The decoy effect is the tendency to pick an option more often when a third, clearly worse option sits next to it.
On a pricing page, the decoy is a plan that is close to your target plan but obviously inferior, such as the same price for less, so the target looks like the easy, sensible pick.
It is also called the asymmetric dominance effect or the attraction effect.
The research behind it is real but narrower than most marketing blogs suggest.
It shows up most reliably when options are compared on clear numbers, which happens to describe most pricing tables.
Our A/B test database has a handful of plan tests that fit the pattern, and a few that warn against adding options at all.
Decoy pricing: how it works on a pricing page
Decoy pricing means adding or reshaping a plan so that it is beaten on every count by the plan you want people to buy, while not being clearly beaten by the other plans.
Nobody is meant to choose the decoy. Its job is to give buyers an easy reason to choose the target.
A simple example with three subscription plans:
| Plan | Price | What you get | Role |
|---|---|---|---|
| Basic | $19/month | 3 users, 10 GB | Competitor |
| Pro (limited) | $45/month | 8 users, 50 GB | Decoy |
| Pro | $49/month | 10 users, 100 GB | Target |
Pro beats Pro (limited) on every line for $4 more.
Basic is not beaten by either Pro plan, because it is much cheaper.
That one-sided comparison is the whole mechanism: the decoy is dominated by the target, and only by the target.
The best-known illustration is the subscription offer Dan Ariely describes in his 2008 book Predictably Irrational.
The offer listed web-only at $59, print-only at $125, and print plus web at $125. As commonly retold from the book, 84% of a group of MIT students chose print plus web when the print-only decoy was present, against 32% when it was removed.
Treat that as a classroom demonstration, not a field test: it was a hypothetical choice by students, and it is not a peer-reviewed study.
Decoy pricing vs. anchoring and the compromise effect
These three get mixed up, and they call for different tests.
- Decoy (attraction) effect: a worse version of the target makes the target look better. The decoy is dominated.
- Anchoring: a high price shown first makes later prices feel smaller. The anchor does not need to be worse at anything.
- Compromise effect: the middle option gains share simply because it sits between two extremes. Itamar Simonson's 1989 paper in the Journal of Consumer Research described both the attraction and compromise effects and linked them to people choosing the option that is easiest to justify.
A three-tier page with a highlighted middle plan usually mixes all three, which is why its test results are hard to pin on any single bias.
The asymmetric dominance effect: what the research shows
The effect was first reported by Joel Huber, John Payne and Christopher Puto in "Adding Asymmetrically Dominated Alternatives" (Journal of Consumer Research, 1982).
Their point was theoretical. Standard choice models assume that adding an option can never raise the share of an option already on the list. A dominated decoy did exactly that.
"Asymmetric" refers to the dominance being one-sided. The decoy is worse than the target on every attribute, but it is not worse than the competitor on every attribute, so it only flatters the target.
What later studies found
A 1995 meta-analysis by Timothy Heath and Subimal Chatterjee confirmed decoys can raise the share of the brand they resemble.
It also found the effect is uneven: decoys reduced the shares of lower-quality competitors more than higher-quality ones.
Then came a serious replication challenge in 2014. In "The Limits of Attraction", Shane Frederick, Leonard Lee and Ernest Baskin argued the effect may be restricted to stylized setups where every attribute is a number.
They found it typically did not appear when people experienced the product, or when even one attribute was shown as a picture, such as a hotel room photo.
In the same issue, Cornell researchers Yang and Lynn reported 91 attempts to produce the effect across 23 product classes.
Only 11 produced reliable effects, fewer than their statistical power predicted.
The original authors replied in "Let's Be Honest About the Attraction Effect".
They argued the effect stays robust when the original conditions are essentially replicated, and that the new papers had found moderators rather than disproved it.
What this means for pricing pages
Both sides agree on one boundary: the effect needs people to see, quickly and clearly, that the decoy is worse.
Pricing tables are close to the numeric setup where it holds best, since price, seats, storage and term length are all numbers.
The research does not tell you how large the lift will be on your page, or whether it will appear at all.
Lab studies measure hypothetical choices. Your buyers spend real money and can leave the page.
What A/B tests show about decoy pricing
We found no test in the LiftArchive database that isolated a pure, textbook decoy on a SaaS pricing table.
What we do have are plan tests that use the same logic, plus tests where adding or removing options moved revenue.
Tests that fit the decoy pattern
An Australian gin subscription brand, tested by Convertibles, is the closest match.
Control priced monthly and bi-monthly boxes the same, at $99. The variant raised bi-monthly to $109 and added a "Go monthly & save $10/box" callout.
Bi-monthly became a dominated option: fewer boxes for more money per box.
Monthly uptake went up, conversion stayed flat, and revenue rose by A$4,091 a month.
Rash ID, an AI skin condition app, went from two plans (monthly, yearly) to three by adding a weekly plan and marking yearly as the best offer.
The result was about +20% ARPU and over +50% conversion to paid.
The weekly plan may have worked as a decoy for yearly, but the badge changed at the same time, so the test cannot separate the two.
Planoly's pricing page test by CodaLift LLC gave the middle Growth plan a contrasting style and showed annual savings per year.
CTA clicks rose 16% on Starter, 20% on Growth and 74% on Professional, with +38% revenue and +24% AOV.
Note that the top plan gained most, which is not what a simple "push the middle" story predicts.
ArchiveSocial's redesign, run by VWO, simplified the plans and marked the mid-priced one most popular, for +64.76% leads.
It was a full redesign, so treat it as a pricing-page result rather than proof of any one bias.
Tests that warn against extra options
A decoy is an extra option, and extra options have a cost.
Mojo, a video editing app, hid AI credit packs on its paywall and showed subscriptions only. New subscription revenue rose 14%.
A party game app's paywall test by Vahe Baghdasaryan moved the monthly plan behind "View All Plans" and shortened the paywall, for +31% install-to-trial and +64% revenue.
Several things changed at once, but fewer visible plans was part of the winner.
See the paradox of choice for more on this side of the trade-off.
Plan order matters too. Microsoft 365 listed its pricier plan first and both plans sold more, the personal plan much more.
No numbers were shared, but it is a reminder that the plan you lead with shapes how the others are read.
ConvertiblesHow to test a decoy on your own pricing page
Run it as a proper A/B test, and decide in advance what counts as a win.
- Pick one target. Know which plan you want more people to buy and why, usually margin or retention.
- Make the dominance obvious. The decoy should lose to the target on every number shown, and the gap should be readable in a second. Hiding the comparison in a long feature list removes the effect.
- Keep it close to the target. The decoy should be priced and specced near the target, not near the cheaper plan, or it may flatter the wrong option.
- Change one thing. Do not add a badge, recolor the CTA and add a decoy in the same variant if you want to learn what worked.
- Measure revenue per visitor, not plan mix. A decoy can shift share to the target while lowering total sign-ups. Watch overall conversion, average order value and, for subscriptions, refunds and churn.
For layout ideas beyond decoys, see our pricing page best practices, the pricing plan tests and, for apps, subscription pricing tests.
When a decoy is the wrong choice
- When you have very little traffic. A shift in plan mix needs a large sample to read, and a small lift will be lost in noise.
- When buyers judge your product by look and feel rather than numbers. That is exactly where Frederick, Lee and Baskin found the effect disappears.
- When the decoy only exists to mislead. A fake plan nobody can actually buy, or one with hidden limits, crosses into dark patterns and can hurt trust and refunds.
- When your page already has too many choices. The Mojo and party game results suggest removing an option can beat adding one.
The decoy effect is a reason to test a plan structure, not a guaranteed lift.
For more biases with the same caveats, browse our psychology hub.