How Long Should It Take to Fall Asleep? The Range, and Why Too Fast Is a Warning

Short answer: for a healthy adult, 10 to 20 minutes from lights out to sleep. The National Sleep Foundation's expert panel rates a sleep latency of 15 minutes or less as appropriate across adult ages, 16 to 30 as acceptable but uncertain, and over 45 as a sign of poor sleep quality. Clinically, taking more than 30 minutes on at least three nights a week for three months is one of the defining criteria of insomnia. And falling asleep in under five minutes as a matter of routine is not evidence of being a good sleeper; on the standard daytime test of sleepiness it is the threshold for pathological sleep deprivation.

That last point is the one worth reading the article for. Most people who ask this question are worried about taking too long. A substantial fraction should be worried about the opposite.

The ranges, with their sources

The literature calls this number sleep onset latency, and it has been measured in laboratories with EEG for seventy years, so the ranges are not guesswork.

A few clarifications the table cannot carry.

The clock starts at lights out, not at getting into bed. Reading in bed for forty minutes and then falling asleep in ten is a ten-minute latency. Lying in the dark trying to sleep for forty minutes is a forty-minute latency. People routinely count the first and worry about it.

One bad night is not a latency. The insomnia definition requires frequency and duration precisely because everyone has nights of staring at the ceiling. The pattern is the diagnosis; the night is noise.

Feeling like it took an hour is not the same as it taking an hour. Laboratory studies consistently find that people overestimate how long they took to fall asleep, and insomnia patients overestimate most. The last thing you remember is being awake, so the awake period feels like all of it.

Why under five minutes is a warning

Sleep medicine has a formal test of sleepiness called the Multiple Sleep Latency Test. Through the day, at two-hour intervals, you lie down in a dark room and are told to try to sleep, while EEG records how long it takes. A rested person takes ten minutes or more, and often does not fall asleep at all in the twenty minutes allowed. The AASM practice parameters treat a mean latency of eight minutes or less as consistent with clinically significant sleepiness, and under five as the range in which narcolepsy and severe sleep deprivation live.

The test exists because the speed at which you fall asleep is a fairly direct readout of sleep pressure: the accumulated drive to sleep, built by time awake and paid down by sleeping. Low pressure, slow onset. High pressure, fast onset. Someone who is "out the moment their head hits the pillow" every night is carrying high sleep pressure to bed every night, and the usual reason is that they are not sleeping enough to discharge it.

This is the point most sleep-hygiene content gets backwards. It treats instant sleep as the goal and a fifteen-minute wind-down as a failure. The lab data says the reverse: a rested brain takes a little while to fall asleep, and a brain that does not is telling you something.

Falling asleep in under five minutes on the daytime latency test is the range in which severe sleep deprivation and narcolepsy are found. It is not the range in which good sleepers are found.

Summarising the AASM practice parameters for the MSLT, Littner et al, Sleep, 2005

If you fall asleep in under five minutes nightly and feel fine in the day, the likeliest explanation is still that you are slightly under-slept and have adapted to it, which is what the caffeine literature repeatedly finds people doing. The test is to add thirty minutes to your sleep window for two weeks and see whether the onset time lengthens. If it does, you were short.

Why over thirty minutes happens

The list of causes is long, but four of them account for most cases in otherwise healthy adults.

Caffeine, later than you think. A 200 mg coffee at 4pm leaves about 76 mg circulating at 11pm, and the study that gave people caffeine six hours before bed found it added time to sleep onset and cut total sleep by more than an hour, without participants noticing. The single most effective intervention for a long latency in a coffee drinker is a cutoff around eight hours before bed.

Insufficient sleep pressure. You cannot fall asleep quickly if you are not tired enough, and two things sabotage that: a nap in the afternoon, and going to bed before you need to. Someone who needs seven hours and gives themselves nine will lie awake for an hour and conclude they have insomnia. Sleep restriction, one of the core components of the treatment below, works by fixing exactly this.

Conditioned arousal. Spend enough nights awake and frustrated in bed and the bed becomes a cue for being awake and frustrated. This is the mechanism that turns a few bad weeks into chronic insomnia, and it is the reason the standard advice is to get up if you are not asleep within about twenty minutes, and go back only when sleepy.

Temperature. Sleep onset requires core temperature to fall, which the body does by dumping heat through the hands and feet. Kräuchi and colleagues showed in Nature that the rate of heat loss from the extremities was the best predictor of how fast people fell asleep, better than core temperature itself. Cold feet in a cold room slow it down. Warm feet, whether by socks, a warm bath an hour before bed, or a warm room that you then cool with a fan, speed it up.

Alcohol deserves a note because it shortens onset and people count that as a win. It also fragments the second half of the night badly enough that the net effect on sleep is negative.

How it changes with age

The intuition is that older people take much longer to fall asleep. The data says the change in onset is small; the change is in staying asleep.

Ohayon and colleagues' 2004 meta-analysis pooled laboratory sleep recordings from healthy people across the lifespan. Sleep latency rose with age, but modestly, on the order of a few minutes across adulthood. What rose sharply was wake after sleep onset: the time spent awake in the middle of the night, which roughly tripled from young adulthood to old age. Slow-wave sleep fell steeply too. So a 65-year-old taking twenty minutes to fall asleep is normal, and a 65-year-old awake for an hour at 3am is also, unfortunately, close to normal.

The NSF panel's latency thresholds are the same for every adult age group for that reason. There is no allowance for age in the "appropriate" band because the panel did not think the evidence justified one.

Your sleep tracker's number is the least reliable one it shows

The onset time on a wrist device is an estimate from movement and heart rate, not a measurement of brain state, and it fails in a specific direction. Chinoy and colleagues (2021) tested seven consumer devices against laboratory polysomnography and found the consistent weakness was detecting wake: the devices were good at recognising sleep and poor at recognising the lying-still-but-awake that makes up most of sleep onset. Lie motionless in the dark for twenty-five minutes and most devices will record much of that as sleep.

Two consequences. The tracker's onset number tends to be shorter than the truth, so a device saying "8 minutes" is not evidence that you fall asleep too fast. And it is not stable enough night to night to diagnose anything. Use it, if at all, for the direction of a trend over weeks when you change one thing, which is the same rule that applies to every other number it shows.

A notebook and the clock are more accurate. Write down the time you turned the light off and your best guess of when you fell asleep, every morning for two weeks. The guess will be biased long, and it will still be more informative than the wrist.

What actually shortens it

The treatment with the strongest evidence for insomnia is not a pill and not a gadget. The AASM's 2021 guideline gives its one strong recommendation to cognitive behavioural therapy for insomnia, a package of which the two most active ingredients are directly about sleep onset:

Stimulus control. Bed is for sleep only. Go to bed only when sleepy. If not asleep in about twenty minutes, get up, do something dull in dim light, return when sleepy. Same wake time every day regardless of the night. This rebuilds the bed as a cue for sleep.

Sleep restriction. Limit time in bed to roughly the amount you are actually sleeping, then expand it in fifteen-minute steps as onset shortens. This concentrates sleep pressure so that when you do lie down, you go under. It is unpleasant for the first week and it is the component that most reliably fixes a long latency.

The rest of the standard list, in order of how much it usually matters:

What to actually do with this

If you take 10 to 20 minutes: nothing. That is what a rested adult looks like. Stop measuring it.

If you take under 5 minutes most nights: add thirty minutes to your sleep window for two weeks. If the onset stretches, you were under-slept. If you also fall asleep in the day unintentionally, or someone has told you that you snore and stop breathing, see a doctor about apnoea.

If you take over 30 minutes, three nights a week, for months: that is insomnia and it has a treatment with strong evidence. Start with the caffeine cutoff and a fixed wake time, then stimulus control, then sleep restriction. If self-directed CBT-I does not move it in six weeks, a clinician who delivers it can.

If it varies wildly: it is supposed to. Look at the fortnight, not the night.

Questions people ask

How long should it take to fall asleep? Ten to twenty minutes for a rested adult. The National Sleep Foundation's panel rates 15 minutes or less as appropriate at every adult age and 16 to 30 as acceptable. Over 45 minutes is rated poor.

Is falling asleep in 5 minutes bad? As a nightly pattern, it usually indicates sleep debt. On the clinical daytime sleepiness test, a mean latency under 8 minutes is the threshold for excessive sleepiness and under 5 is where severe deprivation and narcolepsy sit. Occasionally, after a hard day, it is normal.

How long does it take to fall asleep with insomnia? The diagnostic threshold is more than 30 minutes, on at least three nights a week, for at least three months, with daytime consequences. Many people with insomnia report an hour or more, and laboratory recordings usually show it was somewhat less than they felt.

Why does it take me so long to fall asleep even when I'm tired? The usual causes are afternoon caffeine, going to bed before you are sleepy enough, a bed that has become associated with lying awake, and cold extremities. Being tired and being sleepy are different states; sleep onset needs the second.

Is it normal to take an hour to fall asleep? Not as a routine. An hour on most nights meets the insomnia threshold and responds well to behavioural treatment. An hour on one night after a stressful day is ordinary.

Does sleep onset get longer with age? Slightly. Across adulthood the increase is a few minutes. What increases substantially with age is waking during the night and the amount of deep sleep, not the time to fall asleep.

Is my sleep tracker's "time to fall asleep" accurate? Poorly. Consumer devices are weak at detecting quiet wakefulness, so they tend to record part of the falling-asleep period as sleep and understate the latency. Use the trend, not the number.

What is the fastest healthy way to fall asleep? Arrive at bed sleepy, not just tired: no caffeine after early afternoon, no late naps, a fixed wake time, and warm feet in a cool room. Getting up after twenty minutes awake, rather than lying there, shortens onset over the following nights.

How long does it take to fall asleep after melatonin? Melatonin shortens sleep onset by a few minutes on average in trials, and its main effect is on the timing of the body clock rather than on sleepiness itself. It is more useful for shifting when you sleep than for falling asleep faster at your usual time.

This article covers sleep science for general information and is not medical advice. Loud snoring with pauses in breathing, falling asleep unintentionally during the day, or insomnia lasting more than a few months warrant a conversation with a doctor; both sleep apnoea and chronic insomnia are treatable.