Over the last several years, conjoint analysis has been increasingly proposed as a method to estimate class-wide damages in a variety of consumer class actions including product liability, false advertising, product labeling, and data privacy and data breach matters.
Challenges Relating to Estimating Class-Wide Damages Using Conjoint Analysis
Conjoint analysis is a survey-based marketing research tool developed by academics to understand and estimate consumer preferences. It has been adopted by businesses and industry practitioners to help make decisions on new product development and market segmentation analysis, among other uses. Over the last several years, conjoint analysis has been increasingly proposed as a method to estimate class-wide damages in a variety of consumer class actions including product liability, false advertising, product labeling, and data privacy and data breach matters. However, the technique’s underlying assumptions and limitations render it unsuitable for calculating damages in a class action setting. Demonstrating and effectively conveying these limitations requires an in-depth understanding of conjoint analysis, survey methods, economic damages, and the class certification framework.
Willingness-to-Pay Measures Generated by Conjoint Analysis Cannot Approximate Fair Market Value or Market Price
Conjoint analysis has been increasingly proffered as a method to estimate benefit-of-the-bargain damages (i.e., alleged overpayment or price premium claims). These damages are typically measured as the difference between the fair market value (or price) of the product as warranted and the fair market value (or price) of the product as sold. However, it is not possible to estimate any fair market value or price through conjoint analysis. Price is determined by the interaction of factors impacting product demand as well as product supply. Conjoint analysis is not equipped to account for any supply-side factors (e.g., a manufacturer’s willingness to sell its products at a given price) or competitive activity in the marketplace. It can only generate estimates of consumers’ willingness-to-pay for a product, a measure that is untethered to market prices.
To compensate for these limitations of conjoint analysis, some plaintiffs’ experts have claimed that they can account for supply-side and competitive factors based on the assertion that the number of products sold is fixed as a matter of history. Others have also argued that using actual market prices (which already reflect demand and supply considerations in the marketplace) in conjoint instruments sufficiently accounts for supply-side factors. When rebutting conjoint analysis, it is crucial to explain properly the lack of foundation for these and other economically unsound assertions.
Analysis of Individual-Level Willingness-to-Pay Estimates Can Demonstrate Lack of Common Impact
Plaintiffs’ experts typically use willingness-to-pay measures averaged across respondents when estimating damages. Changes in aggregate measures of willingness-to-pay obscure differences in changes to willingness-to-pay at the individual level. An analysis of individual-level estimates can therefore demonstrate lack of common impact.
Reliability and Validity of Consumer Preference Data Generated by a Conjoint Study Should Be Carefully Assessed
Assessing the reliability and validity of a conjoint survey instrument and the data it generates should include determining whether the survey is properly designed and executed; whether the relevant population of consumers is targeted; whether a representative sample from that target population participates in the survey; whether valid economical and statistical models are used to analyze collected data; and whether the interpretation of the results is consistent with economic and conjoint analysis theory.
Because a conjoint survey should be able to replicate the real-world consumer purchase decision-making process as closely as possible to provide reliable estimates of consumer preferences, it is important to analyze the product features included in the conjoint instrument as well as how those features are described. Conjoint surveys that exclude features that are important drivers of purchase decisions are susceptible to biases well established in the academic literature and generate unreliable results. Similarly, features that are not described clearly and those that stand out in any particular fashion have been found to result in biased evaluations by respondents. Demonstrating the impact of these inappropriate survey design choices on willingness-to-pay estimates (and therefore demonstrating the unreliability of the conjoint study) can involve replicating plaintiffs’ conjoint study after fixing the aspect of the study that is suspected to most bias respondents’ evaluations. It is also important to analyze real market data, where available, as an additional test for the validity of conjoint results.
An Assessment of Analytical Tools in Product Liability Matters—Perspectives from Economics, Marketing, and Consumer Behaviour