Customer Segmentation

Use this skill to run Customer Segmentation as a practical consulting method, not as a generic framework explanation.

Method Notes

• Segment for a decision, not for taxonomy beauty.
• Use needs, behavior, value, buying context, and service cost.

Required Inputs

Collect or infer these inputs before execution:

• customer universe
• needs or behaviors
• value data
• usage or buying evidence

If an input is missing, do not block automatically. Mark it as missing, state the assumption used, and add a validation action.

When Not To Use

Do not use without a defined customer population and evidence of differences that change needs, economics, behavior, or treatment. Avoid segments based on convenient demographics that do not change a decision.

Adjacent Methods

• jobs-to-be-done: understand progress and switching forces within a segment.
• empathy-map: build an evidence-based snapshot of one segment or actor.

Step-by-Step Execution

Output Template

### 1. Segmentation Purpose
Decision to support:
Population:
Available evidence:
Excluded scope:

### 2. Segment Definitions
| Segment | Defining need / behavior | Size / value evidence | Service need | Confidence |
|---|---|---|---|---|
|  |  |  |  |  |

### 3. Segment Profiles
| Segment | Trigger | Desired outcome | Barrier | Best offer / channel |
|---|---|---|---|---|
|  |  |  |  |  |

### 4. Action And Validation
| Segment | Action | Owner | Signal | Re-segmentation trigger |
|---|---|---|---|---|
|  |  |  |  |  |

### Evidence And Next Decision
- Confirmed facts:
- Assumptions:
- Missing evidence:
- Next action, owner, and timing:
- Expected signal and decision threshold:

Quality Gate

• Produce the method-specific outputs for Segmentation Purpose, Segment Definitions, Segment Profiles; do not substitute a generic framework summary.
• Tie every material score, causal claim, or stakeholder judgment to evidence or label it as an assumption.
• Include at least one disconfirming check, sensitivity, alternative explanation, or failure condition appropriate to the method.
• Convert the result into a named action or decision with an owner or stakeholder, timing, and observable signal.
• State missing inputs and the smallest validation action instead of inventing precision.
• Keep wording professional and plain enough that a smart non-specialist can use the output directly.