The decoy effect is probably the most repeated pricing trick in marketing: add a third option nobody wants, and people pick the expensive one. I went to the research to write a how-to. The research said something else.
The short version
The decoy effect (researchers call it the attraction effect, or asymmetric dominance) is real in the setting where it was discovered: hypothetical choices between abstract options described by two numbers. Outside that setting it is much weaker than its reputation. In 2014, one research team made 91 attempts to produce the effect across 23 product classes and got 11 reliable results.1 Another found the effect usually disappears when people see or experience real products, and sometimes runs backwards.2
That doesn’t mean a decoy never works. It means a decoy is a hypothesis to test on your own customers, not a law you can copy into your pricing page.
The famous decoy effect example: The Economist
Most people know the decoy effect through one story: The Economist’s subscription offer, popularised by Dan Ariely in Predictably Irrational. Ariely gave 100 MIT students three options, including a print-only subscription priced the same as print plus web. Nobody chose the print-only option, but 84% chose the bundle. When he removed the decoy, only 32% chose the bundle.5
| Option | Price | With the decoy | Without the decoy |
|---|---|---|---|
| Web-only subscription | $59 | 16% | 68% |
| Print-only subscription (the decoy) | $125 | 0% | Not offered |
| Print and web subscription | $125 | 84% | 32% |
It is a neat demonstration. It is also a classroom experiment with a hypothetical choice, not a record of what The Economist’s real readers bought. As far as we can find, the magazine never published sales data from that offer. That distinction gets lost every time the story is retold as proof that decoys sell subscriptions.
What the original 1982 study actually tested
The effect comes from a 1982 paper by Joel Huber, John Payne and Christopher Puto.4 Their question was theoretical. Classic choice models assumed that adding a new option can’t increase the share of an option that already existed. The researchers showed that adding an option that is clearly worse than one of the existing choices (dominated by it on every attribute) can raise that option’s share. That was an important finding about how people compare alternatives.
The options in that kind of study are deliberately stylised: a product described by two attributes, each a number, chosen hypothetically. That setup is the paradigm where the effect is strongest, and it is a long way from a real pricing page.
91 attempts, 11 reliable effects
In 2014 the Journal of Marketing Research published three papers on the effect in a single issue. In the first, Sybil Yang and Michael Lynn made 91 attempts to produce an attraction effect, using 23 product classes and 73 different decoyed choice sets. Only 11 produced reliable effects, far fewer than their study’s statistical power predicted.1
The detail that matters most for marketers: when the options were described with meaningful words or shown as pictures instead of bare numbers, the effects dropped to chance levels.1 Real pricing pages describe products in words and pictures.
With real products, the effect mostly disappears
In the second paper, Shane Frederick, Leonard Lee and Ernest Baskin tested the effect with products people could see or experience. They argued the effect may be limited to stylised products where every attribute is a number, and found it typically doesn’t appear when consumers experience the product or when even one attribute is shown perceptually, for example with a photo.26
In some of their tests the decoy did the opposite of what the textbook promises: it increased the share of the competing option, not the one it was meant to promote. Researchers call this a repulsion effect. A decoy that pushes customers away from your preferred package is worse than no decoy at all.
Want pricing pages built on your own data?
Our conversion rate optimization team designs package pages around real customer segments, then tests them on your traffic instead of borrowing a trick from a textbook.
Begin a ProjectWhat the original authors said
Huber, Payne and Puto replied in the same issue.3 They didn’t walk back their results. They argued the effect is robust when the conditions of the original study are replicated, accepted that Frederick and colleagues had identified real boundary conditions, and placed their 1982 work where it belongs: as a test of a rational-choice assumption and part of a wider argument that preferences are constructed in the moment, not simply revealed. They also set out the factors that weaken or strengthen the effect.
That is a careful scientific position, and it is very different from “add a decoy and watch revenue rise.” Nowhere do the authors present the attraction effect as a ready-made pricing tactic.
| Claim | What the evidence says |
|---|---|
| Adding an inferior option can shift choices in a lab | Yes, in the original setup: abstract options, two numeric attributes, hypothetical choices |
| It works reliably across product categories | No. 91 attempts across 23 product classes produced 11 reliable effects |
| It works when options are described in words or pictures | Mostly no. With verbal or pictorial descriptions, effects fell to chance levels |
| It works when people experience the real product | Usually not, and sometimes the decoy helped the competing option instead |
| The Economist's sales prove it | We could find no published sales data. The famous numbers come from a classroom experiment |
| The original authors meant it as a pricing tactic | No. They framed it as a test of how people make choices, not a playbook |
Decoy effect vs price anchoring
The two get mixed up, but they work differently. A decoy works through comparison between options: a worse option makes a nearby better one look like the obvious pick. Price anchoring works through a reference number: the first price someone sees, such as a crossed-out “was” price or a premium tier shown first, shapes how expensive the next price feels. Anchoring comes from Amos Tversky and Daniel Kahneman’s 1974 work on how people adjust estimates from a starting value.7
Anchoring has its own, separate body of research, and it is not what the 2014 decoy studies tested. The practical lesson is the same for both, though: a reference price or a premium tier is worth trying, and worth measuring on your own customers, not assuming.
Good, better, best pricing: what to do instead of a decoy
Most service businesses price in three tiers: good, better and best. Three-tier pricing works well when every tier is real. Here is how I’d set it up:
- Build every tier for a real customer. Each package should exist because a real group of customers wants it. A tier designed only to make another look good tends to confuse people, and confusion costs sales. Our guide to customer segmentation covers how to find those groups.
- Make the comparison easy. Clear names, a short list of differences and an honest “most chosen” label (only if it is true) help people decide. Readable comparisons are where the decision actually happens.
- Test, don’t assume. If you want to try a decoy, run it as an A/B test on your own traffic, and judge it on revenue per visitor and refunds, not just which package gets picked. Our conversion rate optimization guide explains how to set one up.
- Be careful with the source. When a tactic’s best evidence is a classroom exercise retold for twenty years, read the papers before you rebuild a pricing page around it.
Pricing tricks are tempting because they promise a result without changing the offer. The same lesson runs through our piece on the Black Friday discount trap: the durable wins come from a clear offer to the right customer, not from a lever pulled on the price list.
Sources
- Yang, S. and Lynn, M. (2014). More Evidence Challenging the Robustness and Usefulness of the Attraction Effect. Journal of Marketing Research, 51(4), 508-513
- Frederick, S., Lee, L. and Baskin, E. (2014). The Limits of Attraction. Journal of Marketing Research, 51(4), 487-507
- Huber, J., Payne, J. W. and Puto, C. P. (2014). Let's Be Honest About the Attraction Effect. Journal of Marketing Research, 51(4), 520-525
- Huber, J., Payne, J. W. and Puto, C. (1982). Adding Asymmetrically Dominated Alternatives: Violations of Regularity and the Similarity Hypothesis. Journal of Consumer Research, 9(1), 90-98
- Decoy effect: overview, including Dan Ariely's Economist subscription example from Predictably Irrational (2008)
- The Limits of Attraction (full text, Wharton)
- Tversky, A. and Kahneman, D. (1974). Judgment under Uncertainty: Heuristics and Biases. Science, 185(4157), 1124-1131
The three 2014 papers appeared together in the August 2014 issue of the Journal of Marketing Research. We summarise their findings in plain language; read the papers for the full methods and the authors’ own wording.
Frequently Asked Questions
What is the decoy effect?
The decoy effect, also called the attraction effect or asymmetric dominance, is the finding that adding an option that is clearly worse than one existing choice can increase that choice's share. In marketing it is usually used as a pricing tactic: a third package or price that exists mainly to make another look like the obvious pick. It was first shown by Huber, Payne and Puto in 1982 using hypothetical choices between options described by numbers.
How does the decoy effect work?
People judge options by comparing them rather than valuing each one alone. A decoy is placed next to one option and is worse than it on every point, so that option looks clearly better, while the comparison with the remaining option stays harder. The easy win against the decoy can pull choices toward the option it was designed to promote.
How effective is the decoy effect?
Much less effective than its reputation. In 2014, Yang and Lynn made 91 attempts to produce the effect across 23 product classes and got 11 reliable results. It is strongest with hypothetical choices between options described by numbers, and with real products, words or pictures it often disappears or even reverses.
Does decoy pricing actually work?
Sometimes, but far less reliably than marketing books suggest. In 2014, Yang and Lynn made 91 attempts to produce the effect across 23 product classes and found only 11 reliable effects. With options described in words or pictures, the effects fell to chance levels. Treat a decoy as something to test, not a guaranteed tactic.
Is The Economist subscription example real?
The numbers come from a classroom experiment Dan Ariely ran with 100 MIT students choosing hypothetically between three subscription options. With the decoy, 84% chose print and web; without it, 32% did. We could find no published sales data from The Economist itself, so the example shows the effect in a classroom, not in real sales.
Why does the decoy effect disappear with real products?
Frederick, Lee and Baskin found the effect may depend on stylised options where every attribute is a number. When people see or experience the product, or even one attribute is shown as a picture, the clean comparison that drives the effect breaks down, and the effect usually doesn't appear.
Can a decoy backfire?
Yes. In some tests by Frederick, Lee and Baskin, adding a decoy increased the share of the competing option rather than the option it was meant to promote. Researchers call this a repulsion effect. That is why any decoy should be tested on your own customers first.
Did the original researchers say the decoy effect is a pricing tactic?
No. In their 2014 reply, Huber, Payne and Puto said the effect holds when the original study conditions are replicated, accepted that later work found real boundary conditions, and described their 1982 study as a test of a rational-choice assumption about how people choose, not as a pricing playbook.
Should I add a decoy tier to my packages?
Not as a shortcut. Build each package for a real group of customers, make the differences easy to compare, and if you want to try a decoy, run it as an A/B test judged on revenue per visitor and refunds. A tier that only exists to steer people can confuse them and cost sales.
What is the difference between the decoy effect and price anchoring?
A decoy works through comparison: adding a clearly worse option makes a similar, better option look like the obvious choice. Price anchoring works through a reference number: the first price someone sees, such as a crossed-out price or a premium tier shown first, changes how expensive the next price feels. They are separate effects with separate research.
What is an example of the decoy effect?
The best-known example is Dan Ariely's Economist experiment: web-only for $59, print-only for $125 and print plus web for $125. With the print-only decoy, 84% of 100 MIT students chose print plus web; without it, 32% did. It was a hypothetical classroom choice, which is exactly the setting where the effect is strongest.
How should I structure good, better, best pricing?
Build each tier for a real group of customers, not to make another tier look good. Keep the differences between tiers short and easy to compare, label a tier as most chosen only if it is true, and A/B test changes on your own traffic, judging them on revenue per visitor and refunds rather than which tier gets picked.
How can you avoid falling for the decoy effect?
Ignore the option nobody would sensibly choose and compare the remaining options against your own needs and budget, not against each other. Ask whether you would still pick the same option if the extra one were not there. If you would not, the decoy is doing the choosing for you.
What is asymmetric dominance?
It is the technical name for the decoy setup. An option is asymmetrically dominated when it is worse than one existing option on every attribute but not worse than the other. The 1982 study showed that adding such an option can raise the share of the option that dominates it.
Add us as a preferred source and Google shows our guides more often in Top Stories and AI answers.
Let’s build what’s next.
Tell us where your brand is headed, and we’ll reply within one business day with where to start.