Friction Engineering

A simple explanation

Friction engineering is the deliberate practice of making unwanted behaviours slightly harder and desired ones slightly easier — not by changing who you are, but by changing what your environment asks of you. A twenty-second delay before Instagram opens. A food delivery app with no saved payment method. A TV removed from the bedroom. A book left open on the kitchen counter.

The mechanism is small and structural. It does not require willpower; it changes what the behaviour costs to start. The friction itself is not the meaning. The meaning lives in what the friction protects, or in what the friction allows to land.

An everyday example

You install a twenty-second delay on Instagram. The first week, the delay catches you four or five times a day. You wait, the impulse dims, you put the phone down. The system works.

The third week, you stop noticing the delay. The fourth week, you've learned to start the timer and switch to email while you wait, returning to Instagram exactly when it unlocks. The friction is still installed. The bypass has become a small loop of its own. The deposit the friction was protecting — the attention you wanted to reclaim — is leaking again.

The design did not fail. The design assumed install-once. Friction engineering is not install-once.

What is friction engineering?

Friction engineering is a subset of environment design and a specific application of choice architecture. Where environment design is the broad practice of arranging your surroundings to bias behaviour, friction engineering is the surgical instrument: the deliberate addition or removal of effort steps on the path between trigger and action.

The interventions are small. The cumulative effect is not. A bowl of fruit on the counter and chips on a high shelf is friction engineering. A separate laptop with no social apps installed is friction engineering. A workout bag packed the night before is friction engineering. Each one moves a single effort step closer to or further from the behaviour, and the body, encountering that step in real time, makes a different decision than it would have otherwise.

The behavioral loop

How a friction intervention runs, end to end:

1. Cold-State design — you, in a reflective moment, identify a loop you want to interrupt or a deposit you want to enable. You install a structural change: a delay, a removal, a placement, a deletion.
2. First contact — Hot-State You hits the friction. The behaviour is not impossible; it is just harder. In many cases, the impulse cools across the delay and the action does not run.
3. Honeymoon period — for one to four weeks, the friction holds. The unwanted loop's frequency drops. The desired loop's frequency rises. The body notices the change.
4. Adaptation — the nervous system, encountering the same barrier repeatedly, begins to route around it. The bypass costs effort, but the reward is preserved, so the loop optimises for the bypass.
5. Decay — the friction is still installed, but its effective cost has fallen. The behaviour returns toward baseline, often with a low-grade frustration tax added by the bypass itself.
6. Maintenance fork — either the design is refreshed (new friction, different friction, escalating friction) or the original loop reasserts itself. There is no third option that runs on its own.

The honeymoon is real. The honeymoon is also not the whole story.

Emotional drivers

Three pulls converge on friction engineering, and they do not always cooperate.

The first is the desire to not be the kind of person who needs friction — to have the inner discipline that makes the design unnecessary. This is a Reward System story about competence, and it sometimes prevents installation.

The second is the relief of the decision already made — the felt ease, after installation, of not having to fight the impulse in real time. The Avoidance System, which fears having to deploy willpower, registers this as safety.

The third is the small grief of seeing the loop honestly — installing friction is a Cold-State admission that Hot-State You is not trustworthy on this front. That admission is healthy, but it costs something.

What your nervous system does

Hot-State and Cold-State are not metaphors. They are different operating regimes of the same nervous system, distinguished by arousal, time pressure, and how much of the prefrontal cortex is currently online.

Cold-State You has bandwidth to model the future, weigh second-order consequences, and notice patterns. Cold-State You designs the friction. Hot-State You has narrower attention, faster reward circuits, and a strong present-bias. Hot-State You hits the friction.

The twenty-second delay works because it is structured to outlast the impulse. The hedonic spike that drove the phone-pickup decays across roughly that window. The friction is not blocking the behaviour; it is buying time for the slow system to catch up with the fast one. This is also why frictions shorter than ten seconds barely work and frictions longer than two minutes start to feel punitive rather than protective.

The DojoWell interpretation

Friction engineering operates on the denominator of the Meaning Density Equation.

Density = (Deposit − Residue) ÷ Effort. The numerator is what the action leaves; the denominator is what it costs to do. Friction engineering does not change what an action leaves. It changes what an action costs to start. That single move is enough to bend behaviour, because B = MAP — behaviour fires when Motivation, Ability, and a Prompt converge, and Ability is what effort buys.

Raise the Effort term on a hollow-reward loop and the verdict tips: the deposit was already near-zero, the residue was already accumulating, and now the denominator is large enough to make the whole expression visibly low-density before the action runs. The Cold-State design front-loads the reading the slow system would have arrived at anyway.

Lower the Effort term on a delayed-harvest deposit — pack the bag, leave the book open, put the meditation cushion in the middle of the room — and a real but quiet deposit becomes accessible to a Hot-State You that would not have summoned the activation energy on its own.

The trap is treating friction engineering as install-once. The substitute is one-off friction without ongoing maintenance. It wears the outer shape of the real thing — barriers are in place, the design looks finished — but the bypass loop learns, the friction decays, and the original behaviour returns wearing the same clothes. Effort was paid in design, the deposit lands for three weeks, residue accumulates in the form of bypass-frustration, and the verdict drifts from high to medium to low. The friction itself becomes a low-density loop.

Friction engineering also exposes a quiet truth about the four Systems: friction does not negotiate with any of them. It does not argue. It does not appeal to reasons. It simply changes the cost of the path. This is why it is often more durable than insight, willpower, or motivation in the short run — and why it requires the slowest of the four (the Meaning System) to refresh it in the long run.

Why does the friction stop working after a few weeks?

Because behaviour, like water, finds the lowest path it can. The friction raised one path; the system searches for another. Sometimes the bypass is intentional (a second device, an incognito tab, a deleted app reinstalled at midnight). More often it is invisible — you simply learn the rhythm of the twenty-second delay and use it for something else, returning to the original behaviour at exactly the moment the gate opens.

The deeper reason is that friction is a structural intervention, and structures decay when the conditions around them change. Your relationship to the behaviour shifts. The reward changes shape. The trigger relocates. The friction, installed against last month's loop, holds the line against a loop that has already moved.

This is not a failure mode. It is the operating reality. Friction engineering is closer to gardening than to construction. The bed has to be tended.

Practical steps

1. Start with the loop, not the friction. Name the behaviour you are trying to interrupt or enable. Name the specific trigger, the specific action, the specific reward. Friction installed against a vague loop will not hold.
2. Calibrate the delay to the impulse's half-life. Twenty seconds for scroll loops. Two minutes for sugar cravings. The night-before for morning habits. Frictions that outlast the impulse work; frictions inside the impulse don't.
3. Prefer subtraction to addition. Removing the TV from the bedroom is more durable than adding a sleep app. Deleting the payment method is more durable than installing a budgeting tracker. Subtractions decay more slowly than additions because there is nothing for adaptation to optimise.
4. Plan the maintenance, not just the install. Put a recurring check on the calendar — every four to six weeks — to ask whether the friction is still doing its job. Adjust, escalate, or remove.
5. Do not over-friction. A house designed entirely against your impulses becomes a hostile environment. Two or three load-bearing frictions, well-tended, outperform a dozen brittle ones.
6. Notice when the bypass becomes a loop of its own. If the workaround you've developed costs you more attention than the original behaviour did, the friction is now negative-density. Redesign.
7. Use friction to protect deposits, not just block substitutes. Lowering effort on the cushion, the page, the call, the walk is the under-used half of the practice. Most people deploy friction asymmetrically.

Reflection questions

• Which of your current habits — good or bad — runs primarily on whether the activation energy is two seconds or twenty?
• Where in your environment has a friction you installed quietly stopped working, without you noticing?
• What is one friction-removal you could install this week that would make a delayed-harvest deposit easier to reach?
• Is there a loop you keep trying to beat with willpower that would yield in one afternoon to a structural change?

Frequently Asked Questions