What does a routine protect that a rule cannot?

Why this matters. Rules are consulted, which requires remembering to consult them. Routines run whether or not you remember.

A rule says what to do when a condition occurs. It depends on you noticing the condition and choosing to apply it — both of which are exactly the faculties this domain has spent two courses showing to be unreliable under load.

What a routine does differently

It attaches an action to a time or an event rather than to a judgement. The monthly holding review from earlier in this course happens because it is the first Monday, not because something felt worth reviewing.

That difference matters most precisely when judgement is compromised: the review most worth having is the one you least want to do.

What the evidence supports, and what it does not

Replicated, with a caveat worth stating: Lally and colleagues found automaticity developing over a wide range — from 18 to 254 days depending on the person and the behaviour. The useful part is the range itself. "Twenty-one days to form a habit" is folklore and the study that is cited for it does not say that.

So a routine takes an unpredictable and often long time to become automatic, and expecting it in three weeks is how people conclude they have failed at something that was proceeding normally.

What a routine is not for

It is not a substitute for the rules — Course 2's artefacts still have to exist. A routine is what makes them get read.

The artefact

A written routine with times attached: what happens daily, what happens monthly, what happens after a loss beyond a threshold.

Three or four items. A routine long enough to be impressive is a routine that will not survive a difficult month, and a difficult month is the only one it is for.

Try it now

Write three items with times. Then remove one — the version you keep is shorter than the version you first write.