessary to estimate the joint effect of the marketing mix on customer behavior. Chapter 7, “Multiple Regression in Marketing-Mix Models,” and Chapter 8, “Design of Price and Advertising Elasticity Models,” provide the definition of elasticity, present the log-log linear model framework for estimating elasticity, and list the control variables that are necessary to include in marketing-mix models to avoid biased elasticity estimates.

You can apply the process of estimating price and advertising elasticity models in the context of the vodka industry. The data provides 13 years of sales, price, and advertising data across channels for 28 vodka brands. Chapter 9, “SVEDKA Vodka,” provides a case-study context to apply the estimated price and advertising elasticity to the context of a brand within the vodka industry.

When you have completed this section, you should have a good understanding of the workhorse of resource allocation: estimating marketing-mix elasticity. The chapters in this section provide ample discussion of unit roots in time series analysis, conditional effects, statistical versus economic significance, and omitted variable bias.