Sleep trackers suck in many ways. Here’s what works.

Not many people remember the Zeo sleep tracker these days — but I still recall how tired I was at its launch event in 2009. The event was in New York, where my then-employer flew me once a month. I was based on the west coast, and asked Zeo’s sleep expert, who coached NBA players on optimal sleep, how to deal with my frequent jet lag. “You should always sleep on east coast time,” was his reply. Hard for a night owl like me to hear, but it was the most useful data I gained that day: Science tells us that sticking to the same sleep schedule every night is one of the best things we can do for our bodies. 

The sleep tracker itself? Not so useful. It had two components: a headband I couldn’t forget I was wearing, and a bedside clock/charger that was so damn bright it broke another cardinal rule of good sleep (make sure you’re bedding down in as dark an environment as possible). Zeo’s tracker was a sophisticated piece of lab-grade tech that broke your sleep down into four stages (awake, REM, light, and deep), and gave you a single consumer-friendly “ZQ” score every morning. Yet it couldn’t persuade enough of us perpetually sleep-deprived humans to strap one on — Zeo went out of business in 2013.

If the current explosion of sleep tracking gadgets is anything to go by, Zeo may simply have arrived a decade too early to cash in: The era of apps is where the money’s at. You can now buy app-connected Bluetooth-based sleep headbands (the Muse S), sleep-tracking smart rings (led by the Oura 3, which I reviewed for a month here), wristbands (like the Whoop), smart watches (the Apple Watch series 6 tracks sleep directly, and its predecessors can use a variety of third-party apps), smart pads that go on your mattress and can detect snoring (the Withings Sleep), apps that promise to turn your phone into a smart pad, entire smart mattresses (the Pod Pro by Eight Sleep), and a not-at-all creepy Google Nest Hub that judges how well you’re sleeping by pointing a camera at your torso and a mic at your mouth. 

Many of these services also require subscriptions — suggesting that if you want the best quality sleep, you’ll have to keep ponying up for the rest of your life. And when it comes to services to sell, we’re just getting started. Amazon won permission from the FCC last year for a curious and/or terrifying device that tracks sleep via radar.

Do sleep trackers actually work?

We’re not great at sleeping in the modern world, admittedly. It’s one of the top four things you can do for your health, and we’re doing it less and less, especially during the pandemic. But is more technology really the best solution for the average sleeper? Is it making the problem worse? And can something as simple and low-tech as a timer work just as well, if not better? Each of the high-tech devices mentioned above uses different algorithms to guess at which sleep stage your body is in, and the results can vary wildly. “It’s known to be a fuzzy standard,” Oura Chief Product Officer Chris Becherer admits. 

Test any sleep tracker for a good chunk of time and you may find a hilariously false reading: a Fitbit that thinks you’re asleep when you’re watching a movie, a smart mattress pad that can’t tell the difference between your movement and your partner’s. Even the supposed gold standard of headband-based sleep tracking — like the Zeo or its descendant, the Muse S, which look directly at brain activity — are inaccurate around 10 percent of the time.

Sleep no more: The defunct Zeo headband.
Credit: MediaNews Group / Boston Herald via Getty Images

And that matters, because an inaccurate sleep reading can ruin your waking life. One study in 2013 deliberately fed its participants the wrong sleep data, and noted that they did exactly as well on cognitive tests as their fake data predicted. A 2020 study, which told participants they’d had 5 hours sleep when they’d actually had 8, and vice versa, came to the same conclusion. In sleep, as in so much, the placebo effect is real.  

Heck, even an accurate reading can increase anxiety. A 2017 paper by sleep tracker researchers coined a new condition among sleep tracker users: Orthosomnia. “The perfectionist quest to achieve perfect sleep is similar to the unhealthy preoccupation with healthy eating, orthorexia,” its authors wrote. They cited sleep study participants who indulged in unhealthy sleep behavior, such as using their phone right up until they fell asleep, and somehow expecting that sleep trackers would fix the problem. 

Sleep trackers can help create habits around good sleep hygiene, in theory, by telling you when it’s time to stop using screens or any other habit that might keep your brain awake. The Oura 3, for example, starts out by studying your sleep patterns for a week or so, suggests your ideal bedtime, then sends a “time to wind down” notification to your phone two hours earlier. But Becherer says many Oura users joke about or roll their eyes at these notifications, which often arrive in the 8pm hour — just when the evening feels like it’s in full swing, in other words. 

I’m not the only one to turn those nagging notifications off, removing one major potential benefit of wearing the thing altogether. (Contrarian resistance to nagging also explains why I’ve never set up the similar Sleep Focus setting on my iPhone.)  

I still use the Oura 3 a month after reviewing it, more out of a sense of inertia than anything else: It’s on my finger, might as well let it track my sleep. I try to ignore the eerie green glow sometimes visible from its inward-facing sensors. Most nights I also wear my Apple Watch to bed, again through sheer force of habit — I’m used to Sleep++, the sleep tracker app I used before the Oura. And then most mornings, I shake my head at the incongruity between the two readouts.

Last night, for example, I spent 8 hours and 8 minutes in bed according to Oura, and 7 hours 33 minutes according to Sleep++. The latter app says I had 7 hours and 11 minutes of restful sleep. Oura clocked it at a mere 6 hours 38 minutes, but also awarded me 97 minutes of restorative deep sleep and a “good” sleep score of 77, which is above average. 

The $399 SmartSleep system from Philips, which claims to make deep sleep more “efficient” via headphone tones.
Credit: Ethan Miller / Getty Images

What am I supposed to do with any of this, really? Is more information better, as we tend to assume in our data-driven world, or am I simply giving myself a case of orthosomnia? Would it be less anxiety-inducing if I simply trusted one source of data above all others, or would that lead to cherry-picking results that make me look better (or worse)? Or should I concentrate entirely on the one metric that is entirely under my control: Time spent in bed?

Prior to using Sleep++, that’s exactly what I did, using the $4.99 ATracker app on my Apple Watch and iPhone (part of a productivity system I wrote about here). I made it a habit to charge my watch overnight, start the sleep timer on ATracker when I put it on the charger, and stop it when I picked up either device in the morning. The very fact that I was recording my sleep, even in this low-tech way, seemed to keep me honest; I tended to avoid going to bed too late or reaching for my phone too early.

I’d have to work out some kind of subjective sleep score of my own to compare the cheap timer app with sleep trackers that are at least a hundred times more expensive — but my instinct is that the average ATracker night was never any worse than the average Oura night. They might even have been better; at least the dumb timer app never gave me an anxiety-inducing sleep score based on incomplete information. 

How much sleep do you really need?

“The right amount of sleep is the number of hours needed for you to wake up feeling refreshed,” says Dr. Guy Leschiziner, clinical lead at the Sleep Disorders Centre in Guy’s Hospital, London. (I spoke to Leschiziner for my story on liminal sleep and dream hacking.) “That [amount] varies for everyone.” Generally speaking, teenagers need more and people over 65 need less. But there is no one right bedtime, right number of hours, or right number of sleep segments. 

You may have heard about the supposedly traditional “second sleep,” for example. History offers many accounts of people waking up for an hour or two in the middle of the night in preindustrial days; they’d read, hang out, pray, or have sex, then settle down for another slumber. But some preindustrial societies made it through the night just fine without what scientists now call biphasic sleep. A 2021 meta-study found that modern-day biphasic sleepers do not feel more refreshed, nor did trackers show them getting any more deep sleep during the night. 

The right amount of sleep is dependent on genetic factors, Leschiziner says, something sleep trackers don’t account for just yet. Perhaps that’ll change if (or when) a DNA-testing service like 23AndMe gets into the sleep tracking game; we’ve already identified two genes associated with naturally short sleepers. 

Until then, Leschizinier isn’t against sleep trackers, exactly, and treads cautiously whenever his patients ask about them. But he also notes they’re more useful for researchers like himself doing large sleep studies. Stick trackers on hundreds of people and noise in the data will tend to even out in the aggregate, allowing scientists to see interesting trends. Stick a tracker on one of Leschizinier’s insomniac patients, and orthosomnia is often the main outcome. He’s also seen people misdiagnose themselves with insomnia simply because their sleep trackers tell them they don’t sleep enough. 

What to try before you buy a sleep tracker

If you have no clue what level of sleep your DNA requires you get, you can in theory find out just as easily with three of the cheapest, most low-tech tools imaginable: a watch, a notebook, and a writing implement. (You might consider using apps for those functions; you might also consider that using actual physical objects will help train your brain away from using screens at bedtime.) Each night, note down what time you went to bed. Each morning, note down what time you got up. And at some point during the day, make a note of how tired or awake you feel. After a few weeks, start looking through what you’ve written. You don’t need to be a data scientist to start noticing correlations. 

If you’re feeling particularly fancy, you might also want to use this low-stress system to try out different evening routines and see the results. For as much as individual sleep trackers disagree on the numbers, individual sleep experts agree on what leads to good sleep hygiene for all of us. You’ve probably heard it all before. Caffeine isn’t the greatest substance for sleep; if you can’t quit it, try not to have it in the evening (or afternoon if you’re particularly sensitive). Reduce or eliminate your alcohol intake, particularly in the two hours before bed.

That’s also the period in which you should stop using screens (yes, sorry, the TV counts), eating, exercising, or doing anything else that causes you to become energized or stressed. Make your bedroom a sanctuary for sleep, sex and little else. (Reading books is fine since they quantifiably reduce stress — just don’t read on a screen.) Reduce light pollution to a minimum (I recommend blackout blinds). And yes, above all else, for your body’s sake, try to keep roughly the same bedtime every night, no matter what coast you’re on. 

All of which is easier said than done, of course. I must have broken most of those rules at least twice in the past week alone. But as with most attempts to fix our habits, it’s the small moves in the right direction that make a difference. Turned off the TV early tonight? That’s a win. Had your last coffee at 4pm rather than 5pm? That’s a win too. You don’t need a stress-inducing sleep tracker strapped to your forehead, or anywhere, to tell you that.