Add a third option nobody wants, and sales of one of the other two can rise. That is the decoy effect. Marketers use it on pricing pages and menus, and loyalty teams meet it in paid tiers and point packs. The research says it is real in some settings and absent in others, so a team should test it before building a program around it.
What is the decoy effect?
The decoy effect is a shift in choice toward one option when a worse third option joins the menu. It is also called the attraction effect. Joel Huber, John Payne and Christopher Puto described it in 1982. A decoy is dominated by one item in the set but not by another. Adding it can raise the chance that the item beating it is chosen (Huber, Payne and Puto, 1982).
The target is the option that beats the decoy on every dimension. The competitor is the option that beats the decoy on only some. The decoy is rarely chosen. Its job is to make the target easy to justify.
How does the decoy effect work?
The decoy gives buyers a simple comparison. Next to the decoy, the target wins on every count. Next to the competitor, the target wins on some counts and loses on others. People lean on the easy win.
Take three phone plans. Plan A costs $30 for 20 GB. Plan B costs $20 for 10 GB. Plan C costs $32 for 15 GB. Plan C is worse than A on both counts, since it costs more and gives less. It is not worse than B on data, since 15 GB beats 10 GB. A buyer torn between A and B finds A easier to defend once C is on the menu.
A 1995 meta-analysis found that decoys raised the shares of higher-quality brands. They typically did not raise the shares of lower-quality brands. In a large test, a traditional group of buyers matched that pattern, and a less traditional group reversed it (Heath and Chatterjee, 1995). So the effect depends on who is choosing.
Does the decoy effect really work?
It works in many lab tests and in some real settings, and it fails in others. The evidence splits by how the options are shown.
| Study | Setting | What it found |
|---|---|---|
| Huber, Payne and Puto (1982) | Lab choice tasks | Adding a dominated option raised the chance of choosing the item that dominates it |
| Heath and Chatterjee (1995) | Meta-analysis plus a large experiment | Effect found for higher-quality brands; reversed in one nontraditional group |
| Doyle and colleagues (1999) | Lab tasks and real in-store purchases | Effect shown in real purchases; called sizable and of practical use |
| Frederick, Lee and Baskin (2014) | Studies with numbers, tastes and photos | Authors contend the effect is limited to options shown as numbers |
| Huber, Payne and Puto (2014) | Reply to Frederick and colleagues | Argue the effect holds when studies are essentially replicated |
| Trendl, Stewart and Mullett (2021) | Movie posters, built to meet Huber’s criteria | A precisely zero attraction effect |
Frederick, Lee and Baskin argue that the effect can be limited to stylized products in which every dimension is a number. One example is a rating of 7.2 and another of 5.5 (Frederick et al., 2014). Huber and colleagues replied that the effect holds when a study is essentially replicated (Huber et al., 2014). The 2021 test with movie posters met the criteria Huber’s side set. It found no effect at all (Trendl et al., 2021).
What conditions does the decoy effect need?
Four conditions come up again and again. Wikipedia sums up Huber and colleagues this way. The buyer must be near indifferent between the target and the competitor. Both dimensions must matter about equally. The decoy cannot be too poor. The dominance must be easy to see (Wikipedia). The evidence adds two more. Show the options as numbers. And aim at the higher-quality option.
What is an example of the decoy effect?
The most quoted example is a subscription menu from The Economist. Wikipedia summarizes it from Dan Ariely’s book Predictably Irrational. The menu listed web access at $59, print at $125 and print plus web at $125. In a class of students, 84% chose print plus web, none chose print alone and 16% chose web. With print removed, 32% chose print plus web (Wikipedia, Decoy effect). Wikipedia marks those figures as needing a citation. Treat them as a classroom demo and not a sales result.
How can a loyalty program use the decoy effect?
The effect needs a menu with comparable options. Loyalty programs have three. Paid tiers are the clearest case, where a decoy tier carries the premium tier’s fee with fewer perks. Our tier strategy guide covers how to set the ladder. Point packs are the second. A pack should not cost more per point than a larger pack unless that is a deliberate choice. Bundled perks are the third, such as a hotel package priced near a richer one.
The research sets the conditions. Show the options side by side. Express the dimensions as numbers a member can compare, such as fee, benefit value and earn rate. Aim at the higher-quality option, since that is where decoys helped most.
Start with an audit, because accidental decoys exist. List each menu and look for an option that costs the same or more than another while giving less. Either remove it or decide that it is a deliberate decoy.
Keep trust in view. A tier that is plainly worse at the same price can read as a trick. Give every tier a real buyer. Read retention after purchase as well as the choice share on the page. Read the result by segment too, because one group in the 1995 experiment reversed the pattern.
How can a loyalty team test it?
Run a randomized test and judge it on margin. Log four numbers per arm. They are the share of buyers choosing each tier, margin per visitor, cost to serve and 12-month retention. A decoy that lifts the premium share but lowers retention has only moved the problem.
For illustration, a program shows 15,000 visitors a two-tier page and another 15,000 a three-tier page. In both, 4% buy, which is 600 buyers. Plus sells for $60 a year and costs $12 to serve. Premier sells for $120 and costs $45 to serve. The third tier, Plus Extra, costs $120 and includes fewer benefits than Premier.
On the two-tier page, 65% of buyers take Plus and 35% take Premier. The margin is 390 x $48 plus 210 x $75, or $34,470. On the three-tier page, 52% take Plus, 48% take Premier and none take Plus Extra. The margin is 312 x $48 plus 288 x $75, or $36,576. The decoy added $2,106 in margin per 15,000 visitors. Because the evidence is mixed, hold the result to a second test before changing the live page.
How GRAVTY handles the decoy effect
A menu of tiers is configuration. Our guide to paid loyalty programs says GRAVTY models a paid program as configuration. Paid tiers, their perks and the earning rules are authored in Visual Rules and go live without an IT ticket. Every earn, burn and benefit event is recorded against a member identity, so a team can compare cohorts. GRAVTY does not set the fee or design the benefits.
Frequently asked questions
What is the decoy effect in simple terms?
The decoy effect is when adding a worse third option makes one of the other two options more popular. The decoy is worse than one option on every dimension and better than the other on some. Huber, Payne and Puto described it in 1982 (Huber, Payne and Puto, 1982).
What is the difference between the decoy effect and anchoring?
Anchoring is the pull of a starting number on an estimate, and the decoy effect is a shift in choice caused by a worse third option. Tversky and Kahneman described anchoring in 1974 (the 1974 paper). Huber, Payne and Puto described the decoy effect in 1982 (the 1982 paper).
Is the decoy effect real?
It is real in many lab tests and showed up in real in-store purchases (Doyle et al., 1999). It is not universal. A 2021 test with movie posters found a precisely zero effect (Trendl et al., 2021), and other work finds it limited to options shown as numbers (Frederick et al., 2014).
When does the decoy effect fail?
It fails most often when attributes are perceived and not read as numbers. Frederick, Lee and Baskin report that it does not typically occur when people taste a drink or see a hotel room in a photo (Frederick et al., 2014). It also depends on the buyer, since one nontraditional group reversed the usual pattern (Heath and Chatterjee, 1995).
How do you use the decoy effect in pricing?
Add a clearly inferior option next to the one you want to sell, and show both with comparable numbers. Target the higher-quality option, where decoys raised shares in the meta-analysis (Heath and Chatterjee, 1995). Then test it against a two-option page and judge the result on margin.
