Google LLC v. Sonos, Inc.

Background

Sonos sued Google for patent infringement, and Google sued Sonos for a declaratory judgment. The related actions were consolidated for trial. The two remaining patents-in-suit, U.S. Patent Nos. 10,848,885 and 10,496,966, concerned customizing, saving, and later invoking overlapping groups of smart speakers or other “zone players.”

Sonos offered James Malackowski as its damages expert. He used the price of a subscription to IFTTT, a third-party scripting application, as the starting point for a hypothetical reasonable royalty negotiation. He relied on technical expert Kevin Almeroth’s opinion that IFTTT could be configured through applets to approximate some of the patented technology. From a $1.99 monthly subscription price, Malackowski derived approximately $90 million in damages: $12,246,294 for the ’885 patent and $77,546,923 for the ’966 patent.

Google filed two motions in limine, which are pretrial motions asking the court to exclude evidence. Its first motion sought to exclude the expert reports and testimony relating to the IFTTT-based damages theory. Its second motion sought to exclude narrower portions of Malackowski’s reports and testimony. Sonos opposed both motions.

The court allowed the witnesses to testify before deciding the motions. After their examinations, the court ruled from the bench that the IFTTT-based damages theory was unreliable. The jury was instructed not to consider IFTTT or the damages figures derived from it, although it could consider Malackowski’s testimony unrelated to IFTTT and other evidence, including admitted license agreements.

Ruling

To the extent stated in the opinion, Google’s first motion in limine was GRANTED. Google’s second motion in limine was DENIED AS MOOT. The opinion explains why the court excluded the IFTTT-based damages theory; it did not exclude all of Sonos’s damages evidence.

The jury ultimately awarded $35,507,183.40 for infringement of the ’885 patent and found no infringement of the ’966 patent.

Reasoning

The court applied the reliability requirement for expert evidence under Federal Rule of Evidence 702. Although calculating a reasonable royalty involves estimation, the underlying method must be reliable and tied to the facts of the case. The court identified three independent problems with Sonos’s method.

First, IFTTT could not serve as a benchmark product. A benchmark product must have technologically comparable features that allow the value of the patented feature to be isolated from the value of other features. In the court’s view, IFTTT did not, as sold, customize, save, and invoke overlapping groups of speakers. Users had to create multiple applets to produce a crude approximation, and the record did not show that anyone other than Almeroth and the Sonos team had used IFTTT for this purpose. The court concluded that speaker-grouping functionality was not an existing IFTTT feature.

Second, the later IFTTT subscription price could not serve as the starting point for the royalty calculation. IFTTT was free at the time of the November 2019 hypothetical negotiation for the ’966 patent, and its paid subscription plans were introduced in September 2020. The court also found that the subscription price covered many possible uses and therefore had no demonstrated relationship to what consumers would have paid for the patented speaker-grouping technology. The court further noted that the proof of concept used a contemporary version of IFTTT and that the record did not establish that the relevant speaker-grouping functions existed at the earlier negotiation date.

Third, the expert’s apportionments—the reductions intended to isolate the value of the patented invention—were not reliable. The court found that the 90-percent reduction based on using two of twenty applets did not account for the many other uses of those applets. It found that the 70/30 Google Play Store revenue split was not tied to the parties’ hypothetical patent negotiation, particularly because Google developed the Google Home application and would have contributed to the application’s other innovations. The court also found that the survey-based reduction did not adequately identify consumers who actually used or valued the patented feature and did not remove other groups of consumers who could not use the feature as accused.

Effect

The ruling removed the IFTTT evidence and the damages calculations based on it from the jury’s consideration. It left other damages evidence available, including testimony unrelated to IFTTT and admitted license agreements. The opinion memorialized the reasons for the earlier ruling for the record on appeal.