Why Doing a Habit With a Friend Works: What Studies Show

A friend helps in a few specific ways. People exercise longer beside a partner, even one on a screen. People who joined a program with friends finished it more often. Seeing friends be active makes people more active. The catch: encouragement alone often did little, and announcing a big goal can backfire. Show up, and let each other see it.

The Köhler effect is the extra push you get from working out with a partner who is a little stronger than you, when your effort counts toward a shared result. In these experiments, the pair's score is set by the weaker of the two, often whoever quits first. If you're the weaker one, the team result is in your hands, and people rise to that.

A 2023 meta-analysis in the journal Kinesiology Review pooled 19 studies with 1,912 people and found a large effect: g = 0.91. People exercising with a partner like that put in far more effort than people exercising alone. Human partners also worked better than partners made by software, a 2014 study in Games for Health Journal found.

Now the honest part. These were lab tasks, like planks and cycling, that lasted minutes, not months, and most of the people were students. A 2019 study with middle-aged adults found no effect on how hard they worked. So the Köhler effect is a real push in the moment. Nobody has shown that it lasts all year.

In the lab, yes. In a 2012 trial published in Annals of Behavioral Medicine, 58 young women rode a stationary bike on six separate days. Alone, they lasted 10.6 minutes on average. Next to a partner shown on a screen, each riding at her own pace, they lasted 19.8 minutes. When the partner was a bit stronger and the pair's score depended on whoever stopped first, they lasted 21.9 minutes.

That's close to double, just from someone else being there. The partner wasn't even in the room, just someone on a screen doing the same thing at the same time.

A 1999 study in the Journal of Consulting and Clinical Psychology followed 166 adults in a group weight-loss program. Of the people who joined alone and got the standard program, 76% finished. Of the people who joined with friends and got extra support, 95% finished.

Two limits are worth knowing. People chose whether to bring friends, so that part wasn't randomized, and these were adults meeting in person. Still, it's a big gap, and it points the same way as the rest of this research: starting next to people you know makes it easier to reach the end.

It seems to. A 2017 study in Nature Communications followed about 1.1 million runners in a global fitness network for five years, and used the weather as a natural experiment to separate real influence from friends simply being alike. When friends ran 1 km more, people ran about 0.3 km more. Less active runners pushed more active ones, but not the other way round. One limit: these were people who already tracked their runs.

Seeing matters more than cheering. A 2016 trial in Preventive Medicine Reports put 790 graduate students into anonymous groups of six for 11 weeks. Groups who could compare their attendance with peers went to about 90% more exercise classes than groups who couldn't. The group that only got support, with a chat and a team prize, did worse than the control group.

A larger trial pointed the same way. In STEP UP, published in JAMA Internal Medicine in 2019, 602 adults joined game-style step programs for 24 weeks. Compared with a control group, the competition version added 920 steps a day, the support version 689 and the collaboration version 637. Twelve weeks after the program ended, only competition still made a real difference, at 569 extra steps.

Even kind words can get in the way. In a 2013 trial in the Journal of Medical Internet Research, with 115 people, an online partner added 79 seconds to how long people held a plank. When that stronger partner also sent encouragement, 31 of those seconds disappeared. Friendly comparison, the quiet kind where you just see what your friend did today, is the part with the most evidence.

A commitment can work, but few people sign up for one. In a field trial published in 2010 in American Economic Journal: Applied Economics, smokers were offered a commitment contract to quit, with their own money at stake. Only 11% took it. Even so, smokers who were offered the contract were 3 percentage points more likely to pass a nicotine test at six months, and the gain held on surprise tests at 12 months.

The lesson: a promise with a clear end and a clear test can change behavior, as long as it's light enough that people actually make it. A pact between friends puts only your name and your friend's attention at stake. Nobody has tested that version the same way, so don't expect the same numbers. It is a much easier yes, though.

The risk is letting the announcement feel like the achievement. In four small experiments published in Psychological Science in 2009, people strongly committed to an identity goal, such as becoming a lawyer, acted on it less after someone else noticed their intention. The researchers' explanation: being noticed gave them a premature sense of already being that person. These studies were small, and we found no large replication, so hold the finding loosely.

The fix is simple. Commit to showing up, not to an identity. 'I'll post a photo of my run every day for 30 days' is something you prove one day at a time. 'I'm a runner now' is credit you get before doing anything.

First, a group isn't magic. A 2026 meta-analysis of 71 studies in Nature Human Behaviour found no significant difference in physical activity between group programs and individual ones, online or in person. And a Cochrane review of 14 trials found that programs adding a partner's support to quit smoking had no effect, mostly because they never raised the support people actually got.

A buddy on paper does nothing. What the studies above point to is a friend who actually sees you do it. In practice: pick one friend you're close to, ideally someone a little ahead of you. Choose the same habit. Agree on a set number of days, so there's a finish line. Then each of you proves it every day in a way the other can see.

Keep the talk short and the proof visible. And if one of you misses a day, just come back the next one. In a well-known 2010 study in the European Journal of Social Psychology, missing a single day didn't materially affect how a habit formed.

Kairo is an iPhone app for building daily habits with your friends, and its pact follows this research. Two friends put their names on the same habit for a set number of days and see each other show up until the last day. A day counts when you prove it, most often with a photo taken in the moment.

No study has tested Kairo itself. It's built to line up with what the studies do show: a close friend instead of a crowd, a clear end date, and proof you can see instead of a promise you announce. The point isn't the signing. It's the proof, every day. Kairo is free.