Customer · Diagnose, Choose

Segmentation analysis

Splits a market into groups that behave differently enough to be worth serving differently. The test is not whether the groups exist but whether you would actually do anything differently for each.

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, 03 Choose
Works at
Corporate, Business
Maturity
Scaling → Large organisation
Time
Two to three days with data; a day without

What it is

Dividing a market into groups whose needs and buying behaviour differ enough that serving them differently would pay. The discipline is in the last clause: a segmentation that does not change what you do for whom is a data exercise, not a strategy input.

Basis of segmentation
What actually distinguishes the groups — needs, behaviour, willingness to pay, service requirements. Descriptive attributes like size or sector are usually identifiers, not the basis.
Distinctiveness
Whether the segments genuinely respond differently to price, proposition and channel. If they do not, the split is cosmetic.
Size and value
How many, worth how much, at what cost to serve. Revenue size and profit size frequently rank segments in different orders.
Identifiability
Whether you can recognise which segment a customer belongs to at the moment it matters. An unidentifiable segment cannot be served differently.
Priority
Which segments you will build for, which you will serve as they come, and which you will let go. A segmentation without exclusions has made no choices.
The mistake people makeMost segmentations produce four or five well-named groups and change nothing, because the business continues to offer everyone the same thing. Judge the work by what it caused you to stop doing for whom.
What it’s forYou are trying to serve everyone and doing it averagely.
What it’s not forYou have one clear customer type. Segmentation will invent distinctions that do not pay.

How you run it

  1. Segment on behaviour or need, not on descriptionFirmographics and demographics are how you find a segment, not what defines it. Two businesses of identical size can buy in completely different ways.
  2. Test that the segments behave differentlyIf two groups respond the same way to price, channel and proposition, they are one group with two labels.
  3. Size each one, in revenue and in profitA large segment you serve unprofitably is not an opportunity. Cost to serve varies more between segments than most businesses realise.
  4. Check you can actually identify themA segment you cannot recognise at the point of sale is an analytical curiosity. If you cannot target it, you cannot serve it differently.
  5. Decide who you are not forSegmentation that ends with "we will serve all of them, slightly better" has not been done. The output is a priority order and an exclusion.

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