Bernard Chao and Sydney Donovan have published a paper titled Does Conjoint Analysis Reliably Value Patents?, 58 Am. Bus. L.J. 225 (2021). Here is a link (the paper is behind a paywall, however) and here is the abstract:
Modern technology products are often covered by thousands of patents. Yet awards for a single component have averaged a surprisingly high 9.98% of the infringing product's price. To curb such disproportionate awards, the law insists that damages reflect the contribution made by the patent. But determining how to apportion damages in this way has proved to be elusive. One emerging technique that appears to offer rigor is conjoint analysis, a type of survey borrowed from the marketing world. This article explores the validity of the conjoint analysis technique by running two conjoint analysis surveys. Unfortunately, we found serious problems. First, the results of our surveys yielded irrationally high numbers. Most survey features suffered from bizarrely high valuations. Second, we demonstrate how experts can manipulate the results by selecting among a number of different ostensibly reasonable statistical choices and picking the one that yields the most desirable outcome. Based on these findings, we provide several recommendations. First, we argue that courts should not allow evidence of conjoint analysis to show the monetary value of specific features. However, we recognize that there is support for using conjoint analysis to provide relative valuations (i.e., feature A is worth significantly more than feature B). To the extent that courts permit this use, we suggest ways to ensure that experts employ the best science available. These recommendations include assuring that experts accurately depict variability in their results and requiring experts to “preregister” the approach they intend to use with the court.
I read this in draft last year and thought it was very good--check it out.