by Paul Alan Levy
Earlier this year, I discussed a troubling decision from the United States Court of Appeals for the Ninth Circuit that reinstated a complaint by Multi-Time Machine, a maker of fancy watches, based on the way in which Amazon’s internal search engine returned results when users entered the search string “mtm special ops.” MTM has deliberately withheld its products from sale on Amazon, suggesting rather snootily that it regards Amazon as a purveyor of cheap goods, and used the litigation to try to force Amazon to tell searchers that it does not carry MTM watches instead of pointing consumers to competing products.
I argued that the court had reversed a recent trend of downplaying the doctrine of “initial interest confusion” in a way that threatened to enable trademark users to subject defendants to the expense of a trial simply by chanting the formula “initial interest confusion” and speculating about the way in which someone might experience confusion looking at the search results. The Ninth Circuit had recently receded from the most extreme approaches to initial interest confusion, and this decision threatened to retract that newer approach. Indeed, in an amicus brief in support of Amazon’s petition for rehearing en banc, Public Citizen and EFF jointly urged the Ninth Circuit to undertake a searching examination of the doctrine of “initial interest confusion” and undo the damage that its sixteen year old decision in Brookfield Communications had wrought.
Instead of leaving the case open for en banc review, the panel has now reversed itself in an opinion by Circuit Judge Barry Silverman, the author of the original dissenting opinion, and affirmed the lower court’s grant of summary judgment. In dissent, Circuit Judge Carlos Bea (the author of the now-discarded majority opinion) angrily charges that the majority “sub silentio overrules this court’s ‘initial interest confusion’ doctrine.” Would that it were so!
Treatment of Initial Interest Confusion
The majority avoids endorsing initial interest confusion, using that term largely in characterizing the plaintiff's argument without ever indicating that a finding of initial interest confusion might alone be sufficient to support judgment for a trademark holder or, indeed, to avoid summary judgment, but the panel does undercut one of its major empirical assumptions, that Internet users are fickle, unsophisticated and too easily fooled. Rather, the panel holds that Internet users are sophisticated enough to look at a set of search engine results and figure out what possibilities are being presented. And the panel confirms that a mere possibility of confusion is not enough – confusion must be "probable" or "likely," and the reactions of "unreasonable, imprudent and inexperienced web-shoppers are not relevant."
Indeed, although one might wonder whether the majority has pared the doctrine back in order to possibly save it, footnote 2 explains that overruling the doctrine wasn't on the table. The opinion noted amici such as Public Citizen that asked the en banc court to revisit the validity of the doctrine but notes that the parties did not contest the application of the doctrine and, in any event, that the panel is bound by existing circuit precedent. I take this footnote as an invitation to litigants to present this question in future panel appeals to preserve the issue for en banc review, and to file petitions on that issue.
The panel opinion bristles with other good news for the development of trademark law. Although the doctrine of initial interest confusion remains alive in the Ninth Circuit, the panel opinion now makes clear that it need not be an obstacle to summary judgment because the court itself can look at a set of search engine results and decide that there is no realistic likelihood of confusion. “Our case can be resolved simply by a evaluation of the web page at issue and the relevant consumer.” And it is the judges' evaluation that counts because "likelihood of confusion is not always a question of fact" and judges can evaluate a set of search results and find no likely confusion as a matter of law.
Irrelevance of the Multi-Factor Test
Even though the panel majority did not take up our invitation to discard initial interest confusion, its ruling largely follows a secondary point that we made in support of en banc consideration, endorsing the approach of the previous dissenting opinion. The new majority holds that the so-called Sleekcraft factors (the Ninth Circuit's version of the common multi-factor test of likelihood of confusion) need not be applied in every case. Indeed, there is some implication that, in the future, it may be confined to “analyzing whether two competing brands’ marks are sufficiently similar to cause consumer confusion.”
The panel opinion ends with a quick run through the Sleekcraft factors showing that they do not, as the previous majority opinion had suggested, support the plaintiff's claim that there is a disputed issue of fact about likely confusion. Unwittingly, it seems to me, this off-hand discussion of the factors just proves the argument of trademark scholars such as Barton Beebe, that Sleekcraft-type multi-factor tests do not provide a route to decision, but only a way of justifying decisions already made.