46% of People Who Intend to Exercise Never Follow Through — What Changes It Is How You Write the Plan
BaroSit · 2026-07-27 · 📝 블로그
“I’ll start working out next month.” How many times has that actually turned into workouts? When it fizzles, we tend to blame willpower. Researchers who have studied this particular failure point somewhere else.
1. Nearly half of the people who mean to exercise don’t — and that number has been measured
There is an analysis pooling 10 studies with 3,899 adults. It singled out the people who had already decided to exercise, then checked whether they actually did it. Among those intenders, 46% never followed through. The authors concluded that intention on its own is a weaker predictor of behaviour than our models assume. Put differently: failing to act isn’t the exception, it’s close to half the room.
2. Adding more motivation did not close the gap
So would wanting it more help? A study of 248 people separated exactly that question. One group received an intervention designed to raise motivation — the risks of inactivity, the reasons to move. Another group got the same thing plus one extra instruction: write down when, where and how you will do it. The motivation-only group showed no meaningful increase in exercise. Behaviour changed only in the group that did both. It ran two weeks on self-report, so it is not a long study, but the direction is clear. What differed was not how much people wanted it. It was whether they had settled on where that wanting would switch on.
3. What made the difference was the format of the plan
The method is simpler than it sounds. Instead of a goal like “exercise more,” you write one line in the form “If (this situation), then (this action).”
• Get more exercise → When I take my shoes off after work, I change straight into workout clothes
• Drink more water → When a meeting ends, I fill a glass on the way back to my desk
An analysis pooling studies of this method reported d=.65 across 94 independent tests — a medium-to-large effect. A larger analysis published in 2024 pooled 642 tests and put the effect between .27 and .66, finding it was bigger when the plan used the if-then format, when people genuinely wanted the goal, and when they rehearsed the plan. Those figures average across many domains, so the effect is smaller when you look at physical activity on its own, though it points the same way. The point is not how badly you want it. It is deciding in advance what counts as the cue, and what you do when it arrives.
4. So how do you use this today
If you already have a plan and you keep it — leave it alone. If you change anything, check the shape of the sentence. If it reads as a frequency (“three workouts a week”), add the moment that will trigger it (“when I put my bag down after work”).
If you make plans but they keep collapsing — put the collapse itself into the plan. One condition that produced bigger effects in the analysis above was planning for the obstacle, not only for the behaviour. If “I can’t when I work late” keeps repeating, write the answer ahead of time: “If I’m still working past eight, I do ten minutes at home.”
If you have no plan at all — make one, not five. Attach it to the most reliably repeating moment of your day. If it doesn’t stick, make the situation more specific rather than adding more plans.
In the end, the useful question is less “how determined am I?” and more “where and when have I arranged for that to switch on?” The effect isn’t dramatic, but the cost is a single sentence, which makes it a reasonable trade.
None of these studies tested posture, or breaking up long stretches of sitting. But it is the same problem underneath — turning something you decided into something you do — so it is worth applying the same approach there. What BaroSit does sits on the cue side of it: telling you when one position has gone on too long. What you do with that signal is better decided in advance, in the way described above.
→ Have a look at barosit.com if you’re curious.
The evidence on sitting time and movement is collected at barosit.com/science.
Sources
• Rhodes & de Bruijn, 2013 · Br J Health Psychol 18(2):296–309 — 10 studies, N=3,899; intention–behaviour gap 46%
• Gollwitzer & Sheeran, 2006 · Adv Exp Soc Psychol 38:69–119 — 94 independent tests, d=.65
• Sheeran, Listrom & Gollwitzer, 2025 · Eur Rev Soc Psychol 36(1):162–194 — 642 tests, .27 ≤ d ≤ .66
• Bélanger-Gravel, Godin & Amireault, 2013 · Health Psychol Rev 7(1):23–54 — 26 physical-activity studies; 0.31 (95% CI 0.11–0.51) post-intervention, 0.24 (0.13–0.35) at follow-up
• Milne, Orbell & Sheeran, 2002 · Br J Health Psychol 7(2):163–184 — n=248
This article is general health information, not medical advice. If pain persists or worsens, please consult a professional.