Why most of what people do is triggered by their surroundings, and why the easiest way to change behaviour is to change the cue, not the person.
Modern habit research agrees on an apparently simple definition: a habit is a learned association between a stable context and a response. When the cue recurs, the behaviour is triggered with little deliberation. Frequency alone isn't a habit. The diagnostic feature is automaticity (it runs efficiently, with low awareness, low control, and no real intention). Wood and Neal's 2007 model in Psychological Review is the modern foundation, and Wood and Rünger's 2016 Annual Review of Psychology synthesis is the best single overview.
The single most important finding for practitioners: context stability is the primary driver of habit strength: same time + same place + same preceding action = same following action. There's even a neural signature. As a behaviour becomes habitual, control migrates from the brain's goal-directed circuitry to cue-driven, sensorimotor circuits (Yin & Knowlton, 2006). The behaviour quite literally stops being "decided."
The popular cue → routine → reward "habit loop" (from Charles Duhigg's bestseller “The power of habit”) is clean and memorable, and partly wrong. The contested piece is the reward. Once a habit is strong, the behaviour often persists even when the reward is removed or devalued. In the lab this is the defining test of a habit: it becomes insensitive to outcome value (Yin & Knowlton, 2006; Wood & Rünger, 2016). Reward matters a lot while a habit is forming, but much less in keeping it running. So the tidy story "find the reward, change the routine" oversells how central reward is to an established habit.
The famous "66 days" number comes from Lally et al. (2010), who tracked 96 people forming everyday habits. Sixty-six days was the median time to automaticity, but individuals ranged across a very wide span depending on how complex the behaviour was. Treating 66 (or, worse, 21) as a law misreads the evidence: the honest takeaway is "longer than you think, and it depends." There's also a measurement problem: the standard tool (the Self-Report Habit Index) infers habit strength partly from how often you do the thing, so "habit predicts behaviour" can be a little circular (Verplanken & Orbell, 2003).
We still lack a clean way to measure habit strength independently of behaviour. We don't fully understand why some behaviours automate quickly and others crawl. And we don't know how durable "window of change" interventions are — most studies follow up over weeks, not years.
Here’s the unsettling bit: after you understand that “blue is left and orange is right” you stop making choices. You just react to the “environment”, in this case the cue you get from the “room” background’s colour. You stay alert, you click, you feel like you chose, but it's just reflex; there's no choice left in there.
So a habit is like an automatism: you do it because your brain picked up this “pattern” and reacts without other kinds of evaluations.
Habits are the mechanism behind an enormous share of everyday behaviour: repeat purchases, commutes, media use, voting routines, health behaviours. If you treat that behaviour as a decision to be persuaded, you'll usually lose. The leverage is in cues and context, and especially in moments when context breaks.
Most repeat buying is habitual, not deliberative: people re-buy without re-deciding. Think about it. How much thought do you put into that product you always buy at the supermarket where you always get your groceries? That has two consequences. First, defending a habit (your customers') means protecting the cue: consistent placement, packaging, routine, and friction-free repurchase (auto-renew, default delivery, one-tap reorder) does more than another persuasion campaign. Second, breaking a competitor's habit is hard unless the customer's context just changed, so targeting life-transition moments (new home, new baby, new job, new device) when old cues are gone and switching is a valuable tactic.
Turnout is substantially habitual: people who vote tend to keep voting, in the same place, the same way. So mobilisation is partly cue-management: stabilise the act (same polling routine, reminders tied to existing routines, "make a plan" prompts that attach voting to an existing cue). And persuadable voters are disproportionately those whose circumstances recently changed (new residents, new graduates, the newly retired) because a disrupted context reopens questions that habit had closed.
Behaviour-change campaigns aimed at the general population tend to underperform because they fight cue-driven behaviour with information. The evidence points somewhere more precise: time interventions to moments of disruption. A field experiment found a sustainability programme worked significantly better on households that had recently moved (Verplanken & Roy, 2016). Practically: reach people at the registry office, the new-tenancy pack, the first-job onboarding, the post-natal visit: the windows where habits are already loosened and new ones haven't set.
Stop asking "how do I motivate this behaviour?" and ask three mechanical questions instead: (1) What is the cue? (2) How stable is the context that delivers it? (3) Is there a context change I can ride, or engineer? Reduce friction for behaviours you want to stick; add friction or remove cues for behaviours you want to break; and spend your intervention budget where the context is already moving.
.png)