Judgment and automation
What would you do if you were Google?
You know that targeted keyword-driven ads will be useful to customers. In fact, your company’s survival depends on keywords working well. At the same time, you know that trademark holders seem to think that they own the string of letters that make up their mark - no matter in what context, or for what purpose, that string of letters is being used. You know this because you keep getting sued when companies get huffy that “their” string of letters is being used by their competitors to target ads. You’d like everyone to relax and compete on the merits (rather than on some pretended magic ownership of strings of letters in all contexts), but these matters get very tense very quickly. These companies don’t want to talk about the theoretical purpose of trademark law.
(By the way, if they’d listen you’d tell them that trademark law is really designed to protect against unfair competition, not uses of strings of letters in the abstract. But they’re not listening.)
So you, Google, have to figure out a way to get through the fire swamp of litigation without falling into the great hole of commercial irrelevance. What do you do? You’re feeling brave, so you convince a few courts that what trademark owners are entitled to worry about is uses of their marks in displayed ads. Not uses of their marks in keywords that trigger the display of ads. It would be nuts to eliminate uses of marks in keywords that trigger the display of ads (you argue elegantly) because the mark isn’t really being “used” in that context - and it certainly isn’t being used in a way that would be likely to confuse a consumer, because the consumer will just be seeing ads after typing in the keyword.
You can’t convince courts in Europe, unfortunately, so with a sigh you don’t allow the use of trademarked keywords there. But in North America you have a different policy:
When we receive a complaint from a trademark owner, we only investigate the use of the trademark in ad text. If the advertiser is using the trademark in ad text, we will require the advertiser to remove the trademark and prevent them from using it in ad text in the future. Please note that we will not disable keywords in response to a trademark complaint.
That’s pretty good, and pretty brave, but the policy has the problem that it might conceivably block perfectly fair uses of trademarks in ads. For example, comparative advertising might not be allowed - if Apple objected, an ad saying “Nokia Phones are Always Better Than the iPhone” wouldn’t fly because it uses the Apple trademark. Or parody advertising might be stymied. Google just can’t decide on a case-by-case basis whether a particular ad using a trademark is fine. Instead, it automates the process; if a trademark holder says “don’t allow use of my marks in ads,” that’s the end of the issue. Google blocks the use of the trademark in the text of ads triggered by particular keywords.
Recently, MoveOn got stuck in a whirlpool of inaccurate reporting about this automated process. Here’s Art Brodsky’s post on the issue. Someone at MoveOn had checked a box on the Google site saying “don’t use my trademark,” so as to avoid frauds and phishing scams. But that check-box-check meant that attack ads directed at MoveOn, or comparative advertising, or parodies would be blocked. When Rep. Susan Collins wanted to place an ad criticizing MoveOn for criticizing General Petraeus, the ad wasn’t allowed by Google. Big squabble and lots of finger-pointing ensued. MoveOn, alarmed, surprised, and chagrined, removed the barrier.
Who’s doing what to speech in this story? Google isn’t a state actor, so the First Amendment of the constitution doesn’t apply to it. It’s finding a way to make lots of fair uses of trademarks in keywords possible, and it’s avoiding the worst liability by blocking keyword use in ads when the trademark holder objects. It’s muddling through. And it has to automate the process - or so it claims - so that it’s not in the position of having to judge which keywords uses are okay and which aren’t.
But default settings, check-boxes, and automation of all kinds can lead to speech being restricted. What’s the right balance? On this one, I think Google is being pretty brave by allowing keyword use to trigger ads at all. They’re taking a position that keyword use - by itself - isn’t trademark infringement. This dispute shows that it’s very difficult to automate judgment.
What’s a fair use, in copyright or trademark? These are judge-run, fact-intensive disputes. Do you want to spend years litigating this question? Or would you rather make your best guess, strike a balance, and automate as much as you can?
What would you do?