consumer survey on patent value admissible despite serious flaws

Sentius Int’l, LLC v. Microsoft Corp., No. 5:13-cv-00825, 2015 WL 331939 (N.D. Cal. Jan. 23, 2015) (magistrate judge)

 

A little different today: this is a survey in a patent case trying to determine the value of a patented feature to consumers.  Despite Microsoft’s challenges, the court allows the survey, finding its flaws go merely to weight and not admissibility.  Sentius’s expert Wecker surveyed customers for preferences for spelling and grammar checking features in the accused products.

 

Surveys conducted according to accepted principles will ordinarily be sufficiently reliable under Daubert to be admitted.  “Unlike novel scientific theories, a jury should be able to determine whether asserted technical deficiencies undermine a survey’s probative value.” Once the survey is admitted, “issues of methodology, survey design, reliability, the experience and reputation of the expert, critique of conclusions, and the like go to the weight of the survey rather than its admissibility.” 

 

The survey asked customers about their preferences for the accused “check spelling and grammar errors as you type” (ed. note: yuck) features compared to alternative user-initiated spell- and grammar-checkers which Sentius’s technical expert identified as the best noninfringing alternatives.  Wecker asked direct, open-ended questions about consumers’ willingness to pay, but, recognizing that responses might overstate WTP, he adjusted his estimates by a “calibration factor.”

 

The survey asked respondents who had used and purchased an Office product and used the background or user-initiated features to suppose that Office included the user-initiated but not the check as you type options.  About 14.7% of spell checker respondents said they wouldn’t have bought the accused products without the background spell checker, and 14.8% for the grammar checker.  Wecker concluded that about 11.2% wouldn’t have bought the products without the accused features.

 

The court agreed that there were “significant concerns” about the structure of the survey and the way in which it was conducted, but these concerns went to weight rather than admissibility.  The survey questions were adequately tied to the subject matter of the asserted patents, and framing of questions is generally an issue of weight.  The description of the features in the survey did not vary so much from the patent that the survey failed to relate to any issue in the case. Microsoft argued that the questions didn’t distinguish between the patented and unpatented features of the background spell/grammar checkers—Sentius only claimed that certain subfeatures infringed.  But this wasn’t enough variation to make the survey unrelated to the case.  Microsoft also argued that Wecker didn’t ask questions about noninfringing alternatives that could implement the accused aspects of the background features.  Microsoft did offer evidence that it could have implemented several noninfringing alternatives, but Sentius could at least argue that its choice of comparator was correct.

 

In addition, the survey was based on reliable methodologies.  True, it asked respondents only about their preferences for the two accused features when the accused products include thousands of features, and thus didn’t accurately capture the value of the claimed inventions, but again this goes to weight.  Direct questioning was a generally accepted methodology even if it may have caused respondents to focus unduly on the accused features; a conjoint survey was not required for admissibility.  Using an “omnibus” survey asking a bunch of other unrelated questions likely impacted the quality of the survey, but that’s again weight.

 

Wecker’s use of a calibration factor to adjust the bias in his survey was also reliable.  He calculated his calibration factor based on a meta-analysis of 29 experimental studies. Microsoft argued that he didn’t show that this approach was generally accepted, but Microsoft failed to show that the presence of hypothetical bias/overvaluation of WTP was a reason to exclude the survey.  Given that, an expert’s adjustment for that bias couldn’t be a reason to exclude a survey either.