court accepts survey universe of respondents who'd logically be interested in D's products despite possible overbreadth

Kodiak Cakes, LLC v. JRM Nutrasciences, LLC, 2022 WL 17340660, No. 2:20-cv-00581-DBB-JCB (D. Utah Nov. 30, 2022)

Extensive discussion of Lanham Act survey admissibility, finding this trademark survey admissible. Kodiak Cakes sells “Protein Power Cakes” products: pancake and waffle baking mixes with added protein. It currently has the second bestselling pancake mix brand in the United States by dollar sales. Kodiak Cakes sued JRM for trademark infringement and unfair competition based on its use of KODIAK SPORTS NUTRITION for nutritional supplements such as protein power.

Kodiak Cakes’ expert Prof. David Franklyn conducted two Squirt/Modified Lineup surveys, an Eveready survey, and a brand recognition survey, all four of which were run online two times.

The opinion usefully recites quality control measures for online surveys, including barring participants from taking the survey on a mobile device to ensure they had a clear view of the stimuli, removing any survey takers who entered gibberish into the open-ended responses were removed, and removing any survey takers who took any survey in less than half the median survey completion time or longer than four times the median survey completion time.

The survey screened for US respondents over 18 who had purchased protein powder or foods with added protein in the past 12 months or were likely to do so in the next 12 months.

Survey one showed respondents six images of products with added protein presented in random order: Oikos Triple Zero Yogurt, Kodiak Cakes Power Cakes, Protein Pretzel Sticks, Kodiak 1Whey, Quest Peanut Butter Cups, and Modern Table Mac & Cheese. In the control cell, the Kodiak name and bear logo had been removed from the shrink band around the cap of 1Whey. Respondents were asked whether they thought any of the products came from the same company/were affiliated and, if yes, why they said that, and could answer “yes” up to four times.

test

control

one display of products

In the test cell, 34% of participants indicated two or more products were from the same (or affiliated or connected) company. In the control cell, 22% of participants did so. But 19% of test cell participants linked the parties’ products, while in the control cell, only 2.3% did so, so Prof. Franklyn concluded that the survey showed a 17.5% level of confusion. [This suggests that, in the test cell, 15% of respondents linked products not in suit, while in the control cell,  19.5% did so, suggesting fairly high levels of noise since those products should have behaved the same in both cells.] In the test cell, 52 out of 53 respondents who linked the parties’ products explained that they thought the two products were related because of the presence of the word “Kodiak” on both products. In the control cell, no respondent explicitly said “Kodiak” as a reason for believing there was a relationship between the products.

Survey two had slightly different products and showed respondents six images of websites offering products with added protein. The questions were the same. The results were similar, with responses running a few points higher across the board; Prof. Franklyn calculated 19.7% net confusion.

Survey 3 was an Eveready survey in which respondents were asked to review an image of 1Whey protein powder; the control again had Kodiak and the bear logo removed. Respondents were asked “What company makes or puts out this product?” and why. They were also asked whether they believed the product is sponsored or approved by another company, is not sponsored or approved by another company, or “I don’t know or have no opinion” and, if yes, asked what other company and why, and asked the same question except with “business affiliation or connection.”

77% of respondents in the test cell responded “Kodiak,” while only 1% of respondents in the control cell did so. When asked what other products were made by the company shown in the image, 14.2% directly referenced products by Kodiak Cakes.

Survey four was a brand recognition survey with generally favorable results for Kodiak Cakes.

The court found that Prof. Franklyn was generally qualified and that the surveys were the product of reliable principles reliably applied. The proper universe is potential buyers of the junior user’s goods or services. While a survey that “provides information about a wholly irrelevant population is itself irrelevant,” usually “the selection of an inappropriate universe will lessen the weight of the resulting survey data, not result in its inadmissibility.”

While the survey targeted individuals over 18 years of age who had either purchased protein powder or foods with added protein in the past 12 months or were likely to in the next 12 months, JRM argued that this was inappropriately overbroad, because its products are protein supplements, not “foods with added protein,” and its target market is 18–35-year-olds. Although this was a logical criticism, in that foods with added protein appear to be a separate product category from protein supplements, consumers of foods with added protein seemed likely to be potential buyers of JRM’s goods given the similarities of the products. “Notably, had the survey screened for whether the potential participant had or intended to purchase protein products, rather than specifying ‘foods with added protein’ and ‘protein supplements,’ the screened participants would have remained consistent—but the hook for Defendants’ argument would have disappeared.” [I’m not convinced that’s true, since then the definition might have been overbroad, but ok.]

Ultimately the potential buyer universe did not need to be defined as granularly as JRM argued. It would have been overbroad to include “respondents who were not in the market for protein food products of any kind (for example, people interested in building muscle mass through exercise alone and not diet)” or “purchasers of food,” but that’s not what happened here. Nor was the survey underinclusive by excluding an entire category of defendant’s purchasers.

In contrast to “surveys that were so overinclusive as to drown out any probative value: there was a probability that none of the participants were potential consumers of the junior user’s products,” even if the survey included only past and potential purchasers of “foods with added protein”—the senior user’s goods—there was “a significant likelihood that that same market will be interested in the junior user’s goods: also protein-focused.”