Use when starting any marketing initiative, entering a new market, refreshing ICP, or when downstream marketing decisions lack customer evidence. Use when messaging isn't resonating, churn patterns are unclear, or positioning feels like guesswork. Foundational skill that feeds all other marketing skills.
Resources
1Install
npx skillscat add amogha-dalvi/marketing-gtm/conducting-customer-research Install via the SkillsCat registry.
Customer Research & Voice of Customer
Overview
Every marketing decision is a bet on what your customer cares about. This skill turns raw customer signal into structured insight that every downstream marketing skill consumes. At $1M-5M ARR, you have 50-500 customers — enough for patterns, few enough to talk to them.
When to Use
- Starting a new marketing function from scratch
- Existing marketing lacks customer evidence
- Entering a new market segment
- ICP needs validation or refresh
- Messaging doesn't resonate (low conversion, poor sales feedback)
- Churn patterns are unclear
- Preparing to build any downstream marketing skill
Don't use when: You have recent (<6 months) validated research and no market changes.
Quick Reference
| Phase | Duration | Output |
|---|---|---|
| Audit existing signal | Day 1 | Customer Signal Audit |
| Design primary research | Day 2-3 | Research Plan + Interview Guide |
| Conduct & synthesize | Week 1-3 | Customer Research Synthesis |
| Build artifacts | Day 1 post-synthesis | Voice Library, JTBD Map, Journey Map |
Core Deliverables
- Customer Voice Library — 50+ categorized quotes by pain, aspiration, objection, value
- JTBD Map — Primary, related, emotional, and social jobs
- Buying Journey Map — Trigger → Research → Evaluation → Decision → Onboarding
- Churn & Expansion Signal Map — Why customers leave/expand, early warnings
Common Mistakes
- Surveying instead of interviewing (surveys give data, interviews give understanding)
- Only talking to happy customers (churned/lost deals have the real insight)
- Letting AI replace interviews (AI synthesizes; humans conduct conversations)
- Treating this as one-time (build continuous cadence: 2-3 interviews/month)
- Skipping language capture (exact customer words > any messaging framework)
Integration
Feeds into: defining-icp-segmentation, building-positioning-messaging, gathering-competitive-intel, running-content-engine, growing-nrr
Refresh: Full cycle every 6 months. Ongoing interviews 2-3/month. Voice library update monthly.
See workflow.md for detailed phase-by-phase execution, interview templates, AI-assisted synthesis workflows, and integration point diagrams.