KEVSTEMPLATES.COM
Habit Loop and Behavioural Drift
Adoption spiked then fell, or you're designing for the long term
Pick the shape your adoption curve made and the signs you recognise, and it names which of the four causes of drift you have and the structural fix.
Roughly 40% of daily workplace actions are habitual rather than decided. You are not competing with people’s opinions — you are competing with their automaticity.
The loop
CUE ──────► ROUTINE ──────► REWARD
▲ │
└────────── CRAVING ◄───────────┘
- Cue — the trigger. Time, place, preceding event, emotional state, other people.
- Routine — the behaviour.
- Reward — what the brain gets. Relief, completion, status, avoiding effort.
You cannot delete a habit. You can only replace the routine while keeping the cue and reward. This is why “stop using the spreadsheet” fails and “when the call ends, log it on your phone — and get your warm-lead list back” works.
Designing a new workplace habit
Fill this in per target behaviour:
Habit design
Behaviour: _______________
Existing cue we can hijack: _______________
(What already reliably happens right before the moment this behaviour should occur? Monday stand-up, end of a call, opening the laptop, ticket closing.)
Old routine it replaces: _______________
Old reward we must still deliver: _______________
New reward, within 24 hours: _______________
Implementation intention: “When _____________, I will _____________.”
⚠️ Attach to an existing cue. New behaviours anchored to new cues almost never stick — you’d be building two habits at once. Find something already automatic and bolt on.
⚠️ The reward must be fast. “This will improve forecasting accuracy next quarter” is not a reward the automatic system can register. “The system now pre-fills your Friday report” is.
Why adoption graphs look like this
Adoption
100% │ ╭─╮
│ ╱ ╲___
│ ╱ ╲______ ← the true baseline
│ ╱ ╲_____
│ ╱
0% └──────────────────────────────► weeks
1 2 3 4 6 8 12
↑ ↑
novelty + attention
supervision leaves
The early spike is not adoption. It is novelty plus surveillance. Everyone knows they’re being watched in week one. The real number is what you see at week 12 when nobody is looking.
Practical consequences:
- Never report week-2 adoption as success. You’ll be held to a number that was always going to fall, and the fall will read as failure.
- Set the drift check dates before go-live — see Drift Check — 30 60 90.
- Plan a deliberate week-8 re-intervention. Most programmes have disbanded by then. That’s precisely when reinforcement is worth most.
The four causes of drift
| Cause | Sign | Counter |
|---|---|---|
| Attention withdrawal | Adoption falls the week the programme team stands down | Hand to the line with a named owner and a live metric before you leave |
| Effort accumulation | “It’s fine normally but not when we’re busy” | The new way must be easier than the old on the worst day, not the average day |
| Reward extinction | Early recognition stops, behaviour stops | Move recognition into the standing management rhythm, not the programme |
| Competing change | A new priority lands and displaces it | Sequence changes; show the sponsor the full list of what’s live |
Effort accumulation is the killer. Test your process at peak load, understaffed, on a Friday. If it fails then, it will fail permanently — because the exception becomes the excuse, and the excuse becomes the norm.
Making it structural
The most reliable form of reinforcement isn’t motivational, it’s structural. Get the behaviour into:
- System defaults and mandatory fields — the strongest lever available
- The standing meeting agenda — if it isn’t discussed, it isn’t real
- Onboarding — new joiners should never learn the old way exists
- Objectives and performance conversations
- Whatever the leadership team asks about — organisations optimise for what gets asked
The blunt version for a sponsor: the behaviour will last exactly as long as someone senior keeps asking about it. Design who that is, and put it in their standing agenda.
Links
→ Embedding and Reinforcement → Drift Check — 30 60 90 → Nudge Library
← Back to the Behavioural Change Toolkit