strategyplaybook.aiSqwyz
Sign in
← All 124 tools
Research · Diagnose, Explore

Analysis plan

Decides what to find out, in what order, and what good enough looks like for each answer. Without one a team researches everything to the same shallow depth and runs out of time on the branch that mattered.

The printable canvas and the handoff into Claude Code are part of a paid plan. See what a plan includes. It needs the Claude desktop app on this machine, and your team and the prompt library already installed in that project — we cannot see your disk, so open it there.

Stage
01 Diagnose, 02 Explore
Works at
All levels
Maturity
Scaling → Large organisation
Time
Half a day, and it saves a fortnight

What it is

A short document turning a tree of questions into a costed, sequenced set of analyses. For each one it names the source, the method, the owner, and what the finished output looks like — plus the part that is usually missing, which is the point at which the answer is good enough to stop.

One analysis per question
Each leaf of the tree gets a named piece of work. If two leaves share one analysis, the tree was not mutually exclusive and should be fixed before anyone starts.
Source and method
Where the evidence comes from and how it is gathered. "Research the market" is not a method; "twelve customer interviews against this guide" is.
The stopping rule
What answer would be enough. Most analysis runs until the time runs out rather than until the question is settled, which is why the last branch never gets done.
Sequence
Cheap and fast before slow and expensive, and anything that could kill the hypothesis before anything that would merely refine it.
The output shape
What the finished analysis looks like — a number, an exhibit, a ranked list. Naming it in advance is what stops the work drifting into something interesting and unusable.
The mistake people makePlanning the analysis the team enjoys rather than the one the decision needs. The tell is a plan where the expensive, interesting work sits on a branch nobody marked as load-bearing.
What it’s forThe issue tree exists and somebody is about to start gathering evidence.
What it’s not forThe question is still moving. Plan analysis against a settled question or you will plan it twice.

How you run it

  1. Start from the load-bearing branchesTwo or three branches decide the answer. Plan those properly and sample the rest, rather than spreading effort evenly and finishing nothing.
  2. Name a method, not a topicThe source, the sample and the technique. A line that says "competitor research" produces a fortnight of work nobody can review or repeat.
  3. Write the stopping rule before you startWhat would be enough. Deciding that once the data is in front of you is how a two-day analysis becomes a two-week one.
  4. Sequence for the cheapest killAnything that could disprove the hypothesis goes first. There is no point refining an answer that half a day of desk work would have ruled out.
  5. Put an owner and a date on every lineUnowned analysis is the first thing to slip, and it is reliably the branch that mattered most.

The prompt

Two ways to run it

Run this tool in your own Claude

The short prompt starts your partner against the library on your disk. The long one carries everything with it and needs nothing installed.

🔒 Prompt locked

Everything above is free — what this tool is, what it is for, what it is not for, and how to run it as a workshop. Three tool prompts a month are free with an account; beyond three, and for the printable canvas, it is the paid part.

See plans — from £19 a month

Your playbook

It lands in the earliest stage this tool suits. Move it on the Playbook page.

You’ll need

  • An issue tree with its load-bearing branches marked
  • The initial hypothesis, and what would disconfirm it
  • The real deadline, and who is genuinely available to do the work

You’ll end up with

  • One named analysis per branch, with a method and an owner
  • A stopping rule for each — the point at which the answer is enough
  • A sequence, so the cheap answers arrive before the expensive ones

Runs well with

Acquisition screen and synergy thesisAdjacency mapAI opportunity mapAnsoff matrix