In this patent infringement action between Finjan and Sophos, the district court had previously granted a motion to exclude Finjan's damage expert. The district court explained that the expert's, Layne-Farrar, "method of applying a royalty rate to an apportioned base for each patent and adding the resulting royalties was not reliable because under her apportionment method she had attributed the full value of certain features of Sophos's products to multiple patents and so had counted the value of these features multiple times."

The district court concluded that this methodology "functioned to inflate Layne-Farrar's royalty base and her final damages calculation." Nonetheless, the district court permitted the expert to submit a supplemental report. Sophos then moved to exclude the supplemental report as well.

The district court explained that the amended methodology, which used a volume discount royalty rate method, suffered from the same problem. "While the volume discount royalty rate reduces the overall damage calculation, it does not resolve the issue I previously identified as it continues to attribute the full value of certain features to multiple patents."

The district court then provided an example of the problem: Layne-Farrar opines that a total of 5/7 modules of Sophos's UTM product are infringed by Finjan's patents. Assuming each module has equal value, Finjan's patents, at maximum, account for 5/7 the value of the UTM's total revenues. The total revenues are $21.5M so the total possible apportioned revenue base for Finjan's five patents, using this infringing module method, is 15.3M. Layne-Farrar also opines that a 16 percent royalty rate is appropriate given Finjan's past licenses. Applying this royalty rate to the apportioned royalty base is (15.3M)(.16) = 2.44M. Based on Layne-Farrar's own methodology, this should be the highest possible royalty for Finjan's five patents on the UTM because this reflects the maximum apportionment base for all five patents and is the highest royalty rate she uses.

The district court explained the problem with the methodology as "Layne-Farrar calculates a 5.6M royalty. She does this by calculating an individual royalty for each patent and then adding the royalties for all of the patents together. Layne-Farrar calculates an apportionment base for each patent based on how many modules that patent covers. She then applies the 16 percent royalty rate to the patent with the lowest apportionment base to calculate the royalty for the first patent. Next she applies a reduced royalty rate of 14.5 percent to the apportionment base for the patent with the next lowest royalty base to calculate her royalty base for the second patent. She continues in this way using a declining royalty and applying it to each patent to calculate individual royalty rates and then adds the royalties together to get her final royalty of 5.6M."

The district court concluded that the amended opinion should be excluded as well because it inflated the royalty number. "While Layne-Farrar's new opinion and calculation is more complex than before, the issue with it is the same. She continues to attribute the full value of certain modules of the Sophos product to multiple patents by including their full value in each patent's individual apportionment base. As discussed before, this is improper; if each patent is valid, then each cannot contribute the full value of a single module."

The district court further explained that "[t]here is nothing wrong with attempting to calculate an individual royalty for each patent. The problem here is that Layne-Farrar has not sufficiently apportioned the value of the various modules to each patent to calculate a per patent royalty without double-counting the modules' value. Layne-Farrar could have attempted to further apportion the modules to each patent."

The declining royalty rate also could not save the opinion. "While this reduces her overall royalty calculation, her total royalty still exceeds the highest royalty possible assuming a maximum 15.3M base and 16 percent royalty rate. While Layne-Farrar's precise calculation is difficult to track, because her ultimate royalty is significantly higher than the maximum possible royalty for her figures, it is clear that she must be double counting the revenues for certain modules. The problem with Layne-Farrar's prior methodology was not her ultimate royalty calculation, but the method she used to calculate it."

As a result, the district court excluded the amended opinion.

Finjan, Inc. v. Sophos, Inc., Case No. 14-cv-01197-WHO (N.D. Cal. Sept. 8, 2016)