TL;DR. A power nap should be about 10 to 20 minutes. That is long enough to lift alertness and short enough to keep you in light sleep, so you wake up clear instead of groggy. The nap length to avoid is roughly 30 to 60 minutes, because the alarm is likely to land in deep sleep and hand you sleep inertia. If you have the time and genuinely need recovery rather than a top up, take about 90 minutes instead, which is close to one full cycle. Nap in the early afternoon, and stop napping late in the day unless you want to pay for it that night.
Twenty minutes. If you only came for the number, that is it, and you can stop reading.
The interesting part is why the number stops working when you move it. Most nap advice online is a list of durations with no explanation of the mechanism, which is why so many people conclude that napping "does not work for them" after trying a 45 minute nap and waking up feeling like they have been hit with a shovel.
They were not doing it wrong. They picked the one length that reliably backfires.
I write this as someone who is bad at sleep for structural reasons. I build a habit tracker in the evenings around contract work and a small daughter, so my nights are shorter than I would design them and my afternoons occasionally need rescuing. The nap is a tool I actually use, and it is also the one I have gotten most wrong.
The short answer, and the mechanism behind it
Sleep is not a switch. You descend through stages: light sleep first, then slow-wave sleep, which is the deep stuff, then eventually REM. A normal cycle through all of that takes somewhere around 90 minutes, though it varies by person and by night.
A short nap keeps you in the shallow end. Ten to twenty minutes gets you into light sleep and no further. You come out of it easily because there was nothing deep to climb out of. Alertness improves, mood improves, and the cost is twenty minutes.
A long nap takes you all the way through. Around 90 minutes, you complete a cycle and surface near a natural transition point, which is why an hour and a half often feels better than an hour.
The gap in between is where the trouble lives. Wake somewhere in the middle of deep sleep and you get sleep inertia: heavy limbs, slow speech, a genuinely worse cognitive state than before you lay down. It clears, but it can take a while, and it is why the 45 minute nap has such a bad reputation.
The power nap calculator encodes this directly. It offers five nap types, and it labels each one for grogginess: the 20 minute power nap and the 90 minute full cycle are marked as no grogginess, the 30 minute nap as mild, and the 60 minute nap as moderate. If you tell it how many minutes you actually have, it picks the longest nap that fits and shows a warning when your available time lands in the awkward middle.
Pick the nap that matches the job
Different lengths do different work. Roughly:
| Length | What it is for | The catch |
|---|---|---|
| 10 to 20 min | Alertness, mood, getting through the afternoon | Does almost nothing for real sleep debt |
| 30 min | Popular for memory consolidation after study | You may surface groggy |
| 60 min | Sometimes used for creative or associative work | The worst grogginess of the set |
| ~90 min | Real recovery after a bad night | Costs 90 minutes and can push your bedtime |
Most days, the answer is the top row. The 90 minute nap is for the day after a genuinely broken night, a day off, or the afternoon before a night shift. If you find yourself needing it several times a week, the nap is not the problem you should be solving. That is a nights problem, and no amount of afternoon scheduling fixes it. The related read there is energy management rather than time management, which is about matching the work to the energy you actually have rather than manufacturing more of it.
When to nap matters as much as how long
Most people feel a dip in the early afternoon. It shows up whether or not you ate lunch, which is the detail that kills the "food coma" explanation.
The rule of thumb the calculator uses is 7 to 9 hours after you woke up. Wake at 7am and your window is roughly 2pm to 4pm. Wake at 6am and it shifts earlier. It is a heuristic, not a measurement, and it moves depending on where you sit on the morning-to-evening spectrum. If you have never worked out which end you are on, chronotypes explained covers the honest version of that, including how much of the four-animal framing is marketing.
The harder limit is the late-afternoon one. Sleep pressure builds across the day, and that pressure is what gets you to sleep at night. A nap spends some of it. Spend too much of it at 6pm and you will be lying awake at midnight, which is a bad trade in every direction. Napping late is the single most common way a good nap habit turns into a bad sleep habit.
The coffee nap, described honestly
The premise: drink a coffee quickly, lie down immediately, sleep for twenty minutes, and the caffeine arrives around the time your alarm does.
The mechanism is real enough. Caffeine takes roughly twenty minutes to have much effect, and it works partly by blocking the receptors that adenosine, the molecule that accumulates while you are awake and makes you feel sleepy, would otherwise occupy. Sleep clears some adenosine. So the caffeine lands with less competition.
The evidence is where I would ask you to lower your expectations. There are real studies behind coffee naps, and they are small ones. It is a plausible, mildly supported trick, not a hack. In my own experience it is worth doing when I already planned to have coffee and already planned to lie down, and it is not worth reorganising an afternoon around. The calculator includes it as one of the five nap types, which is fair, and its own page is more enthusiastic about it than I am.
The other cost is obvious: caffeine that late in the day is a real problem for a lot of people. If you sleep badly, the coffee nap trades an afternoon for a night.
About that NASA study
You will see NASA cited in almost every article on this topic, usually with two precise-sounding percentages attached.
The underlying research is genuine. In the 1990s, NASA researchers studied planned cockpit naps for long-haul flight crews and found that a short nap improved measures of alertness and vigilance compared with no nap. That is a real finding from a real study on real pilots, and it is a large part of why controlled rest on the flight deck became normal practice.
What has happened since is that the numbers have been copied from blog post to blog post, stripped of context, and, in more than one place, swapped around so that the alertness figure and the performance figure change places. Our own tool page repeats a version of them, which I am flagging because it is the kind of thing that gets fixed only when someone says it out loud.
So I will describe it the way I would want it described to me: NASA studied short naps in pilots, the naps helped, and the specific percentages you see quoted online should be treated as folklore until you have read the paper yourself.
Set it up so you do not have to think
The practical version is short.
- Decide the length before you lie down. Deciding while horizontal is how a 20 minute nap becomes 50.
- Set an alarm. Not for the nap, but for the worrying. Half the reason people fail to fall asleep in a nap window is that they are monitoring the clock.
- Add a settling buffer if you need one. The calculator's fall-asleep toggle adds 5 minutes to the timer. If you routinely need longer than that to drop off, add it yourself.
- Dark, cool, flat. Not required, but everything gets easier.
- Do not nap after the mid-afternoon. This is the rule I break most often and the one that costs me most.
The calculator runs the timer for you with a countdown ring and a choice of gentle alarm sounds, keeps a short history you can rate afterwards, and stores all of it in your browser. There is no account and nothing leaves your device. If you are more interested in the underlying 90 minute rhythm than in naps specifically, the ultradian rhythm timer is the daytime version of the same idea, and ultradian rhythms explained is the longer write-up.
When a nap is the wrong answer
Napping is a top up. It is not a treatment for anything, and there are cases where it actively works against you.
- If you have trouble falling asleep at night, standard insomnia guidance is to avoid daytime naps precisely because they spend the sleep pressure you need. Adding naps to a broken night is usually the wrong direction.
- If you are tired every single day regardless of how much you sleep, that is worth raising with a doctor, not solving with a timer. Persistent unexplained fatigue, falling asleep without meaning to, loud snoring, or waking up gasping are all things a website should not be triaging for you.
- If you are using naps to survive a workload that is not survivable, the nap is a painkiller. The early signs of burnout is the more useful read.
None of this is medical advice, and I am a developer rather than anything close to a clinician. What I can tell you is the part I have tested on myself for years: 20 minutes in the early afternoon works, 45 minutes does not, and the difference is not willpower.
Last updated: November 2026.
Written by Eusebiu, the solo founder building Loggd in public. I build the app in the evenings around contract work and a small daughter, which is why the afternoon dip is a topic I have researched with some personal urgency. I share what the aggregate data keeps showing on Threads.
Track the habit, not just the nap. Loggd shows your consistency as a forgiving contribution grid, so an afternoon you skipped stays one lighter square instead of resetting you to zero. Start free.
