Regret aversion
People choose to avoid the feeling of having chosen wrongly, not only to avoid the loss.
Anticipated regret enters the decision alongside expected outcome. Because regret is sharper when the person caused the outcome and when the alternative is easy to picture, choices that could produce vivid counterfactuals get avoided. Regret about action and regret about inaction follow different time courses.
How it shows up in software
Regret aversion shows up as hesitation at irreversible steps: publishing, sending, deleting, committing to an annual plan. It also drives the opposite behaviour, where people over-research a low-stakes choice because picking wrongly would feel bad. Products that make choices reversible take the regret term out of the decision.
Using it well
- Make the risky step reversible and say so on the button's own screen. A visible undo window lowers the cost of deciding more than a reassuring paragraph does.
- Show what the person is choosing against at the moment of choice, for example a plan comparison on the checkout screen, so the counterfactual is answered rather than imagined later.
- Default to the choice that is easiest to change, not the one with the highest immediate value to you.
- For genuinely irreversible actions, slow the step down and state plainly what cannot be recovered.
Where it turns manipulative
- Manufacturing anticipated regret with fake scarcity or invented deadlines so the fear of missing out does the deciding.
- Confirmshaming on cancellation, where the decline option is worded as an admission of a mistake.
- Advertising a free trial as risk free while making the cancellation path slow enough that the reversibility is nominal.
Where you have seen it
Slack
Message editing and deletion stay available after sending, so posting in a busy channel is a reversible act.
Amazon
The returns policy is stated on the product page itself, next to the buy button, rather than only in help documentation.
Notion
Deleted pages go to a trash area and stay restorable, so removing a page does not read as a permanent decision.
What the research says
- Loomes and Sugden, 1982; Bell, 1982Mixed evidence
Independently formalised regret theory, showing that adding an anticipated regret term to expected utility accounts for choice patterns that standard utility theory cannot.
A model that fits anomalies rather than a direct measurement of feeling. Later experimental tests support the direction but not every prediction.
- Gilovich and Medvec, 1995Mixed evidence
Across surveys and interviews, short-term regrets were more often about things people did, while regrets over a lifetime were more often about things they did not do.
Mostly retrospective self-report, and the samples skew American and student-heavy. The time-course asymmetry has held up better than the specific proportions.
- Zeelenberg and Pieters, 2007Mixed evidence
A regret regulation theory setting out how people avoid, and recover from, regret in consumer decisions, including reversibility and justification seeking.
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.