A Claude skill for zero-to-one product discovery. Gets you to a paying user faster than any framework that starts with ideation.
The thesis: Users who are already solving a problem badly — with spreadsheets, DMs, copy-paste, or hired help — are proof that the problem is urgent. Find them. Do the thing manually. Charge them. Then build.
Installs a structured product discovery workflow into Claude that:
- Identifies observable workarounds (not stated preferences)
- Classifies workaround type by willingness-to-pay signal
- Tests for independent parallel invention (pattern vs. anecdote)
- Forces a concierge transaction before any code is written
- Outputs an exact DM script to send to your first customer today
Verdict system:
| Verdict | Meaning |
|---|---|
CHARGE NOW (MANUALLY) |
Paid workaround found — send the DM, charge before building |
INTERVIEW AND CHARGE |
Workaround found — talk to 3 deviants this week, attempt charge |
WATCH AND WAIT |
Pattern emerging, not enough cases yet |
DISCARD |
No observable workaround — move on |
- Download
finding-first-customer.skill - Drop it into your Claude skills directory:
- Claude Code:
~/.claude/skills/ - Claude.ai: Settings → Skills → Upload
- Claude Code:
- Restart Claude
Once installed, trigger it by asking Claude anything like:
- "Help me find my first customer for [idea]"
- "Is this worth building?"
- "What should I build in [market]?"
- "Validate this idea: [description]"
- "How do I get my first paying user?"
Claude will run the full positive deviance workflow and output a deviance table, first customer profile, workaround cost calculation, and a verdict with an exact next action.
- TAM estimates
- Competitive landscape analysis
- Roadmap planning
- GTM beyond first 10 customers
This skill terminates at first paying user. Use a different framework for scale.
Built on positive deviance theory applied to product discovery:
- Workarounds are behavioral proof of urgency — stronger than any survey
- Paid workarounds are direct evidence of willingness-to-pay
- Independent parallel invention separates markets from feature requests
- The concierge transaction (manual delivery, real charge) is the fastest validation possible
Inspired by: Zappos (sold shoes manually before building the store), Airbnb (photographed apartments themselves), Superhuman (onboarded users 1:1 before scaling).
PRs welcome. Especially interested in:
- Additional worked examples across different domains
- Edge cases where the framework breaks
- Refinements to the concierge gate logic
MIT