Experimentation isn't just A/B testing, it's six jobs: find where the funnel actually leaks, map the path to activation, ship the page that fixes it, price it right, route the right variant to the right visitor, and run the statistics properly. gtm-skills' Experimentation Lead pack covers all six in free, open source skills grounded in real testing methodology. Links to the skill files and to the Intempt agent that runs the same job with live guardrails and exit criteria.
Experimentation gets reduced to "run an A/B test," when the actual job starts earlier and ends later: find the real leak, not the obvious one. Map what happens between signup and the moment someone actually gets value. Ship the page. Price it right. Route the right experience to the right visitor. Then run the statistics so the result actually means something.
Six distinct jobs, same shape every quarter. That's how this pack is built.
TL;DR
- gtm-skills' Experimentation Lead pack is 6 free skills: funnel leak diagnosis, activation mapping, landing page design, pricing strategy, personalization routing, and Bayesian test design.
- Install once with npx skills add sidchaudhary/gtm-skills.
- Start with the-leak-finder - know where the funnel actually breaks before designing a test around it.
- Intempt's Experimentation Lead agent uses CUPED to reach significance 20-40% faster and promotes winners automatically.
How to install these in about 60 seconds
- Run npx skills add sidchaudhary/gtm-skills.
- Create a Claude project per skill - "Leak Finder," "Hypothesis Engine."
- Paste the skill's SKILL.md into the project instructions.
- Start a chat with your funnel data, activation metric, or pricing question.
The 6 skills
1. The Leak Finder
Diagnoses funnel drop-offs with conversion benchmarks, root cause analysis, and an optimization roadmap - for marketing, sales, or product funnels. This is where every test should start: know the real leak before you design around it. Read the skill.
2. The First Mile Mapper
Designs the post-signup activation flow: what happens between signup and the aha moment, in what order, and how drop-off gets diagnosed and fixed. Use this when users are signing up but not sticking around. Read the skill.
3. The Page Shipper
Designs conversion-optimized landing pages with copy frameworks and prototype-ready HTML/Tailwind output, built for lead capture, signups, and launches. Read the skill.
4. The Price Point Finder
Designs pricing and packaging: value metric selection, tier structure, price points, and price-increase timing. For setting prices, restructuring plans, or deciding whether and how to raise them. Read the skill.
5. The Variant Router
Designs personalization rules mapping audiences to content variants, with a measurement plan attached - for dynamic content across pages, emails, and in-app experiences. Read the skill.
6. The Hypothesis Engine
Designs Bayesian experiments with Thompson sampling, guardrails, holdouts, and exit criteria - for experience optimization and variant testing done properly, not eyeballed. Read the skill.
The order to build them in
Start with the-leak-finder - every test after this one should be answering a question the leak-finder actually raised, not a hunch. Add the-first-mile-mapper if activation is the specific problem, or the-page-shipper if it's a top-of-funnel conversion issue. The-hypothesis-engine comes in once you have a real test to run properly. Save the-price-point-finder and the-variant-router for when pricing or personalization becomes the actual lever, not a default place to look.
What to never hand over
The decision to ship a pricing change - the-price-point-finder gives you the analysis, a human makes the call given context the skill doesn't have, like a renewal happening next week. And exit criteria on a live test: guardrails matter, but the decision to stop a test early because something looks wrong is a judgment call, not an automated rule.
When you outgrow the manual version
Intempt's Experimentation Lead agent runs this same job with CUPED variance reduction, reaching statistical significance 20-40% faster without inflating false positive rates, and promotes a winning variant to a permanent personalization rule in one click. Same six jobs, running with live guardrails instead of a spreadsheet you check manually.
Frequently asked questions. Answered.
Six free Claude Skills covering CRO and experimentation: funnel leak diagnosis, activation flow mapping, landing page design, pricing strategy, personalization routing, and Bayesian test design with guardrails. Install with npx skills add sidchaudhary/gtm-skills.
