Court Strikes Portion of Patent Expert Witness Report Based on Improper Methodology

Summary: The report of a Patent Expert Witness was limited in a copyright infringement case as to the apportionment of damages

Facts: This case (Oracle America, Inc. v. Google Inc. – United States District Court- Northern District of California –  March 13, 2012) involves a patent dispute. The plaintiff, Oracle America, alleges that the defendant, Google, used code that was copyrighted. Dr. Ian Cockburn wrote a damages report for this case which included five methods of apportionment. First, Dr. Cockburn reviewed the 2006 request for $99 million to Google. Second, he suggested adjusting the pay of the lawsuit to $557 million to match potential lost revenue. Third, he adds on another $28 million by removing a revenue-sharing cap. Fourth, he encourages the use of alternative methods including the “group and value” and “independent significance” approaches. Fifth, he downwardly adjusts because of potential extraterritorial infringement. After the damage report had been reviewed, Dr. Cockburn’s calculated patent damages stood between $18 and $56 million with the copyright lost license fee between $35 and $112 million. Google then acted to strike the report.

Discussion: The court evaluated Dr. Cockburn’s potential apportionment methods. Dr. Cockburn started with the “Group and Value” method which focuses on specific patents and copyrights with the payment from Google to Oracle equaling their importance to the function of the smartphone. Google opines that this was a biased approach because Oracle engineers gave each patent their value, though this ended up being a point for cross examination instead. Dr. Cockburn concluded that there were lower and upper bound patents with the upper bound patents carrying more weight, thus being more of value. The court took issue with the upper bound ranking and ended up striking it with the lower bound and overall group and value method being approved. Additionally, damages did not have to be apportioned on a claim-by-claim basis and Dr. Cockburn’s value distribution curve was allowed to be applied after Google refuted their use.

Dr. Cockburn then suggested another alternative apportionment method called the “independent-significance approach.” With this method in place, twenty-five percent of the 2006 license bundle would be applicable to the current patent claims being disputed. This approach is based on his expertise on the matter as well as information from 2006 regarding use of potential non-infringing alternatives to a Java license. Technical expertise on patent claims relating to Android was also considered. Finally, studies from 2011 on performance benefits of patent claims were noted. This approach was stricken and ruled to be flawed because the 2006 case studies were not of enough relevance to current technology. 

Dr. Cockburn used conjoint analysis as a method of surveying to dictate costs in the damages report. Consumer based surveys on varying smartphone features and their importance to the patents were used to figure out final costs. This survey was used to determine customer’s willingness to buy an Android phone with patented features by Oracle. Google argued that this method was unreliable because results did not match real world consumer preferences. Dr. Cockburn also used econometric analysis to determine Android’s market share as a result of infringement. He used data regarding if consumers would purchase another phone that was not an Android had it not been for the patented features. The court found that Dr. Cockburn failed to adjust for sales prices of Androids sold on Ebay and how one bidder may shift the next buyer’s willingness to pay an amount set by the first bidder. 

Conclusion: The court struck the independent-significance approach, the econometric analysis, the conjoint analysis determination of market share, and the upper-bound calculation in the group and value approach. The court allowed the conjoint analysis study on relative importance to startup time of applications and availability of applications and also the lower bound calculation in the group and value approach.