Is Your Algorithmic Pricing A Lawsuit Waiting To Happen?

Imagine your marketing chief pitches a revolutionary new tool that automatically calculates the best price for whatever you sell, tailored to appeal to specific customers, adjusted second-by-second to account for industry conditions and competitors’ pricing decisions. An entire marketing department in a box, powered by AI. What’s not to like?

A lot, if you don’t ask the right questions. Algorithmic pricing engines are the hottest thing in marketing, but they’re also a hot item with antitrust lawyers and legislators. The Sherman Antitrust Act and state laws prohibit competitors from acting in concert to set prices, and many suspect that’s what’s going on inside AI-driven pricing services.

If your algorithmic pricing engine is just crunching numbers your company supplies plus widely available industry data like publicly posted prices, that’s probably OK. The trouble starts when the engine isn’t just using your internal data but competitors’ internal data as well. Lawyers call it “give to get” and it can be illegal, says Katherine B. Forrest, a partner with Paul, Weiss who concentrates on legal questions swirling around AI.

“If they give certain information, they’ll get back even better,” Forrest explained. “It’s the confidential information that’s getting people in trouble.”

Collection Is Not Collusion

The U.S. Justice Department settled a lawsuit against RealPages, a rent-setting service, claiming it algorithmically revived an old concept in antitrust law, the hub-and-spoke conspiracy. Competitors—the spokes—share confidential information with a vendor or industry organization at the center—the hub— with an understanding that everybody is sharing that information and using it to make coordinated pricing decisions.

It’s that collective understanding, called the “rim,” that makes the whole arrangement illegal. Greystar, the nation’s largest apartment landlord, paid $50 million as part of a $141 million industry settlement with the government that included an agreement to stop using “competitors’ competitively sensitive data.”

Is your pricing engine breaking the law? That’s a tough question to answer without drilling down into the details. The Ninth Circuit Court of Appeals dismissed a lawsuit against Cendyn, a Las Vegas hotel pricing engine, because the judges couldn’t see a rim or explicit understanding among competitors to price their rooms based on confidential information. Simply matching a competitor’s posted price, even when it looks like collusion, is called “parallel pricing” and is allowed under longstanding precedent in antitrust law.

“While antitrust law restricts agreements between competitors regarding how to compete, it does not require a business to turn a blind eye to information simply because its competitors are also aware of that same information,” the Ninth Circuit said.

Know Your Algorithm

The challenge for managers is prying enough details out of a pricing vendor to decide if its AI engine is following the rules. Is the black box crunching numbers other than the ones you supply, and is that information coming from competitors? Does the contract require you to follow its pricing recommendations? Are competitors subject to the same requirement? Perhaps most importantly, do you understand how this mysterious engine works?

“If there’s no hope of anybody explaining it, you’re potentially buying yourself a lawsuit you can’t easily get out of,” Forrest says.

The safest course is to keep a paper trail documenting the answers to the questions above and use separate vendors for analyzing data and making pricing recommendations, she says. Avoid package deals that bundle both services into one inscrutable box.

Managers must ask: “If we get in trouble, can anyone articulate why the algorithm isn’t acting as a price fixer?” she says. “Is there anybody who’s going to be able to say that’s not what’s happening?”