Plan, run, or synthesize customer research: interviews, surveys, call and support transcripts, reviews, public communities, jobs-to-be-done, segments, a customer-language library, and evidence-grounded personas. Use when a positioning, message, conversion, pricing, or retention decision needs what customers actually say and do; use truth-file to record the settled result.
Installs into .claude/skills of the current project.
Are you the author of Customer Evidence?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/mtent-customer-evidence)
---
name: customer-evidence
description: Plan, run, or synthesize customer research: interviews, surveys, call and support transcripts, reviews, public communities, jobs-to-be-done, segments, a customer-language library, and evidence-grounded personas. Use when a positioning, message, conversion, pricing, or retention decision needs what customers actually say and do; use truth-file to record the settled result.
license: MIT
metadata:
version: "2.0.0"
author: MTEnt
---
# Customer Evidence
Research must serve a named decision. Public comments prove that a statement exists; they do not establish population prevalence or private buying intent.
If `.agents/marketing-context.md` exists, read it first; it is the truth file every skill in this hub shares, and its evidence states are constraints. Read [research-modes.md](references/research-modes.md) for source selection, interviews, surveys, synthesis, and the persona and segment contract.
## Operating contract
1. Define the decision, audience, inclusion criteria, current assumptions, evidence needed, and the signal that could change the decision.
2. Choose among existing-material analysis, permitted public-signal research, primary interviews, surveys, or a justified combination.
3. Preserve source, date, context, exact wording, permissions, and counterevidence.
4. Separate observation from interpretation and frequency from importance; report sample and method limits without manufacturing statistical confidence.
5. Produce the requested research plan, guide, coded findings, language library, segment model, or persona.
Do not contact participants, distribute surveys, scrape restricted spaces, or expose sensitive data without authorization.
## Return
Decision and scope, source inventory, method, findings with supporting and conflicting evidence, material segments, customer-language library, implications, limitations, open questions, and the next research step.
