The Decoy Effect: How a Third Option Changes Your Mind

This is one of the most confidently repeated claims in pricing advice, and its replication record is among the weakest of anything on this site.

What the effect claims

The formal name is the attraction effect, or asymmetric dominance. Introduced by Joel Huber, John Payne and Christopher Puto in 1982, it describes a choice between two options where adding a third — clearly worse than one of them on every relevant attribute, but not clearly worse than the other — increases the share choosing the option that dominates it.

The decoy is never selected. Its job is to make one option look obviously superior by comparison.

The theoretical importance was real. Standard rational choice theory holds that adding an option nobody picks should not change the relative preference between the other two. The attraction effect appeared to violate that directly, which is why it attracted forty years of attention.

The famous example, and its problem

The story everyone tells comes from a magazine subscription menu: web-only at $59, print-only at $125, and print-plus-web also at $125. Nobody takes print-only; its presence supposedly drives people to the bundle.

It is a memorable illustration, and it is an anecdote about a real pricing page rather than a controlled result. Whether that menu was designed as a decoy, and what it actually did to purchasing, are not established by its retelling.

That distinction matters more than usual here, because the experimental record turns out to be considerably less supportive than the anecdote suggests.

What happened under scrutiny

Two large research efforts published in 2014 tested how far the effect generalises.

Shane Frederick, Leonard Lee and Ernest Baskin found the effect largely confined to fully numeric, stylised stimuli. It typically did not occur when participants experienced the product, or when any attribute was perceptual rather than a number. Asked to choose between apples and oranges, adding a mouldy orange does nothing.

Sybil Yang and Michael Lynn ran 91 attempts across 23 product categories using realistic descriptions and images. Reliable effects appeared in eleven.

Both groups concluded that the conditions required to obtain the effect are restrictive enough to question its practical relevance — which is a strong claim from marketing researchers about a marketing staple.

Sometimes it reverses

The more awkward finding is that with realistic stimuli, the decoy occasionally reduced the share of the option it was supposed to help — a repulsion effect rather than an attraction one.

The proposed explanation is that a visibly bad option taints that region of the choice space, pushing people toward the opposite extreme on a rule like you get what you pay for. Evidence for the reversal is itself inconsistent, and nobody has established when to expect it.

For anyone applying this commercially, that is the crucial gap. The advice assumes a known direction. The research cannot supply one.

It is also worth being clear about what the 91-attempt figure represents. Those were not eleven successes and eighty failures in a sloppy sense — they were deliberate, well-powered attempts using realistic product descriptions and images, run by researchers who expected to find the effect. The design was a fair test, which is what makes the result informative.

What survives

Not nothing, and the defenders have a case worth reporting.

In abstract choices between options described by two numeric attributes, adding an asymmetrically dominated option does on average shift share toward the dominating one. That much replicates and is not seriously disputed.

Huber and colleagues responded to the failures by identifying boundary conditions: the effect requires that dominance be easy to detect, that neither core option already has a strong prior preference, and that the decoy sit at a plausible distance from the target — close enough to be comparable, far enough to look clearly worse.

That is a reasonable defence and it concedes the practical point. An effect that requires four conditions to be satisfied simultaneously is not a reliable tool for a pricing page, where none of them can be guaranteed.

What the failures do not show

Two things deserve stating for fairness, because the debunking can overshoot as easily as the original claim.

The effect has been successfully replicated in some studies using real prizes and real choices rather than hypothetical ones, so the failures are not uniform. And the theoretical point survives regardless of commercial utility: even an effect that appears only under restrictive conditions still demonstrates that preference between two options is not independent of what else is on the menu.

That was the original contribution and it stands. What does not stand is the leap from a laboratory violation of rational choice theory to a reliable lever for a pricing page.

The commercial appeal also has a structural cause. Advice that a middle option can be added to steer buyers is cheap to act on and impossible to disprove from inside a business, because a pricing change is never run as a controlled experiment. Revenue moves for a dozen reasons and the decoy gets the credit.

Why the advice persists

The decoy effect is genuinely useful as an idea. It is easy to explain, it flatters the person who spots it, and once you know the pattern you see three-tier pricing menus everywhere.

But seeing a middle tier nobody buys does not establish that it was placed there deliberately, or that it works. Menus have three tiers for many reasons, and the inference from a pricing structure to a psychological mechanism is exactly the kind of story that spreads without being tested.

This hub keeps arriving at the same shape. A real laboratory effect, a much stronger commercial claim attached to it, and a large gap between them — compare what actually happened to the choice overload finding, which went through the same process and came out somewhere similar.