Reminders and prompts
Most intended actions fail at recall, so a well timed prompt outperforms more persuasion.
A reminder closes the gap between intention and action by arriving when the action is possible. Planning prompts go further by asking the person to specify when, where and how they will act, which links the intention to a concrete cue. Both are cheap, both work, and both decay with repetition.
How it shows up in software
Reminders are push notifications, digest emails, calendar holds, in-app badges and the nudge: A change to how choices are presented that shifts behaviour without removing any option. a task manager sends at a location. Volume is the killer: each additional low-value prompt reduces the response to every other one, and eventually the user turns the channel off permanently.
Using it well
- Trigger on the user's own stated intent, such as a due date they set or a goal they named, rather than on a marketing calendar.
- Send at the moment the action is possible, and put the action in the notification so it can be done from there.
- Set a prompt budget per user per week, enforce it in code, and make teams compete inside it instead of each adding their own.
- Ask for a plan, not just an intention. Prompting the user to pick a day and time is a small change with a real effect.
Where it turns manipulative
- Reminders sent to hit a daily active user number rather than to help the user are engagement extraction, and the metric will look fine while trust falls.
- Guilt framing in notifications, especially for streaks and missed activity, uses a helpful mechanism to apply pressure a person did not ask for.
- Making notification settings coarse or hard to find, so the only way to stop is to uninstall, turns a prompt into an imposition.
Where you have seen it
Todoist location and time reminders
The prompt fires at a moment the user chose when they created the task, so the cue and the opportunity coincide.
NHS and similar appointment SMS
A short message with date, place and a reply option to cancel, which both reminds and enables the alternative action.
Duolingo notification cap
A limit on how many reminders a lapsed learner receives before the app stops, which treats attention as a finite budget.
What the research says
- Gollwitzer and Sheeran, 2006 (Advances in Experimental Social Psychology)Mixed evidence
A meta-analysis of 94 implementation intention studies found a medium-to-large effect on goal attainment over and above intention alone.
Largely self-reported outcomes in academic samples, and the era and methods make publication bias a live concern. Field effects are much smaller.
- Milkman, Beshears, Choi, Laibson and Madrian, 2011 (PNAS)Well evidenced
A mailing that asked employees to write down the date and time they would get a flu shot raised vaccination rates against an identical mailing without the prompt.
A randomised field experiment at scale. The effect was a few percentage points, which is the realistic size.
- Karlan, McConnell, Mullainathan and Zinman, 2016 (Management Science)Well evidenced
Savings reminders across three field experiments increased deposits, with reminders that mentioned the saver's own stated goal working best.
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.