You know the feeling. A reminder arrives at 2:30 PM about the report you meant to write at 10 AM. You swipe it away. It comes back at 4 PM. You swipe it away again. By 6 PM you've decided the reminder itself is the problem and you turn notifications off. The report still isn't written.

The missing variable in almost every reminder app is the one that matters most. Not what you need to do. Not why you need to do it. When.

Why timing beats motivation

There's a common story about productivity: you fail to do things because you lack discipline, and the fix is somewhere between a system, a journal, and more willpower. It's a clean story. It's mostly wrong.

The more honest story is that adult brains have a few hours a day where meaningful work is possible, a few more where it's hard but doable, and a long stretch where it simply is not going to happen no matter what a reminder says. You already know this from experience. You sit down at 3 PM determined to finish the thing, and you write two sentences in forty minutes. You come back at 9:30 AM the next morning and you finish it before your second coffee.

That difference is not willpower. It is the shape of your attention. The research community has spent sixty years mapping it, and most of it never made it into productivity software.

The attention window

Psychologist Peretz Lavie described what he called ultradian rhythms, roughly 90 to 120 minute cycles of alertness that repeat throughout the day. You don't feel this at a moment-to-moment level, but over a week the pattern is legible: most people have two to three windows per day where focused cognitive work feels easy, and the rest of the day feels like wading through sand.

The windows don't land at the same time for everyone. Chronotype research generally splits the population into morning types, evening types, and a larger middle group, with peaks that land anywhere from early morning to early evening depending on the person. What's more predictive than any clock time, though, is your own past behavior. Your phone can watch for when you actually act on things, and your calendar history knows when the meetings drop and the real work begins.

Gloria Mark's research on interruption at UC Irvine adds the other half of the picture. Her 2008 paper on the cost of interrupted work found that it takes on average 23 minutes to fully refocus on a task after a single interruption. When a reminder lands in the middle of deep work, the cost is not zero-plus-small. It's the full 23 minutes, often repeated if the reminder stacks more on the way.

So now you have two forces working against the fixed-time reminder. First, it probably doesn't land in a window when action is possible. Second, if it does interrupt real work, it pays a large hidden tax whether you act on it or not.

How Nudge measures its own timing

One of the things we had to figure out at Nudge early on was how to measure the health of its notifications. Not "how many did we send" but "how well did they land."

The measure is simple. For any given nudge, we track three things: did it arrive in a window where the user had shown they tend to act, did the user act on it, and did any subsequent nudge get ignored or dismissed. A good nudge scores high on the first two and doesn't have to try again. A bad nudge either lands when the user was unavailable, gets dismissed, or triggers a follow-up that also fails.

The point of measuring is that it lets us degrade rather than escalate. Most notification systems get louder when you ignore them. More pushes. Badges. Sound. Nudge goes quieter. It notices that the current context was wrong and tries a different window next time. Over a week, the rhythm stabilizes around the moments when the user is actually acting.

This is the practical answer to Mark's 23-minute interruption cost. You don't just avoid bad timing. You learn what "bad timing" looks like for this particular user, and you stop doing it.

What this looks like in practice

Take a simple case. You set a recurring reminder to check in with your team every Monday morning. A fixed-time reminder fires at 9:00. In any given week your 9:00 might be a standup, a commute, a deep-work block, or a quiet coffee. Four very different contexts for the same prompt.

A context-aware nudge asks: on past Mondays, when did the user actually respond to personal prompts? Say the answer is 9:45 most weeks, right after the standup ends and before the next meeting begins. The nudge lands at 9:40. You check in. Nothing feels forced.

The same system is what keeps Nudge from sending you the grocery list while you're on the Monday call. It's not that it refuses to interrupt you. It's that the system knows interruption has a cost, and it spends the cost deliberately, on the few moments that actually matter.

Where the science gets thin

Two honest limits to all of this.

First, ultradian rhythm research is real but not precise. The 90-to-120-minute cycles exist, but they are modulated by sleep, caffeine, stress, and week-to-week variation. No timing algorithm is going to land every nudge perfectly. The goal is to be directionally better than a fixed clock time, which is a low bar that still isn't the default anywhere else.

Second, all of the behavioral signal in the world doesn't help if the user hasn't built up a history yet. For the first week or two of using Nudge, the model leans on your calendar availability and sensible defaults. The timing gets noticeably better around week two, once there is real action data to learn from.

What you can do today, with any app

You don't need any particular app to use this. Track yourself for one week: every time you actually complete something a reminder asked for, note the time. Most people find two or three windows where almost all of it happens. Move your recurring reminders into those windows and delete the rest. And stop scheduling reminders inside known meeting blocks; a reminder that fires during your Monday standup is dead before it arrives. A fixed schedule won't adapt week to week the way a system that watches your behavior does, but it already puts you well ahead of the 9:00 AM default.

Timing is the fix, not more discipline

When a reminder feels like an ambush, it usually is one. The app picked a time that seemed reasonable to it and ignored the question of whether you were available, capable, or in the right mood. Most of the "I'm just bad at productivity" feeling people carry around is not a character flaw. It's a timing mismatch between a tool and a brain. If you have ADHD, this mismatch is the whole story: reminders that ignore your attention state fail harder. Your task manager isn't the problem. The moment it interrupts you is.

Getting the timing right doesn't need a breakthrough. It needs a reminder system that pays attention to when you actually act, and deploys its interruption budget accordingly.

Nudge watches your rhythm and lands reminders in the moments you're most likely to act on them. Free on iPhone and web.