What the snooze button actually does
Nine minutes at a time, most people give away between twenty and forty minutes every morning to the snooze button — and get almost nothing back for it. The reason is not moral. It is that fragmented sleep and continuous sleep are not the same substance.
The snooze button is one of the most-used features in consumer technology and one of the least examined. It was designed into mechanical clock radios in the 1950s and has survived, essentially unchanged in concept, into every phone on earth.
It is worth asking what it actually does, because the answer is not what most people assume.
Why nine minutes?
The nine-minute snooze is a genuine piece of engineering archaeology. When snooze was added to mechanical clocks, the gearing had to work with the existing mechanism, and the practical options were an interval slightly under ten minutes or slightly over. The shorter one was chosen, and by the time digital clocks removed the constraint entirely, nine minutes had become what people expected.
There is a widely repeated claim that nine minutes was selected to prevent people entering deep sleep. It is a tidy story and there is no good evidence for it. The constraint was mechanical.
What happens in those nine minutes
You have been woken. Your heart rate has risen, cortisol has been released, you are briefly conscious enough to press a button. Then you lie back down and your body attempts to return to sleep.
The problem is that returning to sleep is not instant. Sleep onset latency does not disappear because you were asleep two minutes ago. A meaningful part of those nine minutes is spent in the transition — drifting, not sleeping. What sleep you do get is light, N1 and early N2, the shallowest stages available.
And then the alarm goes again, and the whole process restarts.
The fragmentation problem
The critical point is that sleep is not a bucket that fills at a constant rate. Its restorative properties depend on continuity, because the useful processes — memory consolidation, the sequencing of stages, the physiological repair that happens in slow-wave sleep — need uninterrupted runs to occur.
Thirty minutes of continuous sleep and thirty minutes chopped into four pieces are not equivalent. This is the same mechanism that makes untreated sleep apnoea so damaging: sufferers may spend eight hours in bed but have their sleep fragmented hundreds of times, and they wake exhausted despite the hours on paper.
Snoozing is a mild, self-inflicted version of the same thing.
You restart sleep inertia every time
Sleep inertia decays on a curve, steeply at first. Wake once and you ride that curve down and it is over. Wake six times and you go through the steepest, worst part of it repeatedly, each time from a fresh start.
This is why a snoozed morning frequently feels worse than a morning where you simply got up — not despite the extra time in bed, but because of how that time was structured.
You are exchanging continuous, consolidated sleep for fragmented, low-quality sleep, and paying for it with repeated sleep inertia. It is a genuinely poor deal, which is why almost nobody reports feeling better after snoozing than before.
What the research actually says
Here is where honesty matters more than a clean story. The evidence on snoozing is not as one-sided as most articles pretend.
The fragmentation argument is solid and well-established. But research that has specifically examined a short period of snoozing — around thirty minutes — in habitual snoozers has found more mixed results than expected: modest improvements in cognitive performance immediately after rising, and no strong evidence of harm to overall sleep quality within that limited window.
Two things can be true at once. A brief snooze period may not be actively damaging for people who have already adapted to it. And it is still a worse arrangement than the alternative, because setting the alarm for the later time and sleeping through gives you the same minutes as consolidated sleep, with one wake instead of four.
What almost nobody defends is the extended version — an hour of snoozing across eight alarms — or snoozing on top of chronic sleep restriction. That combination reliably makes mornings worse.
The real cost, in numbers
Three snoozes at nine minutes is twenty-seven minutes, which sounds trivial. Across a five-day working week it is a little over two hours. Across a year it is roughly 117 hours — around five full days — spent in fragmented, low-value sleep that could have been continuous.
Framed the other way: if you snooze for thirty minutes every morning, you could set your alarm thirty minutes later and get thirty minutes of real, uninterrupted sleep instead. Same wake time, meaningfully better sleep. This is the single easiest improvement available to a habitual snoozer, and it requires giving up nothing.
How to stop, if you want to
Snoozing is a behaviour performed by someone who is not fully conscious, so willpower-based approaches fail for the same reason they fail everywhere else in this subject. Change the structure instead.
- Set the alarm for when you actually get upNot the aspirational time you would like to be a person who gets up. If you snooze for twenty-five minutes daily, your real wake time is twenty-five minutes later than your alarm. Move it and take the sleep.
- Use one alarm, not a sequenceA chain of alarms teaches you that the early ones carry no consequence, so you stop responding to any of them.
- Put it out of reachIf stopping it requires standing up and crossing a room, snoozing stops being a reflex and becomes a deliberate act — and by the time you are upright, most of the argument is over.
- Make stopping it require thinkingA maths problem cannot be solved by a sleeping brain. Our alarm clock has this switched on by default for exactly this reason.
- Put light on a timerA lamp or sunrise clock that starts brightening before the alarm begins the wake process while you are still asleep, so the alarm arrives at someone already halfway there.
- Fix the underlying shortfallCompulsive snoozing is very often just sleep debt. If you cannot get up, the honest first question is whether you slept enough — see how to wake up early.
Hatch Restore 2
A sunrise lamp, sound machine and clock in one. The light ramp starts before the sound, which is the part that actually changes how waking feels.
- Gradual sunrise light
- Large sound library
- App-controlled routines
- Phone-free bedside design
Typically $150–$180. Paid link — we may earn a commission at no extra cost to you.
Loftie Alarm Clock
Built specifically so the phone can leave the bedroom. Two-phase alarm, curated audio, no feed to fall into at 1am.
- Two-phase gentle wake
- Sleep sounds and breathwork
- Wi-Fi updates
- Blackout display mode
Typically $150–$200. Paid link — we may earn a commission at no extra cost to you.
When snoozing is fine
Not everything needs optimising. A single snooze, on a morning where you slept enough, on a day with nothing urgent in it, is not a problem worth solving. Some people genuinely use a short snooze as a deliberate transition period and function well.
The things worth changing are the patterns that carry a real cost: snoozing for forty-five minutes and being late; snoozing on top of five hours of sleep; snoozing so habitually that you have no idea what time you actually wake up any more.
The short version
Snoozing swaps consolidated sleep for fragmented sleep and makes you go through the worst part of sleep inertia several times instead of once. A brief snooze is probably not harmful. An extended one, or one layered on top of insufficient sleep, reliably makes your morning worse.
The best available move is also the easiest: set the alarm for the time you genuinely get up, put it somewhere you have to walk to, and sleep properly until then. Set one now.