Think about how you got to work this morning. You almost certainly don't remember the individual turns, the order you put your shoes on, or the moment you reached for coffee. You weren't being lazy, you were running habits, and your conscious mind was free to think about the meeting you were dreading instead. That offloading is the whole point of a habit, and it is exactly why "just decide to do better" so rarely works.

The quick version

  • A habit is a behaviour that has become automatic, triggered by a context cue rather than a fresh decision each time.
  • Habits form through repetition in a consistent context, not through motivation or insight. The behaviour you repeat in the same place at the same time is the one that sticks.
  • The honest number is "it depends": one well-known study found a median of about 66 days to reach peak automaticity, but the real range ran from roughly 18 to 254 days.
  • The leadership move is to stop relying on willpower and design the cue and the reward, make the good behaviour obvious and easy, and the bad one awkward.

The idea in depth: a habit is a cue, not a choice

The first thing to unlearn is that habits are about discipline. They are about context. In a pair of diary studies in which people reported what they were doing every hour, Wendy Wood and colleagues found that a large share of everyday behaviour was performed almost daily, in the same place, while the person was thinking about something unrelated, roughly 43% of actions, on their figure (Wood, Quinn & Kashy, 2002, Journal of Personality and Social Psychology). The behaviour wasn't being chosen each time. It was being cued, by the kitchen, the commute, the open laptop, and then run on autopilot.

That reframes what a habit actually is. The popular shorthand, made famous by Charles Duhigg's The Power of Habit (2012) and grounded in Ann Graybiel's neuroscience work at MIT, is the habit loop: a cue triggers a routine, which delivers a reward. Graybiel's lab watched this happen in the brain, as a behaviour became habitual, the activity that controlled it shifted toward the basal ganglia and "chunked" into an automatic sequence, with the brain firing mostly at the start (the cue) and the end (the reward) and going quiet in between. The conscious, effortful part drops out. That's why a habit feels effortless once it's formed, and why it's so hard to stop on command: the cue still fires whether or not you've resolved to change.

So the move is: stop hunting for motivation and go looking for the cue. Before you try to build or break a habit, name the trigger, the specific time, place, preceding action or emotional state that sets it off. You can't reliably control a behaviour you only meet at the routine stage; you can often redirect it at the cue. The leader who wants their team to write better standup notes doesn't give a speech about discipline, they tie the note to an existing cue (the calendar reminder that already fires) and make the template one click away.

The idea in depth: repetition in a stable context, not raw willpower

If cues trigger habits, repetition is what builds them. The cleanest real-world study of this is Phillippa Lally's, which tracked 96 people who picked a simple daily behaviour, a glass of water with lunch, a short walk after dinner, and rated how automatic it felt each day for twelve weeks (Lally et al., 2010, European Journal of Social Psychology). Automaticity climbed along a curve that rose fast at first and then flattened, early repetitions buy you the most, and missing a single day didn't wreck the process.

This is also where the famous "66 days" comes from, and where it gets abused. Sixty-six days was the median time for participants to reach their plateau of automaticity. The actual range stretched from about 18 to 254 days. So the popular "21 days to a habit" rule is a myth, and even "66 days" is a midpoint, not a deadline. How long it takes depends on the behaviour, the person and how consistent the context is.

Early repetitions buy you the most. Miss a single day and the habit survives; it's the broken context that kills it.

So the move is: protect consistency over intensity. A behaviour repeated in the same context every day will automate faster than a bigger version done erratically. Practically, that means pinning the new behaviour to a fixed anchor (after I sit down, before the first call) rather than a floating intention ("sometime today"), and treating a missed day as a data point rather than a failure. For a team, it means not redesigning the ritual every sprint, the standup that happens at the same time, same channel, same shape is the one that becomes muscle memory.

flowchart LR
  A(["Cue
(context, time, prior action)"]) --> B(["Routine
(the behaviour)"])
  B --> C(["Reward
(felt payoff)"])
  C -.->|"strengthens the
cue–routine link"| A
The habit loop: a cue triggers a routine that delivers a reward, and the reward wires the cue more tightly to the routine. Leaders Loop

The idea in depth: make it tiny, and make the reward immediate

The third lever is design. B.J. Fogg, who ran the Behavior Design Lab at Stanford, argues in Tiny Habits (2019) that any behaviour happens when three things land at once, Motivation, Ability and a Prompt (his B=MAP model). Motivation is unreliable: it spikes and crashes. So Fogg's advice is to lean on the other two, shrink the behaviour until it needs almost no motivation (floss one tooth; write one sentence), and attach it to a prompt you already have. His formula is "After [an existing routine], I will [a tiny behaviour]," followed immediately by a small celebration, a felt moment of "yes" that, in his account, helps wire the behaviour in.

That celebration point connects back to the loop: the reward has to be felt, and it has to be fast. A payoff that arrives months later, the promotion, the weight loss, is too distant for the brain's habit machinery to attach to the cue. One honest limitation, though. The immediate-reward principle is well-supported, but Fogg's specific "celebrate to install a habit" technique rests more on his practitioner experience than on large independent trials. And habit research as a whole has done far more with simple personal behaviours (water, walking, flossing) than with the messy, multi-step work behaviours leaders actually care about. Treat the models as a reliable lens, not a guarantee.

So the move is: shrink the behaviour and pull the reward forward. If you want a habit to take, make the first version embarrassingly small so it survives a bad day, anchor it to something you already do, and build in a payoff you can feel today, visible progress, a tick on a list, a teammate's acknowledgement, rather than one that only pays out next quarter.

flowchart TD
  S(["Want to change a behaviour?"]) --> Q{"Build a new one,
or break an old one?"}
  Q -->|Build| B1(["Anchor it to a cue
you already meet"])
  B1 --> B2(["Shrink it until it's
almost too easy"])
  B2 --> B3(["Add an immediate,
felt reward"])
  Q -->|Break| K1(["Find the cue
that triggers it"])
  K1 --> K2(["Remove the cue, or add
friction to the routine"])
  K2 --> K3(["Attach a better routine
to the same cue"])
Same machinery, two directions: building a habit and breaking one both start at the cue, not the willpower. Leaders Loop

A worked example

Maya runs a support team that keeps shipping the same fixes twice because nobody writes down what they learn. She's tried the obvious thing, asking people, in the all-hands, to "document as you go." It works for a week, then evaporates. (Illustrative scenario; the figures below are made up to show the shape of the approach.)

So she stops appealing to discipline and designs the loop instead. The cue: she ties documentation to a moment that already happens every day, closing a ticket. The behaviour can't begin until the trigger exists, so she makes the trigger inescapable: the "close ticket" button now opens a one-line box, "What did you learn?" The routine is deliberately tiny, per Fogg, one sentence, not a write-up, so it survives a busy afternoon. The reward is immediate and felt: the answer auto-posts to the team channel, and Maya makes a point of reacting to good ones within the hour, so the payoff arrives in minutes, not at review time.

For the first fortnight, completion is patchy, say, around 40% of tickets (an illustrative figure). Maya resists the urge to add a rule or a nag. Consistency, not intensity, is what Lally's curve rewards, so she keeps the cue, the size and the reward identical day after day. By week six it's the default; people feel the small wrongness of closing a ticket without the line, the way you feel a seatbelt you forgot. She never gave a second speech, she changed the cue, shrank the routine, and moved the reward closer. As our explainer on behavioural levers covers, redesigning the default quietly beats appealing to willpower loudly.

Frequently asked questions

How long does it really take to form a habit?

There's no fixed number, and anyone who gives you one is overselling. The most-cited real study (Lally et al., 2010) found a median of about 66 days to reach peak automaticity, with a range from roughly 18 to over 250. Simple behaviours in stable contexts go faster; complex ones take longer. The "21 days" rule has no good evidence behind it.

Is willpower useless, then?

Not useless, just the wrong tool for the recurring stuff. Willpower is a scarce, fluctuating resource; habits exist precisely so you don't have to spend it. Use deliberate effort to set up the cue and the environment once, then let the habit carry the behaviour so you're not re-deciding every day.

Why do my new habits collapse when life changes?

Because habits are tied to context, and a disrupted context breaks the cue. A house move, a new role, a holiday, each removes the triggers the behaviour was hanging on. That's also the opportunity: moments of change are when old habits weaken and new ones are easiest to install, before fresh cues set.

Can I build habits across a whole team, not just myself?

Yes, and the levers are the same, just made shared. Pin the behaviour to a cue everyone already meets (a recurring meeting, a stage in the workflow), make the action small enough to survive a bad week, and give a fast, visible payoff. Teams form habits through repeated, consistent rituals far more than through policies nobody reads.

How do I break a bad habit?

You rarely delete a habit; you redirect it. Find the cue, then either remove it from your environment or attach a different routine to it. Making the bad behaviour harder (friction) and the better one easier usually beats trying to resist the urge at the moment it strikes, by then the cue has already fired.

Related in the Toolkit

Where to go next