Defaults
The preselected option wins far more often than the merits of the options explain.
A default is whatever happens when a person does nothing. People stick with defaults because switching costs attention, because the default reads as a recommendation, and because leaving it feels less like a choice they own. The size of the effect grows as the decision gets harder or less interesting.
Organ donor consent rates by registration regime
How it shows up in software
Defaults live in every checkbox state, every preselected plan, every notification setting shipped in the on position, and every sort order. Most users never open settings, so the default is the product for them. Teams often set defaults by accident, inheriting whatever the first engineer typed.
Using it well
- Write down, for each default, who it serves. If the honest answer is the company and not the user, change it or justify it in writing.
- Set the default to what a well informed user in the most common situation would pick, then measure how many people override it. Heavy overriding means the default is wrong.
- Keep the switch cheap. One tap to change, one tap to change back, no confirmation gauntlet.
- Review defaults when the user base changes. A default tuned for early adopters is usually wrong for the hundredth thousand user.
Where it turns manipulative
- Preticking consent for data sharing or marketing exploits inattention rather than preference. Under GDPR, preticked consent boxes are not valid consent, as the Court of Justice of the EU held in Planet49 (2019).
- Defaulting to the most expensive plan, adding paid extras to a cart, or turning on a paid upgrade at trial end converts inertia into revenue the user did not choose.
- Making the default easy to accept and hard to reverse turns a nudge: A change to how choices are presented that shifts behaviour without removing any option. into a trap. The test is symmetry: opting out should cost about what opting in cost.
Where you have seen it
iOS App Tracking Transparency
Cross-app tracking is off unless the user grants it in a system prompt. The permission state, not the app's own preference, is the default that governs.
UK workplace pensions
Employers must enrol eligible staff automatically at a set minimum contribution. Staying in requires nothing; leaving requires an active opt-out.
Signal
End to end encryption is on for every conversation with no setting to find. The safe state is the state you get by doing nothing.
What the research says
- Johnson and Goldstein, 2003 (Science)Well evidenced
Countries with opt-out organ donor registration had far higher registration rates than opt-in countries, and an online experiment reproduced the gap with the same people facing different presets.
This is about consent registration rates, not transplants performed. Actual donation depends on family consent, hospital infrastructure and donor eligibility, and the cross-country gap in real donation rates is much smaller than the registration gap. Quoting the registration numbers as lives saved overstates the finding.
- Madrian and Shea, 2001 (Quarterly Journal of Economics)Well evidenced
One firm switched its 401(k) plan to automatic enrolment. Participation jumped, and most enrolled employees stayed at the default contribution rate and default fund for years.
The second half matters more than the first. A default that sets a low contribution rate anchors people there, so a good default has to be right, not just present.
- Jachimowicz et al., 2019 (Behavioural Public Policy)Well evidenced
A meta-analysis of default effects across domains found a reliable but variable effect, larger for consumer and environmental choices than for savings and health.
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.