Add a decoy tier and buyers pick your target. That is the most-copied pricing advice of the past twenty years. It does not work.

The advice traces to a single classroom demo. Dan Ariely stood at the front of an MIT lecture hall and showed three Economist subscription prices. Fifty-nine dollars for web only. A hundred and twenty-five for print only. A hundred and twenty-five for print and web. Nobody picked print-only. Eighty-four percent of the room picked the combo. When he pulled the print-only option, the combo dropped to 32 percent. One dummy tier seemed to steer the whole decision. That result launched a thousand pricing pages.

If you have built a pricing page in the past ten years, you have seen this example. Add a dummy middle tier. Make one option look so bad that the target looks good by contrast. Every SaaS blog, every pricing guide, every "just add a decoy" recommendation traces back to that slide in that lecture hall.

The advice sounds mechanical. It is not. It is decorative.

The research on this is worth reading, because it says the opposite of what most pricing guides claim.

The original finding came from Huber, Payne, and Puto in 1982. They called it asymmetric dominance. In a lab, when you add a third option that is clearly worse than one of two real choices, people shift toward the one that looks better. Clean result. Tight conditions. Two numeric traits on a grid. No real product names. No buying habits in play.

That is where it stayed clean.

Frederick, Lee, and Baskin tried to copy the effect 38 times. They published in the Journal of Marketing Research. The effect held only with bare numbers on a screen. The moment they added real product descriptions, the kind a buyer would read on an actual pricing page, the effect died. In some cases it flipped. People moved away from the target.

Yang and Lynn ran 91 attempts across 23 product types. Eleven produced a reliable effect. Eighty did not. Even the original authors responded in the same journal issue and admitted the limits of their own finding.

The decoy does not fail people. It fails the system they are trying to run.

A hundred-plus lab attempts told one story. The real market told it louder.

Devine and colleagues published a study in 2025 in npj Science of Learning. They tracked 3.6 million grocery store transactions. Real buyers, real shelves, real prices. The decoy shifted preference by roughly one percent. And the part that matters most: frequent buyers, the ones who shopped most often, showed little to no effect. The people who buy the most are the ones the trick does not touch.

Rafai and colleagues tested the same idea on a real digital marketplace. More than 140,000 search sessions on a flight booking site. No attraction effect found. None.

The problem is not that the decoy is a bad idea in theory. The problem is that the conditions it needs do not exist on any pricing page worth building. Two numeric traits. No real descriptions. Buyers with no prior experience. That is a lab bench. It is not your checkout page.

The evidence suggests something most people find hard to sit with. The dummy tier on your pricing page is not pushing anyone toward the target. It is just sitting there.

Stop trusting the decoy as a lever. Test whether the middle tier moves buyers up or splits their focus. Once that is clear, three moves follow from it.

Pull your own numbers. Look at tier conversion data for the past 30 to 90 days. What percent of buyers pick each option? If the middle tier draws less than five percent of sales, it is not a decoy. It is dead weight. This is the part most people skip, because the answer might confirm what they already suspect.

Run a removal test. Take the middle tier off your pricing page for two weeks. Do not announce it. Just pull it. Track what happens to conversion on the top tier. If the top tier holds or goes up, the middle was not doing its job.

Replace the dummy with a real offer. If the middle tier is not moving the number, kill it. Build a mid-range option with its own margin, its own value, its own reason to exist. Three tiers should each make money. Not two tiers and a prop.

Running this for 30 days does something the advice never did:

It shows you which tier is carrying revenue. It shows you whether buyers are choosing or just defaulting. It shows you where price sensitivity sits in your actual market, not in a lecture hall. And it tells you whether your page was built for your buyer or for a theory that needed a controlled room to work.

Your pricing page is not a lecture slide. A prop tier does not move buyers who have already compared and decided. Now you know what the dummy was doing all along: nothing.