Decoy effect
Adding a clearly worse third option can shift which of the first two people pick.
Introduce an option that is worse than one existing option on every dimension but not worse than the other, and preference shifts toward the option that dominates it. The decoy is never meant to be chosen. It exists to make one neighbour look like an obvious win, which violates the assumption that adding an option should not reorder the others.
What a decoy set looks like
How it shows up in software
The classic move is a middle pricing tier priced close to the top tier so the top tier reads as the bargain. It also appears in storage upgrades, seat bundles, and add-on packaging. The honest version of this is much less exotic: most pricing pages are not running a decoy, they are running plain anchoring: The first number you see shapes every judgement that follows. and a compromise effect, and calling it a decoy flatters everyone involved.
Using it well
- If you are curious, run it as an experiment with a real holdout, and measure revenue per visitor rather than mix shift alone.
- Prefer removing a confusing tier over adding one. Fewer, clearly distinct tiers beat a set engineered for contrast.
- Make every tier something a real segment would buy. A tier nobody should ever pick is a tier you will end up supporting.
- Check whether your 'decoy' is doing anything at all. Mix shift with flat revenue usually means you moved people sideways.
Where it turns manipulative
- Listing an option priced so nobody chooses it purely to distort comparison is a form of misleading presentation, and regulators in the EU and UK have taken an interest in pricing structures designed to confuse rather than inform.
- Deliberately crippling a mid tier so it is unusable rather than differently scoped punishes the customers who do choose it.
- Repackaging tiers every quarter to keep the comparison unfamiliar makes the product harder to buy and harder to trust.
Where you have seen it
The Economist
The subscription page has historically listed digital-only, print-only, and a print-plus-digital bundle priced at or near the print-only price.
Apple
Device configuration pages list storage upgrades at fixed increments, so each step's marginal price is read against the step above it.
Cinema concessions
Three drink sizes are listed where the largest carries a small price gap over the medium.
What the research says
- Huber, Payne and Puto, 1982Mixed evidence
Adding an asymmetrically dominated alternative to a two-option choice set increased the share going to the dominating option across several product categories in hypothetical choices.
- Frederick, Lee and Baskin, 2014Contested
Across dozens of studies using real, familiar consumer goods and perceptual stimuli, the authors found the attraction effect largely vanished, and argued it depends on numerically described, unfamiliar attributes.
- Yang and Lynn, 2014Contested
A large set of replication attempts using varied designs failed to reproduce the effect reliably, reporting it as fragile and sensitive to procedure.
This is the weakest entry in this file and you should treat it that way. The effect is real in a narrow lab setup with abstract, numeric, unfamiliar options, and it repeatedly fails with real products people recognise. The famous Economist subscription story is a demonstration from a book, not a controlled result. If someone proposes a decoy tier, ask them to test it rather than assume it.
Grades are a judgement about the evidence, not about how useful the idea is. Plenty of contested effects are still worth knowing, as long as you do not cite them as settled.