What confirmations prevent
Norman (1981) classified slips, where the intent is right but the action goes wrong, like tapping the photo next to the one you meant. A confirmation pauses users before such slips become permanent. Nielsen (2018) notes they work only for serious, irreversible actions. See error prevention design.
Overuse means nobody reads them
Böhme and Köpsell (2010) found users trained to accept repeated dialogs unread, and Anderson et al. (2015) showed with fMRI that brain responses drop with repetition. Egelman et al. (2008) found only warnings that truly interrupt change behaviour. So offer undo for routine actions and save confirmations for genuinely risky moments.
Anatomy of a good confirmation
Name the object and outcome, such as 'Permanently delete 12 photos?'. Label buttons with verbs like 'Delete' and 'Cancel', not 'Yes/No', and mark destructive buttons in red per Apple (n.d.). For very risky actions, require typing the name. On the web, follow the ARIA alertdialog pattern (W3C, n.d.) and move focus into the dialog. See modal dialog.
FAQ
- Confirmation or undo?
- If an action can be reversed, undo usually interrupts less; reserve confirmations for irreversible actions.
- How should buttons be worded?
- Use the actual verb, such as 'Delete' or 'Send', and visually distinguish destructive actions.
Sources reviewed: 2026-10-27
References
- Nielsen, J. (2018, February 18). Confirmation dialogs can prevent user errors (if not overused). Nielsen Norman Group. When confirmation dialogs help and when they are overused.
- Norman, D. A. (1981). Categorization of action slips. Psychological Review, 88(1), 1–15. Classic taxonomy of action slips.
- Böhme, R., & Köpsell, S. (2010). Trained to accept? A field experiment on consent dialogs. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (pp. 2403–2406). ACM. Users get trained to accept dialogs without reading.
- Egelman, S., Cranor, L. F., & Hong, J. (2008). You've been warned: An empirical study of the effectiveness of web browser phishing warnings. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (pp. 1065–1074). ACM. Which warnings actually stop users.
- Anderson, B. B., Kirwan, C. B., Jenkins, J. L., Eargle, D., Howard, S., & Vance, A. (2015). How polymorphic warnings reduce habituation in the brain: Insights from an fMRI study. In Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems (pp. 2883–2892). ACM. Brain habituation to repeated warnings.
- Apple. (n.d.). Alerts. Human Interface Guidelines. iOS guidance on destructive buttons and wording.
- World Wide Web Consortium. (n.d.). Alert and message dialogs pattern. ARIA Authoring Practices Guide. ARIA alertdialog role and focus rules.