STRATEGY

Decide what not to do

Use when the team is spread across too many activities and you need a defensible stop list.

decide-what-not-to-do.md

Download .md

You are a strategy advisor whose job is to find things to stop. You are biased towards stopping.

Everything we currently do: {{CURRENT_ACTIVITIES}}
Results or evidence for each, where we have it: {{EVIDENCE}}
Strategy for the next year: {{STRATEGY}}
Team size and hours: {{CAPACITY}}

Produce:
1. A table with columns: Activity | Ongoing cost per month in hours and money | Evidence it works | Fit with strategy | Verdict (keep, cut, shrink, pause and revisit).
2. For each cut or shrink, name what we give up and who will object.
3. Identify activities kept only because of sunk cost, habit, or one internal stakeholder, and say which of those three applies.
4. Rank the cuts by hours freed per unit of risk taken.
5. Write the one paragraph explanation for each cut that could be sent to the person who owns it, written plainly and without blame.
6. State what we will do with the freed capacity. If the answer is unclear, say the cut should wait.

Constraints: recommend at least two cuts. If the evidence for an activity is missing rather than negative, say MISSING EVIDENCE and propose a 30 day measurement instead of a cut. No em dashes and no softening language.

Fill in before running

Replace each placeholder with your own detail. The more specific you are, the less the model invents.

{{CURRENT_ACTIVITIES}}
{{EVIDENCE}}
{{STRATEGY}}
{{CAPACITY}}

Getting a better result

Include ongoing maintenance work in the activity list, not just campaigns; that is where the hours hide.
The "who will object" column is what makes the meeting go well, so do not delete it.
If nothing gets reallocated, the cut usually reverses within a quarter - fill in step 6 properly.

Questions about this prompt

When should I use this rather than a normal prioritisation pass?

When capacity is the constraint and everything currently running has a defender. Prioritisation ranks what to start next. This is biased towards stopping: it costs each activity in monthly hours and money, tests fit against the stated strategy, and is required to recommend at least two cuts.

What should the activity list include?

A full list of current activities including ongoing maintenance, not just campaigns, because that is where the hours hide. Whatever evidence exists per activity, the strategy for the next year, and team hours. Missing evidence is handled explicitly, so leaving a gap beats filling it with a guess.

What comes back, and which column saves the meeting?

A keep, cut, shrink or pause verdict per activity, what each cut gives up and who will object, the ones kept only through sunk cost, habit or a single stakeholder, cuts ranked by hours freed per unit of risk, and a plain paragraph for each owner. The who will object column is what makes the meeting survivable.

What is the mistake that makes a cut reverse?

Leaving step 6 vague. If freed capacity is not assigned to something specific, the cut reverses within a quarter and the activity returns with its original owner. The prompt says the cut should wait when that answer is unclear, which is worth honouring rather than banking the saving on paper.