How to Stop Binge-Watching

Nearly everything written about stopping binge-watching is advice about willpower. Set a limit. Notice how you feel. Ask yourself whether you really want another episode.

There is one intervention here with an actual randomised experiment behind it, and it takes about ninety seconds. Everything else in this article is either weaker than it is usually presented, or a correction to something you have probably been told.

The one experiment worth knowing about

Researchers took 76 regular Netflix users and randomly assigned half of them to turn autoplay off for a period of ten to seventeen days. The other half left it on.

What makes the study unusual is how it measured the result. Rather than asking people what they watched, participants requested their own Netflix viewing records under data-protection law, which gave the researchers real timestamps for real sessions. A browser extension independently checked that autoplay was actually off. Each person's previous six months of viewing served as their own baseline.

Turning autoplay off reduced average daily viewing by 21 minutes and shortened the average session by about 18 minutes. The gap between one title ending and the next starting grew by about 24 seconds.

The most telling number is the one that did not move. Time between sessions was unchanged. Autoplay was not making people open Netflix more often. It was making them stay once they were already there.

That is exactly what you would expect from a default rather than from a craving, and it is why changing the default beats deciding to be more disciplined.

The caveats are real: 76 people, all in the US, for a fortnight, and the researchers could not tell active watching from a television left running in an empty room. Nobody should treat 21 minutes as the amount they personally will save. But as evidence that the design does the work, it is the best thing in this literature.

Turning it off, exactly

Netflix has two separate settings, and most articles conflate them. Both are per profile, not per account, so switching one off for you does nothing for anyone else in the house. Neither can generally be changed from the television app itself.

• In a browser: go to your Account page, select Profiles, choose a profile, then Playback settings. Toggle Autoplay Next Episode. The same screen carries Autoplay Previews, which is the one that starts trailers playing while you are browsing.
• In the mobile app: open Netflix, tap My Netflix at the lower right, tap the profile name at the top, tap Manage Profiles, select the profile, then toggle Autoplay Next Episode. On an iPhone or iPad, tap Done.

Netflix notes that the change propagates automatically but that a device may need refreshing to pick it up. Switching profiles and back, or signing out and in, does that.

YouTube works the opposite way round. The setting is per device rather than per account, which catches people out: turn it off on your phone and it is still on in the living room.

• On a computer: the Autoplay switch sits at the bottom of the player on any video.
• On a phone or tablet: the switch is at the top of the player, or under your profile picture, then Settings, then Autoplay.
• On a smart TV: open the YouTube app, go to Settings, scroll to Autoplay.

Google also notes that autoplay is off by default for viewers aged 13 to 17.

Prime Video and Disney+ both have an equivalent setting, and we are not going to print steps we could not verify. Neither company's help documentation could be read at the time of writing: Amazon's pages were unreachable, and Disney's help centre search was unavailable, while its player-controls page documents a next-episode button and no autoplay toggle at all. Look under playback or profile settings in each service, and expect the same pattern as Netflix, where the change is made on the web or in the mobile app rather than on the television.

If the phone is the bigger problem than the television, limiting the apps designed to be endless is a separate job with its own settings.

What the sleep research actually found

This is where most coverage goes wrong, and the real finding is more interesting than the exaggeration.

A 2017 study in the Journal of Clinical Sleep Medicine surveyed 423 adults aged 18 to 25 and found that self-identified binge viewers had roughly twice the odds of poor sleep quality, along with higher fatigue and more insomnia symptoms.

The part worth stopping on: regular television viewing was not associated with any of the three sleep outcomes. Only binge viewing was. Whatever is happening, it is not simply that watching television keeps you up.

The authors' explanation is cognitive pre-sleep arousal. Not the screen, not the light, but a narrative that leaves your mind still running when the lights go off. In their analysis that mental arousal accounted for the entire relationship.

Now the hedging that has to come with it. This is a cross-sectional survey, so it cannot say which way the arrow points, and poor sleepers having more waking hours available to watch television is at least as plausible as the reverse. It was a self-selected Facebook sample of mostly Belgian students, all self-reported, with no sleep measured by any instrument.

The pooled picture is weaker still. A 2022 meta-analysis of 16 studies covering 8,077 people found the association between binge-watching and sleep problems at an effect size of 0.16, with a confidence interval whose lower bound almost touches zero, extreme disagreement between the studies pooled, and nearly two-thirds of them rated poor quality. Every single one was cross-sectional.

So the honest sentence is that binge-watching is weakly and inconsistently associated with worse self-reported sleep in studies that cannot establish direction. It is not that binge-watching wrecks your sleep.

The pattern matters more than the episode count

There is no agreed definition of binge-watching. Some studies use two episodes, some three, many just ask people whether they consider themselves binge viewers. Most sessions are smaller than the word implies.

In the Belgian survey cited above, just over half of sessions ran to three or four episodes, 25.6 per cent were two, 16.2 per cent were five or six and only 5.9 per cent went beyond six. Three episodes of a drama is the typical case, not an outlier.

A study of 482 Italian young adults found something more useful than any threshold. Participants classified as problematic binge-watchers had significantly worse sleep than everyone else, taking longer to fall asleep, sleeping less and functioning worse in the day. But non-problematic binge-watchers were statistically indistinguishable from people who did not binge-watch at all.

The behaviour was not the problem. The pattern was.

You are probably not addicted, and there is a study about that too

If you have taken an online quiz that told you your viewing is pathological, here is the paper that should reassure you.

Researchers ran 7,983 people through the standard binge-watching questionnaire and tested whether any cut-off score could reliably separate problematic from non-problematic viewers. Five of its seven components performed poorly. Even the two best ones misclassified roughly three people in ten and missed nearly half of the cases they were supposed to catch.

Their conclusion was that no reliable threshold could be established, and they warned that cut-off-based work on comparable behaviours has produced implausible rates of ten to twenty per cent pathological cases, which promotes over-pathologising and stigma. Binge-watching is not a recognised disorder in any diagnostic classification.

It is worth knowing that the evidence does not all point one way, either. A German study of 668 people found binge-watching associated with greater feelings of autonomy, competence and relatedness, and through those with better well-being, including recovery and enjoyment. It was a partial replication that could not reproduce the original model cleanly and was run during pandemic lockdowns, so it is not strong evidence. It is a reminder that watching four episodes of something you love on a Saturday is not a symptom.

What the literature does distinguish is planned from unplanned viewing. Loss of control shows up in the unplanned case. That is the distinction worth applying to yourself, rather than an episode count.

The Netflix quote everyone repeats is not quite what he said

Reed Hastings is endlessly quoted as saying Netflix's biggest competitor is sleep. It makes a good headline and it is not a real quote.

What he said, in a 2017 earnings interview and as reported by Entrepreneur and NPR Illinois, was that when you get addicted to a show you stay up late, and so "we're competing with sleep, on the margin", and that this is a very large pool of time. The qualifier at the end is always stripped out, and the point he was making was about the size of the market rather than a boast about keeping people awake.

The design intent is real enough without the embellishment. Netflix only added the ability to switch off autoplaying previews in February 2020, after years of complaints.

What about the research on television and dying earlier

You may have seen this one. A meta-analysis of ten cohort studies covering 647,475 people found that those in the highest category of television viewing had about a third higher all-cause mortality than those in the lowest.

Handle it carefully. These are observational studies, and television viewing is bound up with income, education, employment, physical activity and being ill already, which is a serious reverse-causation problem: people who are unwell sit and watch more. The dose-response curve was not a straight line, so this is not an hourly tax on your lifespan. It is a signal concentrated at the heaviest end.

What actually to do

In descending order of how much evidence sits behind it.

1. Turn autoplay off. The one thing here with a randomised experiment behind it, and it takes a minute. Do it on every device and every profile, because both services handle the scope differently.
2. Decide the end before you start. The strongest distinction in the research is between planned and unplanned viewing, and the loss of control sits with the unplanned kind. Choosing two episodes in advance is a different activity from finding out afterwards that you watched five.
3. Put a gap between the last episode and bed. The sleep finding that survived was about a mind still running, not about the screen, so the useful move is to end on something that lets the story settle rather than on a cliffhanger at midnight. The rest of the bedroom setup matters more than any of this.
4. Watch the pattern rather than the count. Four episodes of something you chose is not the same event as four episodes you did not decide to watch, even though the clock cannot tell them apart.

Two things you may have read that are not supported by anything reliable: that watching with other people reduces how much you watch, and that feeling regret afterwards will change your behaviour. Neither has evidence behind it, and in the research, social motives are among the reasons people binge rather than a brake on it.

When it is not really about the television

Streaming is unusually good at absorbing an evening you did not want to spend thinking. If what you are avoiding is a decision, a conversation or a feeling, autoplay is not the mechanism and turning it off will not touch it.

If your sleep stays bad after you have moved the viewing earlier, or if you are watching to avoid something you already know about, those are worth addressing directly rather than through a settings menu. Persistent insomnia in particular has plenty of causes no amount of screen hygiene will fix, and is worth taking to a doctor. Nothing here is medical advice.

If you want to go further than one setting, our digital declutter guide covers doing this across everything rather than one service at a time.