funnel-analysis
Skill0 starsChanged 7 months agoMITView on GitHub
Use to diagnose drop-off in an acquisition or activation funnel. Produces step-by-step conversion rates, identifies the single biggest leak, hypothesizes causes, and recommends one experiment to run. Always declares assumptions and data source.
10 other projects use a version of this.
Adopted by
Most recently taken up by
- undark-ai/Meta-Ads-OStheir own version, first seen Nov 15, 2026
- shisuidata/ss-data-skillstheir own version, first seen Nov 8, 2026
- RBraga01/builder-growththeir own version, first seen Nov 1, 2026
- astDeniss/business-skillstheir own version, first seen Nov 1, 2026
- sp1e/claude-skillstheir own version, first seen Oct 12, 2026
- Zafer-Liu/Data-Analysis-Agenttheir own version, first seen Oct 8, 2026
Used with
Projects with it also use ab-test-analysis.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
About-IntelligenceSkillab-test-analysis
Analyze A/B test results with statistical significance, sample size validation, confidence intervals, and ship/extend/stop recommendations.
Also in 3 projects that use this
Other versions10 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 10
- Zafer-Liu/Data-Analysis-Agent★ 2.6ksame name
- liangdabiao/claude-data-analysis-ultra-main★ 282same name
- astDeniss/business-skills★ 8same name
- shisuidata/ss-data-skills★ 8same name
- liangdabiao/claudesdk-data-chat★ 4same name
- RBraga01/builder-growth★ 2same name
- nandanosql/all_in_one_pm_skills★ 1same name
- sp1e/claude-skills★ 1same name
- chaakresh/claude-skills★ 0same name
- undark-ai/Meta-Ads-OS★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository8 other files it gives its agents.
- code-reviewSkill
- cold-email-sequenceSkill
- content-calendarSkill
- market-researchSkill
- qualify-leadSkill
- release-notesSkill
- seo-outlineSkill
- write-linkedin-postSkill
The file
We split this skill into the rules it gives an agent
4 rules. 0 are common, 4 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 6 edited versions we read keep these.
4
Distinctive
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 62
Always declare the funnel definition explicitly.
- Distinctive: line 63
Never average step rates — multiply them for end-to-end conversion.
- Distinctive: line 64
Don't recommend an experiment without a sample-size estimate.
- Distinctive: line 65
If a step has < 200 users in the window, flag the rate as "low confidence."
“Common” is measured: kept by at least 6 in 10 of the 6 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
funnel-analysis
The point of a funnel is not to celebrate the top number — it's to find the single biggest leak and fix it.
Inputs to gather
- The funnel definition — ordered list of events (e.g. visit → signup → activate → paid).
- Window — the date range (typically last 30 days).
- Segment — all users? a cohort? a channel?
- Data source — PostHog, Mixpanel, Amplitude, GA4, custom SQL.
Method
- Pull conversion at each step.
- Compute step-to-step rates and the overall rate.
- Identify the step with the lowest conversion rate, not the lowest absolute count.
- Compare that step's rate to industry benchmark (if known) and to last period.
- Hypothesize 3 causes. Rank them by likelihood × ease-to-test.
- Recommend one experiment.
Output format
## Funnel
| Step | Users | Step-to-step | Cumulative |
|---|---|---|---|
| Visit | 12,400 | 100% | 100% |
| Signup | 1,488 | 12.0% | 12.0% |
| Activate | 446 | 30.0% | 3.6% |
| Paid | 67 | 15.0% | 0.54% |
Window: <date range>
Segment: <segment>
Source: <data source>
## Biggest leak
**Activate → Paid (15.0%)** is the lowest step rate. Industry benchmark for B2B SaaS at this stage is 20-30%. We are 5-15 points below.
## Hypotheses
1. **Pricing-page friction** (high likelihood, easy to test) — users land on pricing without seeing ROI proof.
2. **Onboarding incomplete** (medium, medium) — users hit the paywall before reaching their "aha moment."
3. **Wrong ICP traffic** (medium, hard) — top-of-funnel is bringing tire-kickers.
## Recommended experiment
A/B test: insert a one-screen ROI calculator between activate and pricing.
Hypothesis: lifts Activate → Paid from 15% → 20%.
Sample size needed: ~3,500 users per variant for 95% confidence at 5pp lift.
Expected runtime: 4 weeks at current traffic.
## Caveats
- "Activate" is currently defined as <event>; some teams define it differently.
- Last week of data may be incomplete due to <reason>.Hard rules
- Always declare the funnel definition explicitly. "Activated" means nothing without the event.
- Never average step rates — multiply them for end-to-end conversion.
- Don't recommend an experiment without a sample-size estimate.
- If a step has < 200 users in the window, flag the rate as "low confidence."
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.

