Friday, January 21, 2011

A yummy survey case: Pom Wonderful again

POM Wonderful LLC v. Organic Juice USA, Inc., 2011 WL 70562 (S.D.N.Y.)

(An earlier post was misread, wonderfully, as “somebody should start the porn litigation blog.” I’d read that too! But one could keep busy with the Pom litigation blog as well.)

Pom alleged that Organic Juice violated federal and state laws by selling adulterated pomegranate juice, falsely labeled as "100% pure." Organic’s juice is produced, bottled, and labeled in Turkey, and sold under two varieties, 100% Organic 100% Pure Pomegranate Juice and Natural 100% Pure Pomegranate Juice. For background on Pom’s claims that its in-house production is more reliable than overseas production, see here.

Organic counterclaimed that Pom deceptively concealed the fact that Pom's juice is made from concentrate, made unsubstantiated health claims, and, from 2002 to 2008, added elderberry juice concentrate to its "100% Pomegranate Juice." Pom argued that the elderberry juice, added to the concentrate prior to reconstitution, was so miniscule as not to render the 100% label false under FDA regulations. The parties cross-moved for summary judgment on the counterclaims and Pom moved to dismiss several of the counterclaims. All motions were denied.

Other possibly relevant facts: Pom’s ads use its bottle, distinctively shaped like two pomegranates stacked on top of each other. The bottom bulb uses the words “from concentrate,” but the writing on the bottom bulb doesn’t appear in Pom ads, despite the consistent prominence of the bottle in those ads. Pom argued it removed that writing because it is "difficult to read when reproduced as a bottle image in print, and distracts from the main messages the advertisement is intended to convey." (I note that Google Images didn’t produce very good images of the writing on the actual bottles either, though of course many of the pictures I found were taken by nonprofessional photographers.)

Organic argued that the “from concentrate” was removed in an effort to deceive consumers. In 2003, a third-party marketing company produced, at Pom’s request, a document describing the results of focus group research relating to Pom's products. Under the heading "From Concentrate," the report states:
POM is "from concentrate," and this phrase is a severe negative. Fortunately, none of the respondents knew this, and the same undoubtedly is true of the market in general. Therefore, POM should be very careful to do nothing that would allow this knowledge to spread.
Pom responded that the focus group results played no role in Pom's advertising, noting that the bottle image used in its ads was the same before and after the focus group research.

Pom also touts its vertical integration in its marketing: a “tree to bottle” production method. Pom created a promotional video, which shows a worker cutting a pomegranate, then boarding a truck and adding his haul to other pomegranates. “Next, the video depicts pomegranates on a conveyor belt being washed and sorted, followed by some pomegranates being funneled into large pipes. Juice then flows from the pipes. The video then cuts to a machine molding plastic tubes into the distinctive shape of Pom's bottles. Next, the completed bottles move down a conveyer belt where they are filled with deep red pomegranate juice, capped, and date stamped. Finally, the full bottles are placed into cardboard boxes, presumably ready to be shipped to stores.”

The voiceover says great things about Pom and its processes, but, like the visuals, does not cover concentration, freezing, reconstitution, or pasteurization. (“Much of the healthy antioxidant goodness comes from crushing the whole fruit, husk and all. We even make and decorate our own distinctive bottles, shaped like two pomegranates stacked on top of each other. Then, we fill the iconic bottles with 100% natural pomegranate juice. … Once filled, the bottles are sealed and date coded to maintain freshness. They are sorted into cases, loaded into refrigerated trucks, and shipped to a store near you.”). Pom argued that the video was not meant to show the entire production process, but to highlight its vertical integration, focusing on those aspects of the production process that exemplify Pom's "Tree to Bottle" production method.

Organic offered internet surveys to test whether consumers understood from the ads that POM is made from concentrate. The surveys concluded that the print ads and the video were misleading, and that consumers believe that pomegranate juice not made from concentrate is better in terms of freshness, healthfulness, taste and purity, and that each of these categories is important to consumers considering whether to purchase a particular juice. Pom’s expert, whose qualifications, like Organic’s, were “unassailable,” critiqued the surveys.

Three surveys were done online through an internet survey platform, Participants had to be 21 or older, and had to have purchased pomegranate juice in the past year or "would consider purchasing it." The sample group was limited to New York, New Jersey and Connecticut.

The print ad surveys divided 200 respondents into test and control cells. The test group looked at an unaltered Pom print ad, while the control group looked at an ad that was edited to include the words "from concentrate" on the bottle. Respondents were instructed to "look at the ad and bottle [ed. note: already I see an issue with this question] as you would if you were considering buying pomegranate juice." They were then asked, "What does the ad say or suggest about this pomegranate juice?" and required to write in four responses. Then the close-ended question, "Based on what is said or suggested in the ad, how is POM Wonderful made?" The choices: (1) "From fruit without any processing," (2) "From concentrate without any processing," (3) "From fruit but with some processing," (4) "From concentrate but with some processing," (5) "From artificial flavors," or (6) "Other, please specify." And then: "Based on what is said or suggested in the ad, is POM Wonderful Pomegranate juice made from concentrate?" The answer choices were (1) "Yes," (2) "No," or (3) "Don't know/Not sure.”

Respondents were then asked to rank juice attributes (freshness, healthfulness, taste, purity, and color) from "Not at all important" to "Very important." They were then asked, "How does juice made from concentrate compare to juice that isn't from concentrate on each of the factors below?" The same attributes were ranked on a spectrum from "Concentrate is much better" to "Not from concentrate is much better."

For the open-ended question ("What does the ad say or suggest about pomegranate juice?"), none of the test group used the word "concentrate," while 5.5% of the control group did in one of their four responses. (This is an interesting design in that previous surveys have occasionally have been criticized for digging too deep with repeated “anything else?” questions; by requiring four responses from the get-go, you avoid that repeated cuing, but not necessarily the problem of forcing respondents to come up with things they wouldn’t otherwise have thought about.)

For the first closed-ended question ("Based on what is said or suggested in the ad, how is POM Wonderful made?"), 9% of test cell respondents answered "from concentrate," either with or without some processing. The corresponding percentage from the control group was 32%. For the second closed-ended question ("Based on what is said or suggested in the ad, is POM Wonderful Pomegranate juice made from concentrate?"), 19% of the test group answered "yes," as did 41% of the control group. (I’m interested in that 19%. Do some consumers have a background expectation that concentrate is the standard, such that the producer will tout fresh-squeezed status if it exists? Also notice how relatively low the pickup was from the explicit disclosure in the control ad, even though the control was designed by someone with an interest in making the disclosure work—well under half.)

Organic’s expert concluded that a statistically significant—Wait, stop! This is not the right concept!  What is so wrong with “significant”? Does it not sound scientific enough? Please stop doing this, lawyers! Ahem.  I'll come in again. Organic’s expert concluded that a “statistically significant,” by which the court most certainly means “practically significant,” number of consumers are likely to be confused by the ads. Respondents in the control group were more than 20% more likely to believe that the juice was made from concentrate than respondents in the test group. Moreover, both groups believed that pomegranate juice not from concentrate was better across the various attributes.

The video survey apparently lacked a specific control, but gave respondents the same questions. In response to the open-ended question, nobody mentioned concentrate. Eleven percent chose “from concentrate” in response to the first closed-ended question, and in response to "Based on what is said or suggested in the ad, is POM Wonderful made from concentrate?", 27% said yes, 71% said no, and 2% didn't know or weren't sure.

Pom’s expert Henry Ostberg thought that the survey suffered from fatal defects. First, it was limited to internet users part of a prerecruitied panel, even though the subject matter was not internet-related. Second, the proper universe was prospective purchasers, but past purchasers were also included. Third, the screening for who counted as a prospective purchaser was insufficient. Fourth, the “key questions of the interview were unclear, confusing and almost incapable of being answered.” Fifth, respondents weren’t offered a “don’t know” option for a key close-ended question, forcing them to pick even when they had no opinion. Sixth, the answer options weren’t rotated. Seventh, there were execution problems. Finally, the video survey additionally suffered from want of a control.

Here’s a survey design whoops: The test cell accidentally first answered the questions relating to the test ad, but then answered them a second time while reviewing the control ad. Though Organic’s expert excluded their responses to the control group questions from his results, Pom argued that there was nothing stopping respondents from going back and changing their initial answers once they’d seen both ads. (Again, if this happened, it’s an unintentional illustration of how bad most ads are at conveying more than a scrap of information, given the relatively modest differences between test and control; I would have thought that more respondents would have taken a harder look to figure out how the ads differed.) The parties, however, disagreed about whether it was technically possible for respondents to change their initial answers after seeing the second ad.

The court declined to resolve the issue, because even if test cell respondents could have changed their original answers after seeing the control ad, Pom benefited from the error. “Presumably, the ‘test’ group respondents would change their answers to suggest that Pom is made from concentrate, because the only difference between the ‘test’ ad and the ‘control’ ad is the presence of the ‘from concentrate’ designation on the bottle featured in the ‘control’ ad. The result would be less discrepancy between the answers given by both groups, meaning that the surveys would show less evidence of consumer confusion.”

I’m not convinced. If I was taking part in a survey and saw two ads that differed only in the “from concentrate” label, I’d be more inclined to say that the first ad was trying to communicate that it wasn’t made from concentrate, by contrast with the second. In other words, the court seems to be assuming that respondents would use the second ad to figure out what the first was trying to say, but that’s not what contrasts usually mean—the contrast highlights difference, especially in the context of marketing surveys, which often ask respondents to evaluate multiple pitches or multiple sellers. I’m not sure what you could get out of looking at the accidentally acquired test-cell-reaction-to-control-ad data, but if it turned out that those respondents answered “from concentrate” in response to the control ads at a much higher rate than the respondents in the actual control cell, then that could indicate some contrast effect at work.

The court did concede that this error might evidence a lack of diligence in survey design/implementation.

Pom also argued that the ads shouldn’t have been visible to respondents while they were answering questions, converting the task into a “reading” test, and that the closed-ended questions were impermissibly leading. Moreover, Pom argued that the video was wrongly taken out of context; Organic’s expert admitted that he mistakenly believed the video was a TV ad, not from Pom’s website. Pom argued that taking the video out of context made the survey unreliable because the webpage text clearly showed that the video was meant to highlight the “tree to bottle” method, not the entire production process. (Presumably Pom would make the same argument with respect to its YouTube page?)

Following the general rule that most criticisms of surveys go to weight rather than admissibility, the court found that the surveys were not so flawed that they had to be excluded. “Pom's criticisms of the Maronick surveys are thorough and well-articulated, and Pom may very well convince a trier of fact that the surveys should be given little or no weight,” especially for the video survey. But the rules favor admissibility unless the probative value is substantially outweighed by the danger of unfair prejudice, confusion of the issues, or misleading the jury. Cross examination, presentation of contrary evidence, and careful instruction on the burden of proof are the appropriate means of attacking shaky but admissible evidence.

I think there’s a fascinating conceptual issue here which Organic has, so far successfully, circumvented with its survey evidence. In general, failure to disclose material information, even though it is material, is not actionable under the Lanham Act except in limited circumstances, mostly when the affirmative statements in the ad mislead about the omitted information.

Arguably, at least in the video which focuses on parts of the production process, the affirmative statements here implicate the “from concentrate” enough to trigger that exception. And I could even construct an argument about the print ads, though I’m not sure how I feel about that: in general, failure to disclose material information is not actionable because we understand that many things are material and allow advertisers to pick their areas of strength. It is indubitable that price is material, and yet the Pom ads are not deceptive for failing to disclose price. Arguably, however, the difference between “price” and “from concentrate” is that no reasonable consumer, asked what the ads said about Pom’s price, would feel willing to give an answer: Responses should cluster around “don’t know” or “nothing.” By contrast, consumers were relatively willing to answer questions about “from concentrate,” suggesting that they received some implication about that from the ad. However, full confidence in this story would require a control question about price, and probably ideally an option to say “the ad said nothing about this” along with just a “don’t know.”

The court then turned to Organic’s motion for partial summary judgment on its theory of falsity by necessary implication.

Though Organic brought NY and Lanham Act claims, the court treated the standards as the same, meaning that it applied the Lanham Act’s explicit/implicit falsity distinction to NY state law. I think this is the result of reflex/unwillingness to address the different features and history of state law, but there is no doubt that it’s common enough to be unremarkable, especially when the parties don’t bother to argue to the contrary. (So why do plaintiffs bother to plead state law claims? Your guess is as good as mine ….)

Organic argued that the video necessarily implied that Pom's juice is not made from concentrate, “because the video purports to depict the entire production process of Pom's 100% Pomegranate Juice, but omits any visual depiction or discussion regarding the fact that the juice is pasteurized, concentrated, frozen, and reconstituted before it is bottled.” Unsurprisingly, Organic relied on Coca-Cola Co. v. Tropicana Products, Inc., 690 F.2d 312 (2d Cir. 1982), which found literal falsity when the ad showed Bruce Jenner squeezing an orange into a glass, and then pouring the contents of the glass into a Tropicana carton, when in fact the juice was pasteurized (heated).

The court found this situation distinguishable. The “tree to bottle” video didn’t explicitly claim to represent the entire production process. “A reasonable alternative implication is that the video--which is just over one minute in length--depicts only those parts of the production process that highlight Pom's vertically integrated production model.” Organic would need to rely on its survey evidence.

Pom’s motion to dismiss the additional counterclaims (related to elderberry juice/health claims) was denied. It relied on substantially the same arguments the court had already rejected when granting Organic leave to amend: the court explicitly found that the amended counterclaims were not futile, meaning they could withstand a motion to dismiss. Thus, Pom’s motion was frivolous, and Organic was entitled to attorneys’ fees and costs incurred in connection with it.

1 comment:

  1. Anonymous11:07 PM

    I think courts use words like "statistically significant" mistakenly for two reasons. First, as you suggest, they want to sound like they know what they are talking about when they don't. Second, lawyers probably misuse the term when presenting materials to the judge. It also doesn't help that judges and lawyers typically have no statistical training.