Making Survey-Based Price Experiments More Accurate

Abstract

Marketers frequently want to know how the sales of their brand will respond to a change in price. Survey-based price experiments are a convenient and cheap method of finding this out but they have a reputation for giving inaccurate estimates of price effects. These inaccuracies can include over-estimation of price effects, non-significant price effects and even positive price elasticities. We sought to identify some of the causes of these inaccuracies by analysing both the existing literature and the results of a recent mail intercept survey based price experiment (n=382). We found that; (i) when price effects were overestimated, pictures of the relevant brands had not been included; and (ii) when non-buyers of the category were included in the sample, price effects were increasingly non-significant and elasticities were sometimes even positive. Consequently managers should be able to use the results of survey-based price research with greater confidence when some form of pictorial representation is used and non-buyers are not forced to make a choice.