Eric Goldman thinks the beverage substitution hypo is past its prime:

For example, we need to retire the example of a consumer ordering Coke and the retailer silently delivering Pepsi instead from trademark law. This example is so distracting because we already know that the retailer can’t do this—at minimum, such a delivery is a breach of warranty and perhaps even a fraud. Trademark law does provide producers with a remedy (more on this in a moment), but from a consumer protection standpoint, what does trademark liability add to the analysis? Personally, I don’t think it adds much.

Sorry, Eric. I’m going to keep using it. And not just because it makes proponents of the trademark use doctrine uncomfortable. I’m going to keep using it because this is core trademark doctrine, and if we lose sight of it while arguing about search keywords and co-shelving and popup ads, we’re throwing out the baby while keeping the bathwater.

The retailer who delivers Pepsi when the customer ordered Coke is engaged in passing-off (or, for the more poetically inclined, palming-off). This is the central evil that trademark is meant to prevent. Whether you think that trademark is about preventing consumer confusion, or about encouraging investments in goodwill, or about protecting businesses from unethical competition, this kind of silent substitution is bad, bad, bad.

Even confining ourselves to consumer protection, this is a case in which we want competitors to have a remedy. We could call that remedy “trademark,” or we could call it “false advertising,” but we expect competitors to be substantially more vigilant in looking out for passing-off than governmental actors would be. If we’re prepared to throw in the towel on trademark law here, I humbly submit we ought to be prepared to throw it in everywhere.

The point of the beverage-substitution cases is simply to show that retailer silence can be all but indistinguishable from retailer speech in terms of its effect on consumers. It shouldn’t matter whether the waiter says, “One Coke, coming right up,” or not—and it doesn’t. That’s why I’m skeptical of the trademark use doctrine, at least as it’s applied to search engines and other online actors. Too often the analysis seems to turn on arbitrary characteristics of the technical configuration at issue; where do bytes representing the mark show up, and who sees them? 1-800 Contacts comes to mind as a case that may well be right on its facts, but still manages to draw a line that’s almost incoherent.

I don’t like most of the reasoning by analogy that’s dominating these debates. Offline metaphors are dangerous in reasoning about online trademark law because they make lawyers and courts get the facts wrong. This is like a billboard, they say, when it’s not like a billboard at all. But—and this is something that Eric has been very good at—we need to keep the offline analogues to online activities in mind, because otherwise, online trademark law becomes dangerously untethered, a collection of ill-justified sui generis rules that lead to inconsistent and unstable results.

Eric is exactly right that it would be a very good thing to have some better-understood limiting doctrines. He’s also exactly right that trademark use is more promising than many traditional trademark defenses precisely because it puts the burden on the plaintiff. And he’s even more exactly right when he says that “[T]he trademark use in commerce doctrine can serve the requisite channeling function, but it doesn’t do it very well.” And I’ll definitely support his call for the Goldman Act of 2025.

But don’t go trying to sub out beverage substitution, okay?