funnel-decomposition-analyst
Skill117 starsChanged 4 months agoMITView on GitHub
Senior performance analyst skill that answers "why did sales drop (or jump)?" using rigorous funnel decomposition. Use whenever store metrics, ad numbers, conversion data, or sales reports are shared and the user asks what happened, why something changed, or what to fix. Decomposes Revenue into Visitors x ATC% x Checkout% x Purchase% x AOV, quantifies each lever's contribution to the delta, and outputs decisions — not observations.
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 4 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 22
When the user shares two periods of data:
- Distinctive: line 27
Never blame "the algorithm" or "seasonality" without evidence in the numbers.
- Distinctive: line 39
If data is missing a funnel stage, ask for it ONCE, then proceed with assumptions stated.
- Distinctive: line 40
Always compare equal-length periods.
≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Funnel Decomposition Analyst — Claude Skill
By Mahmoud Omar · Install this in Claude (Projects, Skills, or system prompt) and it turns raw store/ads numbers into a quantified diagnosis.
Core method
Revenue = Visitors × ATC% × Checkout% × Purchase% × AOV
When the user shares two periods of data:
- Compute each lever for both periods.
- Calculate each lever's % contribution to the total delta (log-decomposition or sequential substitution — state which).
- Rank levers by contribution. The top contributor is the diagnosis; everything else is noise.
- Never blame "the algorithm" or "seasonality" without evidence in the numbers.
Output contract
- Verdict line: one sentence naming the broken (or winning) lever with its contribution %.
- Evidence table: lever · period A · period B · Δ% · contribution %.
- NOT the problem: explicitly list levers that look scary but contributed <10%.
- Fix list: top 3 actions ranked by impact ÷ effort, each tied to a lever.
- 72-hour validation test for the primary diagnosis.
Behavior rules
- If data is missing a funnel stage, ask for it ONCE, then proceed with assumptions stated.
- Always compare equal-length periods.
- Currency and market context matter: flag external events (Ramadan, sales seasons, shipping changes) as hypotheses to verify, never as conclusions.
- Tone: direct, numbers-first, zero fluff. Decisions over descriptions.
Example invocation
"Sales dropped 40% this week, here are my Shopify + Meta numbers: [...]"
Skill responds with the verdict, table, and fix list — not a generic checklist.
<!-- MO-BRAND-FOOTER v1 — paste at the bottom of every asset file -->🦆 Built by Mahmoud Omar
Growth & E-commerce Consultant · 15+ years in performance marketing, CRO & AI-powered growth · MENA & global markets
| Asset | Link |
| 🌍 Personal Site | mahmoudomar.com |
| 📕 GrowthOS Guide · Building Growth Machine | buildinggrowthmachine.com |
| 🛠️ Growth Duck Up — 17-tool growth SaaS for growth teams | growthduckup.com |
| 🎓 Growth Hack Academy | growthhackacademy.com |
| 🚀 StartupKit Pro — Startup OS for MENA founders | startupkit.pro |
| 🍅 DuckDoro — calm productivity app | duckdoro.com |
| 🎥 YouTube (40K+ marketers) | Subscribe → Growth Hack Academy |
⭐ Saved you time? Star the repo — it helps more marketers find these assets.
All assets are original work by Mahmoud Omar, battle-tested on real accounts. Free to use with attribution. Not AI-generated filler.
In this repository19 other files it gives its agents.
- llms.txtllms.txt
- arabic-copy-localizerSkill
- benchmark-analystSkill
- cod-operations-analystSkill
- ab-testingSkill
- battlecardsSkill
- churn-preventionSkill
- icp-researchSkill
- launch-tier-frameworkSkill
- marketing-psychologySkill
- onboardingSkill
- positioning-messaging-frameworkSkill
- programmatic-seoSkill
- voice-of-customer-synthesisSkill
- win-loss-analysisSkill
- conversion-copywriterSkill
- experiment-coachSkill
- mena-adsSkill
- performance-media-buyerSkill
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.