This class action lawsuit arises out of Amazon’s practice of using smart-speaker technology (“Alexa”) to surreptitiously: (a) intercept; (b) eavesdrop; (c) record; (d) disclose; or (e) use millions of Americans’ voices and communications, all without their knowledge or consent. Such conduct blatantly violates Washington’s wiretapping law, which applies nationwide to Plaintiffs and all members of the Class.
Defendants here, Amazon.com, Inc. and Amazon.com Services LLC (collectively, “Amazon” or “Defendants”), are therefore liable as a result of their egregious violations of the State Wiretapping laws – and are also liable for their violations of the Washington Consumer Protection Act (“CPA”), the Electronic Communications Privacy Act of 1986 (“Federal Wiretap Act”), and the Stored Communications Act of 1986 (“SCA”). Plaintiffs Kaeli Garner, Jodi Brust, Diane McNealy, Michael McNealy, Ricky Babani, Jeffrey Hoyt, Lorlie Tesoriero, Ronald Johnson, Selena Johnson and Caron Watkins (collectively, “Plaintiffs”) brought this action individually, and on behalf of a Class of similarly situated individuals, to redress those violations of law.
Plaintiffs sought to exclude Amazon’s expert, Dr. Dominique Hanssens‘ opinions because he is not qualified to opine on privacy issues or the law and because the surveys he conducted are neither scientifically valid nor relevant to the issues in this case.
Marketing Expert Witness
Dominique Hanssens is a Distinguished Research Professor of Marketing at the UCLA Anderson School of Management and has been on the UCLA faculty since 1977. Hanssens’ undergraduate degree in applied economics was earned at the University of Antwerp. His M.S. and Ph.D. degrees are in Management from Purdue University. His focus is on strategic marketing problems, and both his research and his legal consultations have involved surveys and analysis regarding consumer perceptions and their response to alternative information disclosures.
Discussion by the Court
A. Knowledge and Expertise
Plaintiffs argued that, because Hanssens’ expertise is in marketing, not privacy, he should not be permitted to assess consumers’ understanding of Amazon’s privacy policies. But Hanssens does not opine regarding how or whether consumers understood Amazon’s disclosures. Rather, he designed and conducted two surveys, one to see whether providing additional information related to Amazon’s use of Alexa voice recordings impacted consumers’ decisions to register an Alexa device (“Materiality Survey”) and the second to test consumers’ awareness of information regarding how Alexa works and their satisfaction regarding the ease of finding that information (the “Awareness Survey”). Hanssens has knowledge and expertise in consumer marketing and survey design sufficient to offer an expert opinion in this case.
Plaintiffs also objected to Hanssens’ opinion that “individualized inquiry is needed to assess a particular Proposed Class Member’s awareness of and attitude toward a particular Alexa feature.”
The objection was abandoned in reply and is overruled. Hanssens describes the data and evidence that led him to that conclusion, including his own survey results.
B. Materiality Survey
Plaintiffs argued that Hanssens’ Materiality Survey is irrelevant because it tests a factual scenario that does not apply to any potential class member, one in which the Alexa-enabled device has already been purchased and the consumer must now decide whether to register the device. Although it is possible that a consumer could review Amazon’s policies before purchasing a device, the scenario Hanssens tested appears to be the most common way in which a user becomes familiar with those policies.
Hanssens ultimately concluded that “the likelihood to register an Alexa Device was statistically indistinguishable between the Test Group (which was shown additional information about Amazon’s practices) and the Control Group.” While the survey results and the conclusion drawn therefrom do not definitively resolve the issue of whether consumers would alter their behavior in the face of effective disclosures, they do suggest that the specific alterations Hanssens tested would not impact consumer behavior. To that extent, they are relevant.
The problem is that the survey that generated the data underlying Hanssens’ conclusions is unreliable. Hanssens used a participant panel that was already predisposed to share information, excluded anyone who had never purchased or lived with an Alexa device (decisions that may correlate with a heightened concern for privacy), excluded anyone who was uncomfortable sharing personal information (a defect compounded by the fact that demographic information was requested at the beginning of the survey), and excluded anyone who took the time to read the disclosures. Having disqualified 95% of the respondent pool, many for reasons that could bias the responses to the main question of the survey, the results of the survey did not reliably reflect consumer reaction to the tested changes in disclosures.
C. Awareness Survey
Hanssens’ awareness survey was designed to test Plaintiffs’ contention that consumers are unaware that Alexa records, transcribes, and stores voice interactions even when the recordings are not intended for Alexa, that the recordings and transcriptions are stored and used by Amazon forever, and that human reviewers listen to and annotate the voice recordings. The survey also assesses whether consumers find these practices acceptable and whether consumers are satisfied with the availability of information regarding Alexa’s features. Plaintiffs argued that the survey results are irrelevant because it is based on consumer knowledge in 2024 and/or at the time they registered their first Alexa device (if in or after 2019).
While this choice means that the survey results reflect consumer knowledge that has been bolstered by five to ten years of disclosures and may not accurately reflect consumer knowledge when Alexa devices were first introduced, it did not make the results irrelevant. The proposed classes undoubtedly include individuals who first purchased their devices in the studied time frame. While the survey is not coextensive with the class period and cannot reflect the information known to all class members, the results shed light on consumer awareness of the practices covered by the survey.
Plaintiffs further argued that the Awareness Survey is unreliable because Hanssens did no research or follow-up questioning to confirm the respondents’ recollections, as represented in their survey responses. He did, however, limit the time frame of the questions to five years. As long as a “survey was conducted in accordance with generally accepted survey principles and that the results were used in a statistically correct manner . . ., technical inadequacies in the survey, including the format of the questions or the manner in which it was taken, bear on the weight of the evidence, not its admissibility.”
Held
The Court granted in part and denied in part Plaintiffs’ Daubert motion regarding Dr. Dominique Hanssens’ opinions.
Key Takeaway:
Shaky but admissible evidence is to be attacked by cross examination, contrary evidence, and attention to the burden of proof, not exclusion. That the opposing party can poke holes in a survey’s design and construction is not surprising: surveys are a scientifically constructed sampling method and, like any scientific method applied in the social sciences, there are bound to be limitations, restrictions, and flaws.
Please refer to the blog previously published about this case:
Computer Science Expert’s Testimony on the Value of Data Admitted
Privacy Expert’s Testimony on Alexa Users Limited
Case Details:
| Case Caption: | Garner V. Amazon.Com, Inc. |
| Docket Number: | 2:21cv750 |
| Court Name: | United States District Court for the Western District of Washington |
| Order Date: | March 30, 2026 |
Leave a Reply