Can't sleep? Get out of bed. And don't look at the clock

You have been awake for a while. You do not know how long, and you are trying not to check, which is its own kind of work. The instinct is to stay exactly where you are, keep your eyes closed, and try harder.

The best-supported advice in behavioural sleep medicine says the opposite. Get up. Leave the room. Come back when you are sleepy.

It sounds like a punishment, and the first few times it feels like one. It is worth understanding why it exists, because the version of this rule that circulates online has a mistake in it, and the mistake is the part that keeps people awake.

The instruction, in short

If you are in bed and not sleeping, and it no longer feels like drifting, get out of bed. Go to another room if you have one. Keep the light low. Do something calm and undemanding. Return to bed when you feel sleepy, not when you feel like you have served your time. Repeat as many times as the night requires.

That is it. The rest of this article is about the details that decide whether it helps or backfires.

Why leaving the bed helps

The idea comes from Richard Bootzin, who in 1972 described insomnia in terms of conditioning. His argument was that for a good sleeper, bed is a strong cue for sleep: you lie down, and the association does most of the work.

For someone with insomnia, the bed has been paired with something else entirely. Hundreds of hours of lying there worrying, planning, checking the time, calculating how much sleep is left, and being frustrated. The bed stops being a cue for sleep and becomes a cue for wakefulness and effort. Bootzin’s first case study was a man who spent four hours a night in bed worrying about money and his job, then turned on the television to escape the worrying, and often fell asleep with it running.

The treatment is to break that pairing by refusing to do anything in bed except sleep. If you are awake, you are somewhere else. Over a few weeks the association starts to reverse.

This is not a trick for tonight. Tonight, getting up will probably feel like it made things worse. It works on the scale of weeks.

The part almost everyone gets wrong

Search for this and you will find it called the 15-minute rule or the 20-minute rule, phrased as a time limit: if you have not fallen asleep in twenty minutes, get up.

Here is the problem. Bootzin’s original instruction contained no number at all. It said simply: if you find yourself unable to fall asleep, get up and go into another room. A 2025 review of how clinicians actually implement the technique notes that this vagueness was deliberate. Bootzin left it unspecified in order to prevent clock-watching.

The numbers were added later by other people, and they have been inconsistent ever since. Some clinical descriptions say fifteen minutes, some twenty, and some published protocols reference the original paper without specifying anything.

So the most quoted instruction in the field has, in its popular form, imported exactly the behaviour its author was trying to avoid. And clock-watching is not a minor detail. Monitoring your own sleep, and trying to produce it, is one of the better-described mechanisms by which insomnia keeps itself going. A rule that has you checking the time every few minutes to see whether you are allowed to get up yet is a rule that manufactures the problem.

Use the feeling, not the clock. Current best-practice guidance suggests a more useful cue: get out of bed when you notice you are annoyed, frustrated, thinking about other things, or ruminating. That moment is unmistakable, and it arrives at a different time every night. It also sidesteps the fact that people with insomnia are famously poor at estimating how long they have been awake, usually overestimating badly.

If you want a number as a rough anchor, twenty minutes is a reasonable one. Just do not measure it.

What to do once you are up

The goal is to be awake somewhere that is not bed, in a way that does not wake you further.

Keep the light low. A lamp, not the ceiling. Not the kitchen at full brightness.

Choose something boring and finite. A dull book, folding laundry, a podcast you have heard before. Boring matters more than pleasant.

Avoid anything with a hook. Email, news, messages, anything with a feed, anything that generates a decision or an emotion. The point is to pass time, not to fill it.

Do not turn it into an achievement. If you get up and start being productive, you have taught yourself something worse: that waking at three is when you get things done.

Go back when you are sleepy, not when you are bored. Sleepy means heavy eyes, drifting attention, a pull toward lying down. Tired is not the same as sleepy, and going back because you are fed up is how the cycle repeats.

If you get back in bed and nothing happens, get up again. Bootzin’s original instructions said to repeat this as often as necessary, and in his case study the man was getting up four or five times a night at first.

When getting up isn’t practical

The instruction assumes a spare room, a warm flat, and a body that gets up easily. Plenty of people have none of those.

Reasonable adaptations, roughly in order of preference: move to a chair in the same room; sit up in bed with the lights off and the covers back; or stay where you are but stop trying to sleep.

That last one is a technique in its own right rather than a compromise. If you cannot leave the bed, the next best thing is to remove the effort, because the effort is most of what is keeping you up. Lying still and gently trying to stay awake is a recognised approach with a clinical history, and we built a guided version of it for exactly this situation.

If your mind is racing rather than churning with worry, a stream of unrelated images is the other option worth having: the cognitive shuffle is designed for that specific problem.

And if you want to skip the deciding entirely at three in the morning, we made a page for that: one question, one tap, no reading.

What this is part of

Stimulus control is not a standalone hack. It is one component of cognitive behavioural therapy for insomnia, and it sits alongside a few others that do the heavy lifting together. The most important companion instruction is a fixed wake time, every day, regardless of how the night went. Getting up at the same hour is what rebuilds the pressure to sleep the following night.

You will also see stimulus control mentioned alongside sleep restriction, which limits time in bed to roughly the time you actually sleep. That one is effective and also the component most likely to backfire when self-administered, since it involves a period of deliberate sleep deprivation. It belongs with a clinician, not with a blog post.

How well supported is it

Better than most things you will read about sleep, with the usual caveats.

Controlled work from the late 1970s onward found that stimulus control instructions reduced time to fall asleep relative to placebo and no-treatment conditions. It has remained a standard component of behavioural insomnia treatment for fifty years, and appears in current clinical guidance including brief behavioural treatment protocols delivered in four sessions.

The honest caveats: much of the foundational work is old, with small samples by modern standards. It is usually studied as part of a package, so isolating its individual contribution is difficult. And it asks a lot of people. Adherence is the main practical limitation, which is unsurprising given that it requires getting out of a warm bed repeatedly on nights you feel worst.

There are also people for whom getting up repeatedly in the dark is not advisable, including anyone at risk of falls. If that is you, ask before adopting it.

When technique is no longer the question

If you have been sleeping badly three or more nights a week for more than three months, you are looking at chronic insomnia, and it has treatments that work and keep working. The rule above is one piece of one of them, and piecing together a treatment from articles is a slow way to get there.

Waking unrefreshed after a full night is a different signal again, and worth raising with a doctor rather than solving with a bedtime instruction.

We have written about what to do at that point and how to ask for it.

Common questions

Is it 15 minutes or 20 minutes? Neither, originally. Bootzin’s instruction had no number, deliberately, to stop people watching the clock. Later versions added fifteen or twenty. Use the feeling of being annoyed or alert instead, and treat twenty minutes as a rough sense of scale rather than a timer.

Does this apply to waking in the middle of the night too? Yes. The instruction is the same whether you are trying to fall asleep at the start of the night or lying awake at four in the morning.

What if I get up and then can’t get back to sleep at all? Then you get up again. That is the procedure, not a sign it is failing. The first week or two often produce less sleep, not more, which is why it is worth knowing in advance.

Can I read in bed if the reading is boring? Not while you are doing this. The whole point is that the bed stops being a place where you do anything except sleep. Read in the chair.

How long before it helps? Think in weeks, not nights. It is reconditioning an association built over months or years.