StoweA reminder you set for yourself should be the easiest instruction in the world to follow — you wrote it, you meant it, it arrives signed by you. And yet it is usually the first notification you swipe away. The reason is not weakness. It is that most reminders are built, precisely and accidentally, to be ignored: they repeat until they mean nothing, they arrive when acting is impossible, and they name a topic instead of an action.
I build a planning app and had to decide how often it interrupts people. That decision forced me through the mechanics of why reminders die, and the mechanics turn out to be specific enough to design against — in any app's settings, or on paper.
Repetition teaches you to ignore the signal
Habituation is one of the oldest findings in psychology: a stimulus repeated without consequence stops producing a response. It is not a failure of attention; it is attention doing its job, filtering out what experience says is safe to skip. The same pattern shows up at life-and-death stakes in hospitals, where the safety literature calls it alarm fatigue — clinicians surrounded by frequent alerts, most of them non-actionable, begin missing real ones.
A daily reminder is a habituation experiment run on yourself. It fires at 9am; you are mid-something; you swipe. Nothing bad happens. Tomorrow the swipe is faster. Within two weeks the alert is processed and dismissed below the level of conscious reading — you will sometimes catch your thumb doing it. The reminder still fires every day. It just no longer arrives.
The bitter implication: adding more reminders to a system you are ignoring makes the ignoring stronger, because every additional ignorable alert is another training rep. Reminder systems do not usually fail by being too quiet. They fail by being too loud, then getting louder.
A reminder that arrives at the wrong moment can only be dismissed
A reminder is a bet that a particular moment is a good one to act. Most are terrible bets — set for round numbers ("9:00") rather than for situations where the action is possible. The alert fires; you are in a meeting, driving, mid-conversation; the only available response is dismissal. And each forced dismissal does double damage: the task goes unhandled and you get another rep of swipe-without-acting.
The fix is to anchor reminders to moments, not clock times, wherever the moment is predictable. "After I park at the office" beats "8:45". This is the territory of if-then planning — deciding when-and-where in advance — which is the best-evidenced technique in the planning literature (Gollwitzer and Sheeran's meta-analysis across 94 studies found a medium to large effect). A well-anchored reminder is an if-then plan with an external trigger: the situation named, the action pre-decided, nothing left to figure out at ping time.
Where no situational anchor exists, at least choose a clock time when the action is actually performable. A "call the pharmacy" reminder belongs inside pharmacy hours, at a time you are usually between things. Obvious, and almost nobody does it.
A vague reminder hands you the hard part at the worst time
"Gym." "Taxes." "Mum." Each of these names a topic and silently assigns you a job: translate the topic into a first physical action, right now, mid-interruption. That translation — from taxes to find last year's return in the email — is exactly the step that stalls in the first place; a reminder that skips it has reminded you of the mountain and left the climbing unplanned. Topic reminders get swiped not because you forgot caring but because acting on them requires thinking, and the ping arrived during something else.
Phrase the reminder as the first step, small enough to do or to start within a minute of reading it: "put shoes by the door", "open last-year-taxes folder", "call Mum — walk to kitchen". If writing the action version feels hard, that is the discovery that the task was never actually planned — better to find out at writing time than at ping time.
Why this bites harder with ADHD
External cues are precisely the right kind of ADHD support: when internal prompting is unreliable and time passes without being felt, a cue at the moment of action substitutes for the sense that was supposed to fire. The research on time perception in ADHD (Ptacek et al., 2019, for a review) is one reason at-the-moment cues matter more, not less.
But the failure modes above bite harder too. Novelty fades fast, so a nagging system becomes wallpaper sooner. An interrupted moment is harder to return from. And the standard advice — set more alarms — runs the habituation experiment at a larger dose. The conclusion is not fewer supports; it is a higher design bar for each one: rare, anchored, action-phrased.
What a reminder has to be to survive
- Rare. Every alert you can safely ignore weakens all the others. Prune the recurring ones that have become wallpaper — deleting a dead reminder is not giving up on the task, it is un-training the swipe.
- Anchored to a moment you can act — a situation where possible, a realistic clock time where not.
- Phrased as the first action, doable within a minute of reading.
- Backed by a plan, not substituting for one. A reminder is the last line of defence. If the day itself is small and chosen — 3–5 things you can actually finish — most of it needs no defending.
When I built Stowe I took the scarcity principle literally: the app sends at most three calm nudges a day, plus only what you explicitly set — no streak-at-risk alarms, no "we miss you". Not generosity; mechanics. An interruption budget is like any other budget — spend it everywhere and it buys nothing. That maths holds for whatever app is on your phone right now: open its notification settings, cull to the few alerts you would never want to miss, and let the plan carry the rest.
Frequently asked questions
Why do I ignore reminders I set myself?
Three mechanics stack up. Habituation: a signal repeated without consequence gets filtered out, and most reminders repeat daily with no consequence for swiping. Bad timing: a reminder that fires when you cannot act (in a meeting, driving) can only be dismissed, and each dismissal practises dismissing. Vagueness: an alert that names a topic ("gym") rather than an action leaves the hard translation work for the worst possible moment. None of these are character flaws; all three are properties of the reminder.
How do I make reminders I actually follow?
Make them rare, timed for a moment you can act, and phrased as the action itself. Rare, because every ignorable alert trains ignoring — a system that interrupts you three times a day stays meaningful in a way thirty times a day cannot. Timed to a genuine decision point, not a round number. Phrased as the first physical step ("put shoes by the door"), not the goal ("exercise").
Are reminders bad for ADHD?
No — external cues are exactly the right category of support when internal prompting is unreliable, and time-blindness makes at-the-moment cues more necessary, not less. But ADHD also makes habituation to a nagging system faster and more complete, so the design bar is higher: fewer alerts, better timed, action-phrased. A wall of notifications is not scaffolding; it is wallpaper.
How does Stowe handle reminders?
Frugally, on the theory that scarcity is what keeps an interruption meaningful: at most three calm nudges a day, plus only the reminders you explicitly set. No streak-at-risk alarms, no re-engagement pings. The plan itself does most of the reminding — a 3–5-thing day is small enough to hold in mind, so the app interrupts less because the day needs less rescue.