Motivation & Habits: Why You Do What You Do (and How to Change It)

Roughly 40% of your daily actions aren’t decisions at all — they’re habits, running on cues while your mind is elsewhere. That’s the single most useful fact in behaviour change, because it means the game isn’t “become more motivated.” Motivation is the ignition; habit is the engine. This doc covers both systems — what actually drives behaviour, and the mechanics of rewiring it.

Two kinds of motivation (and how they interact)

• Intrinsic — the activity is its own reward: play, curiosity, craft.
• Extrinsic — the activity is instrumental: salary, grades, likes, gold stars.

The interesting finding is that they don’t simply add. The overjustification effect: pay children to draw — something they already loved — and when the payments stop, they draw less than kids never paid at all. The reward converted play into work; the mind re-filed “why I do this” from because I love it to because I’m paid, and removing the pay removed the reason. This replicates best when rewards are expected and tied to mere participation; unexpected rewards or praise for quality do much less damage. Design implication: be careful gamifying what people already enjoy.

Self-determination theory: the three nutrients

The most robust framework for sustained motivation (Deci & Ryan). Intrinsic motivation thrives when three needs are met, and starves when they’re not:

1. Autonomy — I chose this; it isn’t imposed.
2. Competence — I’m getting visibly better; the challenge fits my skill.
3. Relatedness — this connects me to people I care about.

Audit any motivation collapse — a job gone stale, an abandoned hobby, a disengaged team — and you’ll almost always find one of the three missing. (Games are motivation machines precisely because they industrialize all three: chosen goals, calibrated difficulty curves, guilds.)

What dopamine actually does

Pop culture calls dopamine the “pleasure chemical.” Wrong — and the correction matters:

• Dopamine signals reward prediction error: it fires for rewards that are better than expected, stays flat for fully predicted rewards, and dips when an expected reward doesn’t arrive. It’s a learning signal — “update your predictions” — not a pleasure meter.
• The systems for wanting (craving, drive — dopamine) and liking (enjoyment — different circuitry) are separate. This is why you can compulsively scroll a feed you don’t even enjoy: the wanting system is fully engaged while the liking system shrugs.

Now the design-consequence everyone should understand: variable rewards. Skinner found that unpredictable reinforcement produces the most persistent, extinction-resistant behaviour — a rat rewarded sometimes presses the lever far longer than one rewarded always. Prediction error explains why: variability keeps the “better than expected!” signal alive forever. Slot machines run on this. So does pull-to-refresh, and every feed where maybe the next scroll has something great. Once you see the mechanism, you can’t unsee it — in the products you use, and the ones you build.

Habits: the cue–routine–reward loop

A habit is a behaviour that has migrated from deliberate (System 2) to automatic (System 1), triggered by context rather than choice:

Facts about the loop worth knowing:

• “21 days to form a habit” is a myth. The real research (Lally et al.) found a median of ~66 days, ranging from 18 to 254+ depending on the behaviour’s complexity. Drinking a glass of water automates fast; 6am gym does not.
• Missing one day doesn’t matter. The same study found single lapses had no measurable effect on habit formation — what kills habits is the “what the hell” spiral after a miss, not the miss.
• Habits don’t die; they get overwritten or de-cued. The old wiring stays (why relapse is easy in old contexts) — you either avoid the cue or attach a new routine to it.

The evidence-backed toolkit

Ranked roughly by strength of evidence:

1. Implementation intentions — the single best-supported trick in the literature (meta-analyses show medium-to-large effects). Replace vague goals with a cue-wired plan: “When X happens, I will do Y.” Not “I’ll exercise more” but “after I close my laptop at 6, I put on running shoes.” You’re pre-installing the cue→routine link instead of hoping in-the-moment motivation shows up.
2. Environment design — the strongest lever in practice. Habits run on cues, so edit the cues: fruit on the counter, phone charging in another room, junk food not in the house (the decision happens once, in the store, when you’re strong — not thirty times a night when you’re weak). Friction is destiny: make good behaviours one step easier and bad ones one step harder.
3. Habit stacking — anchor new habits to existing ones (“after I pour my morning coffee, I write one sentence in my journal”). Existing habits are free, reliable cues.
4. Start absurdly small — two minutes, one pushup, one sentence. The goal early on is casting the vote, not the workout; automaticity grows from repetitions, and repetitions require showing up. Scale intensity only after showing up is automatic.
5. Make it satisfying immediately — the loop needs a reward now, not in ten weeks. Tracking streaks works because the checkmark is the immediate reward (and don’t-break-the-chain adds loss aversion to your side).
6. Goal gradient — motivation accelerates near a finish line; people with a 10-stamp coffee card that comes with 2 free stamps complete it faster than an 8-stamp card. Break long projects into near-finish-lines.

What the evidence is against

Willpower as a strategy. Beyond ego depletion’s replication troubles (see the overview), the deeper finding is that people who look highly self-controlled mostly experience fewer temptations — they’ve arranged their lives so the fight rarely starts. Discipline is downstream of environment. And motivation is a state, not a trait — it fluctuates daily, which is exactly why systems that don’t depend on it (cues, defaults, stacks) outperform systems that do.

Takeaways

• Protect intrinsic motivation: audit for autonomy, competence, relatedness; gamify carefully.
• Dopamine = prediction error, not pleasure; variable rewards are why feeds are sticky — respect the mechanism as a user and wield it ethically as a builder.
• Habits are cue-driven: design the environment, wire if-then plans, stack on existing routines, start tiny, reward immediately.
• ~66 days, misses don’t matter, spirals do.

Next: the strongest force acting on your behaviour that hasn’t appeared yet — other people.