Optimism bias
People expect good things to happen to them more than to comparable others.
When judging their own future, people rate their chances of positive events above average and their chances of negative events below it. The comparative version (better than others) is the best documented. The absolute version, where a person's own risk estimate is wrong against a known base rate, is harder to establish and easier to dispute.
How it shows up in software
Optimism bias is why users skip backups, ignore security warnings, and decline the insurance they later want. It also drives signup for products bought against an intended future self: gym memberships, learning apps, and finance tools all capture people who genuinely expect to use them more than they will. Internally it shows up as launch forecasts that assume adoption curves nobody in the category has achieved.
Using it well
- Make protective behaviour the default rather than an informed choice, since a warning is competing with a belief about one's own luck.
- Show the user their own data instead of a population risk: unbacked-up files, reused passwords, actual sessions last month.
- In forecasting, require a comparable product's adoption curve as a reference before the plan's curve is accepted.
- Offer a downgrade prompt when usage does not match the plan someone bought, and accept the revenue hit.
Where it turns manipulative
- Selling annual plans to people who will not use them, then making cancellation hard, monetises the optimistic self against the realistic one. This is the core of the FTC's case for easy cancellation.
- Auto-renew without a usage-based reminder profits from the gap between intended and actual use.
- Marketing that promises outcomes for a typical user while only being achievable by the top few percent.
Where you have seen it
Apple iCloud
Device backup is offered during setup and runs automatically rather than requiring the user to initiate it.
1Password
Watchtower flags the user's own reused and breached passwords rather than reporting general breach statistics.
Strava
Weekly summaries report actual activity against the user's own set goal.
What the research says
- Weinstein, 1980Mixed evidence
College students rated themselves more likely than their peers to experience positive life events and less likely to experience negative ones, across a long list of events.
- Sharot, Korn and Dolan, 2011Contested
Participants updated their risk beliefs more when given better-than-expected news than worse-than-expected news, an asymmetry the authors linked to specific prefrontal activity.
- Harris and Hahn, 2011Contested
Argued that apparent unrealistic optimism can arise as a statistical artifact from rare events, scale attenuation, and base rate regression, without any underlying bias.
Take this critique seriously. The comparative finding is robust as a description of what people say, but the claim that people hold systematically miscalibrated beliefs about their own absolute risk is contested. Shah et al., 2016, also failed to replicate the optimistic update asymmetry under tightened controls.
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.