# Decision Rule Template

Documenting tasks removes execution work. Documenting **decisions** removes the waiting,
which is usually the larger bottleneck.

## Write it from your last ten actual decisions
Not from principle. A rule invented from principle will not match your behaviour, so you
will overrule it — and that teaches everyone the rule is not real.

    Last ten [decision type]:
      Approved ___ . What they had in common:
      Declined ___ . What they had in common:

## The three parts — all required

    DECISION: ______________________________

    Threshold:  [the number or condition that separates cases]
    Default:    [what happens — an ACTION, never "use judgement"]
    Escalate:   [when to come to me anyway]

**No threshold** = a sentiment, not a rule.
**No default** = you have not removed the wait.
**No escalation point** = people will not use it, because they cannot see its edge.

## Worked example

    DECISION: Refunds

    Threshold:  Inside 30 days, OR we failed to deliver what was agreed
    Default:    Approve, note it on the record
    Escalate:   Above £2,000, or where the client is threatening a public complaint

    Outside 30 days with work delivered and accepted: decline, offer a credit.

## Set the decision level explicitly

| Level | Sounds like |
|---|---|
| Ask me | "What should we do about this?" |
| Recommend | "I think X — okay?" |
| Decide, tell me first | "I'm going to do X unless you object" |
| Decide, tell me after | "I did X, here's why" |
| Decide | It's in the rule; no notification needed |

Most decisions with a written rule should sit in the bottom two. Someone bringing you options
is still a queue with you at the end of it.

## Never make a rule for
- Rare decisions (twice a year — the rule will be stale by the second one)
- Expensive and irreversible (large spend, hiring, firing, ending a client relationship)
- Strategy — direction, positioning, what you don't do
- Anything depending on relationship context you cannot write down → use a playbook

## Review quarterly
- **Did the escalations need to come to me?** Approved every one without hesitation? Loosen.
- **Did anything go wrong that the rule allowed?** Tighten, and say why.
- **Never escalated at all?** Either perfect, or nobody is using it. Check which.

## When someone applies it and the outcome is bad
Change the rule and back their decision. Overruling a correct application teaches everyone
the rule is fiction, and undoes the whole exercise.