When And How To Use Conjoint Analysis For Patent Damages

Transcription

1 Portfolio Media. Inc. 111 West 19 th Street, 5th Floor New York, NY Phone: Fax: When And How To Use Conjoint Analysis For Patent Damages By Daniel Korczyk (January 18, 2018, 1:32 PM EST) Conjoint analysis continues to find meaningful application in the litigation arena. As recent as July of 2017, a Miami federal judge in the case of Kenai Batista v. Nissan North America gave final approval of a nationwide settlement where conjoint analysis had been prominent in awarding class certification.[1] In litigation settings, conjoint analysis is best known as a technique applicable to measurement of patent damages. Patent infringement cases frequently involve patented components that are part of a multicomponent system or device. The damages expert in a multicomponent infringement case must isolate the subject technology component s damage. This may entail determining the relative value contribution of the infringed component separate from other components and nontechnology drivers of product sales and profits. Daniel Korczyk Rather than relying upon the opinion of an individual or small group of arguably vested persons to establish relative component contributions to the overall product, counsel may be better served to commission a rigorous methodology that incorporates survey findings tied to opinions of unbiased product users.[2] To accomplish these objectives one appropriate damages methodology may be conjoint analysis, and there are several factors to consider in determining when and how to use this methodology. Conjoint Analysis Overview Conjoint analysis is a form of multiple regression analysis used for decades by marketing analysts to determine the product characteristics that drive consumers preferences for products.[3] This same methodology has application in litigation, particularly cases with potentially large damages. As with any multiple regression, the dependent variable (in this case product sales) is predicted based on a number of independent variables (i.e., logical sales drivers, including the infringed technology component). Predicted outcomes are then compared to actual historical sales to measure the strength of the multiple regression equation. Refinements may be necessary, but once the expert has a solid equation in hand, he can then apply it to measure the relative contribution of each independent variable. Conjoint Analysis Strengths Under the right conditions, conjoint analysis is appealing as a mechanism to help disaggregate royalty

2 rates consistent with relative contributions of technology components to a product s overall value. Part of the technique s appeal is its statistical nature, but also the manner in which its output easily dovetails with royalty rate damages assessments. Conjoint Analysis Facilitates Statistical Validity Checks Conjoint analysis relies upon survey responses as a source for multiple regression data. Using conjoint analysis, the expert is able to test whether his independent variables statistically explain sales. If this correlation is not present, then comparing the infringed component, as one of the sales drivers, to other attributes of the overall product is a meaningless exercise. Therefore, conjoint analysis provides the expert a basis for commenting on statistical validity. To help explain the conjoint survey design concept and its statistical validity, assume the only independent variables to the multiple regression analysis are technology attributes to the exclusion of price, advertising expenditures, number of sales distribution points, and product warranties. Under this formulation it may be determined that the resulting multiple regression equation is unreliable in explaining sales, and thus user preferences for the various technology components. To rectify the situation, a redesigned survey and additional sampling will be necessary. Once a strong regression equation is developed, conjoint analysis can quantitatively measure the contribution of each component attribute included in the analysis rather than simply providing a ranking. These measurements can then be coupled with royalty rate analysis to measure damages. Conjoint Analysis Supports the Royalty Rate Damages Computation When conjoint analysis is linked to royalty rate analysis, a widely accepted damages measurement technique, the expert is provided a valid basis upon which he can apportion the total technology royalty rate that a product might reasonably command.[4] The royalty analysis portion of the expert s work requires additional assessments such as: an examination of comparable royalty rates; the complexity of the subject component; the importance of the product to the overall product portfolio; pre-tax profit levels; levels of research and development expenditures; corporate rates of return; alternative technologies; applicable infringer sales; competitive factors; and supply/demand conditions.[5] A Simple Example By way of example, a conjoint analysis may inform the expert that buyers of the product ascribe value to each of three patented components, known as independent variables. Assume the following: Relative Technology Contributions to Sales 20 percent to Feature A;

3 50 percent to Feature B; and 30 percent to Feature C. If it can be demonstrated that an overall 10 percent technology royalty rate applies to the licensing of the product, and the infringed attribute is Feature C, then a 3 percent royalty may be applied to the applicable royalty sales base.[6], [7], [8], [9] A Few Words of Caution Courts have acknowledged the inclusion of full product sales as the royalty base in a damages analysis where the royalty rate is adequately lowered to reflect only the infringed technology component. Here conjoint analysis may be used to help determine the appropriate rate. However, caution must be exercised to understand whether or not the applicable royalty base is represented by full product sales, or whether some smaller base is appropriate.[10] Smallest Salable Patent-Practicing Unit Rather than changing the royalty rate in accordance with conjoint findings, it may be more appropriate to change the royalty sales base. Indeed, the federal court has indicated that royalties should be based not on the entire product, but rather on the smallest salable patent-practicing unit.[11] In other words, if the infringed component is periodically sold on a stand-alone basis, separate sales quantification attributed directly to the infringing technology component may be possible. Data Gathering Data gathering in the conjoint survey process is another area where strong attention is necessary to achieve supportable outcomes. Our experience indicates there are two broad categories of data gathering under the conjoint approach that can lead to challenges. First, sample questions may be biased, vague, confusing, lacking adequate depth, fail to tie to the facts of the case, or erroneous.[12] Equally troubling, survey respondents must have a bona-fide basis for replying to the survey. Sample size must be statistically adequate, and respondents must be selected in an unbiased manner. The bottom line is that respondents should be representative of the desired population. While randomness and lack of bias is important to building a pool of respondents, there is nothing to prevent the expert from staging his survey work so as to first screen out individuals that are not truly eligible to participate in the process of assigning preferences to the component attributes. Survey Timing May Prove to Be Problematic Beyond the process of data gathering, damages experts must also examine whether the survey results can be set forth to truly represent the willingness of parties to pay royalties at the time of infringement.[13] If, for example, a damages assessment is being made two years after the date of infringement, survey responses may be stale or tainted simply by the passage of time. How a respondent answers survey questions at a given point in time may not be the same as how they would have replied years earlier, particularly if intervening technological changes render relevant technologies partially or fully obsolete. What Do Courts Think About Conjoint Analysis?

4 As courts have changed the standard they apply to calculation of reasonable royalty damages, litigants have responded by turning to conjoint surveys.[14] These surveys have been used to establish the value of patented attributes in a number of recent high-profile cases, including Microsoft v. Motorola, Applev. Samsung, and Oracle v. Google. [15] Likewise, according to Sidak and Skog, U.S. courts generally consider surveys to reveal evidence of consumer preferences. Any shortcoming in methodology or survey design generally affects the weight, rather than the admissibility, of survey evidence. That general principle applies even in complex patent infringement cases. [16] Conclusion Where conjoint analysis assists the damages computations, experts must stand ready to defend their work with respect to a number of analytical and sampling issues. Key questions will focus on design of the survey, attributes and noninfringing attributes incorporated in the study, sampling design and rigor, use of constrained data, instructions and definitions given to survey respondents, whether survey pretests were conducted, and more. Because case law is central to arguing patent damages, the author believes that counsel should consult with its experts before a final damages measurement methodology is decided. If the method selected leans toward inclusion of full product revenues with conjoint analysis contributing to the quantification of the royalty rate, the expert must attend to the design of the conjoint survey. Survey attributes that make up the independent variables to the conjoint equation must include those elements thought to be predictive of sales, with one attribute being the infringed component.[17] Running the conjoint multiple regression analysis will tell the expert whether the product and nonproduct attributes, including the infringed component, are actually predictive of sales. Some amount of experimentation may be necessary to formulate independent variables that explain sales levels related to the subject product. Experts must find unbiased product users (buyers) who understand each of the attributes included in the regression equation and have a solid basis for weighting their importance to the purchase decision. Conjoint analysis in damages cases continues to evolve. Daniel J. Korczyk is a senior managing director in the Atlanta office of GlassRatner Advisory & Capital Group LLC. The opinions expressed are those of the author(s) and do not necessarily reflect the views of the firm, its clients, or Portfolio Media Inc., or any of its or their respective affiliates. This article is for general info rmation purposes and is not intended to be and should not be taken as legal advice. [1] Case No. 1:14-CV CIV-SCOLA/OTAZO-REYEZ [2] J. Gregory Sidak and Jeremy O. Skog, Using Conjoint Analysis to Apportion Patent Damages, The Federal Circuit Bar Journal, Volume 25, No.4, June 2016: To support their calculations of reasonableroyalty damages in patent infringement cases, experts increasingly introduce survey evidence to provide real-world evidence of consumers valuations of patented technology. [3] Paul Green and Yorman Wind, New Way to Measure Consumers Judgments (Harvard Business Review, August 1975); and Wind, Green, Shifflet, and Scarboroug, Courtyard by Marriott: Designing a

5 Hotel Facility with Consumer-Based Marketing Models (Interfaces, 1989). [4] Section 284 of the Patent Act requires that courts award patent damages that are adequate to compensate for the infringement, but in no event less than a reasonable royalty for the use made of the invention by the infringer. See 35 U.S.C. 284 (2012). [5] See Georgia-Pacific Corp. v. United States Plywood Corp., 318 F. Supp. 1116, (S.D.N.Y. 1970), modified and aff d, 446 F.2d 295 (2d Cir.) for 15 factors in determining a reasonable royalty rate. [6] 30 percent of 10 percent. [7] Example assumes the applicable royalty sales base is defendant s sales that include the infringed component. [8] I have simplified the example as many Features (also called Attributes) typically consist of multiple Levels. [9] Alternatively, where reasonable royalty analysis is not a primary damages approach, it may be possible to conduct the conjoint survey to directly assess buyers willingness to pay for the infringing feature. [10] An exception to partial sales treatment is where the Feature (or Attribute ) is shown to create the basis for consumer demand for the whole product. This principle of law is known as the Entire Market Value Rule. See: Paper Converting Machine Company v. Magna-Graphics Corporation, 745 F.2d 11 (Fed. Cir. 1984). Also see: State Industries, Inc. v. Mar-Flo Industries, Inc. 883 F. 2d 1573, Court of Appeals, Federal Circuit (1989). [11] LaserDynamics, Inc. v. Quanta Computer, Inc. 694 F.3d. 51, U.S. Court of Appeals, Federal Circuit (2012). [12] For example, questions may be erroneous if not directed to the appropriate industry where the subject technology is used. [13] I use the term royalty to also mean licensing fees and other forms of payment structures for the use of technologies owned by others. [14] Quoted: Lisa Cameron, Michael Cragg and Daniel McFadden, The Role OF Conjoint Surveys In Reasonable Royalty Cases, Law 360, New York (October 16, 2013) [15] Ibid. [16] J. Gregory Sidak and Jeremy O. Skog, Using Conjoint Analysis to Apportion Patent Damages, The Federal Circuit Bar Journal, Volume 25, No.4, June 2016, p.586. [17] Where attributes are independent variables in multiple regression terminology.