You know a survey's been done badly

... when it produces a negative net confusion rate.

Water Pik, Inc. v. Med-Systems, Inc. No. 12-1065 (10^th Cir. Aug. 29, 2013)

Highlights: when SinuCleanse, SinuSense, and NeilMed neti pot packages were shown to survey respondents, 29 of the 198 total respondents, or 14.6%, expressed confusion between SinuCleanse and SinuSense. But 30 respondents (15.2%) expressed confusion between SinuCleanse and the supposed control, NeilMed—producing a net negative confusion number.  Plaintiff did better with the other product tested, with net 6.5% confusion, but that wasn’t enough to create a factual issue on likely confusion.  (There were also flaws with the survey construction, but frankly, who cares?)

The court accepted a theory of “unconscious” confusion, whatever that means (apparently it means association, though association alone isn’t actionable), despite the relevant consumer’s testimony that she made a typographical error rather than being actually confused.  Still, this was anecdotal evidence and not sufficient to show likely confusion.

Also, the court rejected the district court’s analysis of intent, which was that intent favored the firstcomer because the secondcomer knew about SinuCleanse before adopting SinuSense as a trademark, and only developed its own product after an offer to buy SinuCleanse was rejected.  Instead of focusing on intent to copy the product, the proper inquiry is whether the secondcomer intended to deceive consumers about source—and there was no evidence of that, even though intent to copy a mark may justify an inference of confusion in the 10^th Circuit.  Knowledge plus intent to compete is not intent to mislead.