The Option You Were Never Meant to Buy
The third option on a pricing page is often there to lose. The effect is real, it is narrower than the marketing books say, and the place it works best is exactly where modern shopping happens.
Key takeaways
- Huber, Payne and Puto showed in 1982 that adding an option nobody picks can raise the chance of choosing the option that beats it, which standard choice models say should not happen.
- The paper itself pointed at the commercial use, suggesting product line strategies that would not otherwise seem plausible.
- Frederick, Lee and Baskin found the effect is largely restricted to stylised products where every attribute is a number, and typically does not appear when people taste, hold or see the product.
- They question the practical significance of the effect, which is a serious caveat almost never carried in the retellings.
- It does survive where attributes genuinely are numbers, which is where subscription tiers, insurance excesses and broadband plans live.
Open any software pricing page and count the tiers. There will be three. One of them will be worse than the tier above it on every line of the comparison table and cheaper by an amount that looks small. Almost nobody buys it, and it is not there to be bought.
Dominance
In 1982 the Journal of Consumer Research published a short paper by Joel Huber, John Payne and Christopher Puto with an odd finding in it. Take a choice between two products where each beats the other on something: one is cheaper, the other is better. People split. Now add a third option that is clearly worse than one of the first two and not clearly worse than the other. That third option is [asymmetrically dominated](https://thewhyweproject.com/articles/what-is-the-decoy-effect). The decoy effect, and how well it has replicated, is the subject of What is the decoy effect. Very few people pick it.
But its presence increases the probability that people choose the option that dominates it.
That result is stranger than it first looks. Standard choice models say adding an option can only take share away from the existing ones. Here, adding a loser gives share to a specific winner. Huber, Payne and Puto were explicit about the commercial implication in the paper itself: it suggests product line strategies that would not otherwise seem plausible.
Which is why the pricing page you looked at this week has three tiers and why one of them is there to lose.
Mechanism
The usual explanation is comparison. Two options that beat each other on different things give you no easy way to decide. A third option that is worse than one of them on every dimension creates a comparison you can complete, and completing it produces a winner, and that winner then carries the momentum into the real decision.
Evidence on attention supports something like this. Michal Król and Magdalena Król tracked eye movements during incentivised choices and argued that the decoy's inferiority to the target, not its resemblance to it, drives the transfer of attention behind the effect.
Time pressure appears to matter too. Jonathan Pettibone tested four levels of it against both the asymmetric dominance effect and its cousin the compromise effect, and reports both as larger under greater pressure. The result is one study and the two effects did not behave identically across his manipulations, so treat it as a direction rather than a measurement. It is worth knowing anyway, because every commercial setting that uses a decoy is also built to hurry you.
Limits
Now the part the marketing books leave out, and it is a serious one.
Shane Frederick, Leonard Lee and Ernest Baskin looked at the shape of the evidence and noticed something. Nearly every demonstration of the attraction effect uses stylised products where every attribute is a number: a toaster oven with a durability rating of 7.2 and an ease-of-cleaning rating of 5.5. When people actually experience the product, by tasting a drink for instance, or when even one attribute is shown perceptually, such as hotel rooms whose quality is conveyed by a photograph, the effect typically does not appear.
Their conclusion is stated plainly in the paper. They question the practical significance of the effect.
That is not a small caveat, and it cuts against the examples everybody reaches for. The cinema popcorn story, the one told in every book on this subject, is about a good you are about to taste, bought at a physical counter, and it is exactly the case Frederick and colleagues say should not produce the effect. We have no evidence that it does.
What the critique leaves standing is the decoy on spreadsheets, which is where the demonstrations come from.
It is worth naming what has just happened, because it is the test this publication says it applies and rarely shows itself applying. A 1982 finding, from the small-sample era, met a large critique four decades later and came out smaller. That is the replication process working, not a scandal, and the correct response is to keep the narrower claim rather than to keep the original one or to throw both away.
Where it has been shown again, and where it has not
The obvious next question is whether anybody has demonstrated the effect outside the spreadsheet. On the evidence we can find, no.
Ian Bateman, Alistair Munro and Gregory Poe took it into choice experiments and contingent valuation for non-market goods, and found that introducing an asymmetrically dominated option significantly affected choices between the non-dominated options, and that this carried through into the values people then placed on them.
That is a genuine extension of the stakes and not of the territory. Contingent valuation presents attributes as numbers in a table, which is the paradigm Frederick indicts, and the paper predates his critique by six years, so it cannot answer it. What it establishes is that the consequences reach further than retail, because the method is used to value public goods and set policy. What it does not establish is that the effect survives where Frederick says it fails.
So the honest map looks like this. Where you are comparing listed specifications, subscription tiers, insurance excesses, mobile plans, broadband speeds, storage sizes, the demonstrations apply and you should assume the layout is deliberate. Where you are choosing by looking, tasting or holding, the evidence that it works is much thinner than the confidence with which it is sold.
The verdict, and one inference we are less sure of
The attraction effect is real, narrower than its reputation, and concentrated precisely where modern buying has moved.
That last point is the one we would emphasise, and it is an inference rather than a finding. Frederick's critique lands hardest on physical retail, where attributes are perceptual and comparison is messy. A great many consumer decisions now happen on a screen displaying a table of numbers, which is the environment the demonstrations used. Nobody has run the study on a subscription pricing page, so what we are offering is a reason to be suspicious of one and not evidence about it.
What we would stop saying is that a decoy can sell you anything. It cannot. It nudges a choice between two options you were already weighing.
What we would keep is the diagnostic. Three tiers where the middle one is worse than the top on every dimension, and cheaper by an amount that looks small, is an argument wearing the clothes of a menu.
What it costs to ignore
If you buy the tier the decoy was pointing at and you wanted it anyway, nothing has been done to you. That is the honest position and it is why this is a mild manipulation rather than a scandal.
The cost lands elsewhere. A pricing page built this way is designed to stop you asking the question that actually matters, which is not which of these three, but whether any of them is worth it. The decoy works by making a comparison easy, and an easy comparison is a comfortable place to stop thinking.
So the practical move is small. When you meet three options and one of them is obviously pointless, cover it and look again at the two that are left.
Sources
Huber, J., Payne, J. W., & Puto, C. (1982). 'Adding asymmetrically dominated alternatives: violations of regularity and the similarity hypothesis.' Journal of Consumer Research, 9(1), 90-98. Adding an asymmetrically dominated alternative to a choice set can increase the probability of choosing the item that dominates it; the authors note this points to the inadequacy of many choice models and suggests product line strategies that might not otherwise be intuitively plausible. doi:10.1086/208899
Frederick, S., Lee, L., & Baskin, E. (2014). 'The limits of attraction.' Journal of Marketing Research, 51(4), 487-507. The authors contend the attraction effect may be restricted to stylised product representations in which every dimension is a number, note that such effects do not typically occur when consumers experience the product or when even one attribute is represented perceptually, and question the practical significance of the effect. doi:10.1509/jmr.12.0061
Krol, M., & Krol, M. E. (2019). 'Inferiority, not similarity of the decoy to target, is what drives the transfer of attention underlying the attraction effect.' Journal of Neuroscience, Psychology, and Economics, 12(2), 88-104. Eye-tracking with incentivised rather than hypothetical choices. doi:10.1037/npe0000104
Pettibone, J. C. (2012). 'Testing the effect of time pressure on asymmetric dominance and compromise decoys in choice.' Judgment and Decision Making, 7(4), 513-523. Both decoy effects were found to be larger as time pressure increased. doi:10.1017/s1930297500002849
Bateman, I. J., Munro, A., & Poe, G. L. (2008). 'Decoy effects in choice experiments and contingent valuation: asymmetric dominance.' Land Economics, 84(1), 115-127. Introducing an asymmetrically dominated option significantly affected choices between non-dominated options and translated into significant impacts on subsequent valuations. doi:10.3368/le.84.1.115