Ambiguity aversion
People prefer a known risk to an unknown one, even at a worse expected value.
Risk is when the odds are known and ambiguity is when they are not. People consistently prefer bets with known probabilities over bets with unknown ones, and will pay to remove ambiguity even when doing so lowers expected value. The preference is not reducible to pessimism about the unknown odds, since people avoid the ambiguous option on both sides of the same bet.
How it shows up in software
This is why opaque pricing kills conversion. 'Contact sales' on every tier, shipping costs revealed at the last step, usage-based billing with no estimator, and permission prompts that do not say what will be accessed all ask a user to accept unknown odds. It also explains why free tiers work: they convert an unknown outcome into a bounded one.
Using it well
- Publish prices, or publish a range with the variables that move it. A band with stated drivers beats a blank.
- Give usage-based products a live estimator and a hard cap the user can set, so the downside has a number.
- State what happens on failure before the user commits: refund terms, data retention on cancel, how long a migration takes.
- Replace vague permission copy with the specific data and the specific use, since 'improve your experience' reads as unbounded.
Where it turns manipulative
- Manufacturing ambiguity so the safe choice is your product: security and insurance marketing that implies an unquantified threat is selling the feeling, not the coverage.
- Hiding total cost until checkout is drip pricing, and it converts by removing the user's ability to compare rather than by being a better deal.
- Offering a paid 'certainty' add-on for something that should be standard, such as charging for a guaranteed response to a bug you caused.
Where you have seen it
AWS
The pricing calculator lets a customer model a configuration before committing, and Budgets can raise alerts at set thresholds.
Wise
Transfer quotes show the exchange rate, the fee, and the amount that will arrive before the transfer is confirmed.
Linear
Pricing is published per seat by tier rather than routed to a sales conversation at every level.
What the research says
- Ellsberg, 1961Well evidenced
Given an urn with 50 known red and black balls and another with an unknown mix, most people bet on the known urn for red and also for black, a pair of choices no single probability assignment can justify.
- Fox and Tversky, 1995Well evidenced
Ambiguity aversion was strong when ambiguous and clear bets were evaluated side by side and largely disappeared when each was evaluated alone, which located much of the effect in comparison.
- Trautmann and van de Kuilen, 2015Mixed evidence
A review across many designs found ambiguity aversion for moderate and high likelihoods but ambiguity seeking for low-likelihood gains, so the direction is not universal.
The effect is well replicated but conditional on the likelihood level and on whether options are compared. 'People hate uncertainty' is too broad a statement of 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.