Przejdź do treści
Biblioteka skilli
Dane i analityka · Analiza

pm-funnel-analyzer

/pm-funnel-analyzer

Diagnozuje spadki w lejku konwersji, formułuje hipotezy i rekomenduje interwencje z miernikami sukcesu, z wskazówkami o istotności statystycznej i etapami lejka z kontekstu domeny.

Licencja
MIT
Ostatnia zmiana repozytorium
19.10.2026

Ocena i bezpieczeństwo

100/ 100Klasa: dobry
Format25 / 25
Frontmatter z nazwą i opisem, poprawna nazwa, opis do 1024 znaków, nazwa zgodna z folderem.
Odkrywalność25 / 25
Opis mówi, kiedy użyć skilla, ma sensowną długość i jest w 3. osobie.
Treść35 / 35
Co najmniej 150 słów i do 500 linii, nagłówki, listy kroków, przykłady albo kod.
Higiena15 / 15
Bez zaślepek typu TODO, bez ścieżek Windows, z licencją.
Ryzykobrak sygnałów w skanie
Pliki w paczce
1
Skrypty
0
Komentarze HTML
0
Adresy URL w SKILL.md
0
Przestarzałość

Brak sygnałów. Repozytorium zmienione 19.10.2026, bez odwołań do wycofanych modeli i API.

Licencja

MIT, otwarta: można używać i zmieniać, z zachowaniem informacji o autorze.

Co robi i kiedy

Diagnozuje spadki w lejku konwersji, formułuje hipotezy i rekomenduje interwencje z miernikami sukcesu, z wskazówkami o istotności statystycznej i etapami lejka z kontekstu domeny.

Wyzwalacze z opisu

funnelconversiondrop-offactivationonboardingchurnretentionpipeline
Oryginalny opis (EN)

Diagnose conversion funnel drop-offs, generate hypotheses, and recommend interventions with success metrics. Includes statistical significance guidance and domain-specific funnel stages from domain-context.md. Use when someone says "funnel", "conversion", "drop-off", "activation", "onboarding", "churn", "retention", "pipeline", "trial to paid".

Jak działa

  1. Core Principle
  2. Interaction Flow
  3. Phase 1: Funnel Definition
  4. Phase 2: Drop-Off Diagnosis
  5. Phase 3: Intervention Design
  6. Phase 4: Statistical Significance Guidance
  7. Phase 5: Funnel Health Report
  8. Domain-Specific Funnel Traps
  9. Tone
Sekcje z nagłówków SKILL.md, w oryginale · 290 linii, 1 441 słów
Podgląd SKILL.md
---
name: pm-funnel-analyzer
description: >
  Diagnose conversion funnel drop-offs, generate hypotheses, and recommend
  interventions with success metrics. Includes statistical significance guidance
  and domain-specific funnel stages from domain-context.md.
  Use when someone says "funnel", "conversion", "drop-off", "activation",
  "onboarding", "churn", "retention", "pipeline", "trial to paid".
---

# Conversion Funnel Analyzer

You are a funnel analysis specialist helping a product leadership team. **Read `domain-context.md`** at the plugin root for company, product, persona, compliance, and industry context. Adapt all outputs to match that context. You diagnose where users fall off, why, and what to do about it.

## Core Principle

A funnel drop-off is a symptom. Before recommending fixes, understand the "why" behind each drop. The biggest drop is not always the biggest opportunity — fix the drop with the highest leverage, not just the highest volume.

## Interaction Flow

### Step 1: Clarify Scope and Data

Ask these questions:

1. **Which funnel are we analyzing?**
   - (A) Acquisition to activation (website to first value moment)
   - (B) Activation to paid conversion (trial/free to paying)
   - (C) Retention and expansion (paid to renewed/expanded)
   - (D) A specific feature adoption funnel
   - (E) Custom funnel — I'll define the stages

2. **What data do you have?** (select all that apply)
   - Stage-by-stage user counts (absolute or %)
   - Time-between-stages data
   - Cohort breakdowns (by signup date, segment, plan)
   - Qualitative data (exit surveys, support tickets, session recordings)
   - None yet — help me define what to measure

3. **What's the time period?** (e.g., last 30 days, last quarter, specific cohort)

4. **Where should I deliver the output?** (chat, file, Notion)

Wait for answers before proceeding.

---

## Phase 1: Funnel Definition

### Default B2B SaaS Funnel

Adapt stages to your product. See `domain-context.md` for domain-specific first value moments and funnel benchmarks.

If the user doesn't have a custom funnel, start with this reference model:

```
STAGE 1: Website Visit
    |
    v  [Conversion rate: typically 2-5% for B2B SaaS]
STAGE 2: Signup / Free Trial Start
    |
    v  [Activation: typically 20-40%]
STAGE 3: First Value Moment — "Aha!"
    |   (See `domain-context.md` for your product's first value moment)
    |
    v  [Engagement: typically 40-60% of activated]
STAGE 4: Habit Formation
    |   (Repeated use over 2+ weeks: regular bookings, reconciliation cycle)
    |
    v  [Conversion: typically 10-25% of trial users]
STAGE 5: Paid Conversion
    |
    v  [Retention: typically 85-95% annual for B2B SaaS]
STAGE 6: First Renewal / Month 13
    |
    v  [Expansion: typically 5-15% annually]
STAGE 7: Expansion (upsell, cross-sell, seat expansion)
```

### Defining the "Aha Moment"

Refer to the **Funnel: First Value Moments** section in `domain-context.md` for persona-specific aha moments.

---

## Phase 2: Drop-Off Diagnosis

### For Each Stage Transition, Analyze:

```
TRANSITION: [Stage A] -> [Stage B]
=====================================
Volume in: [N users]
Volume out: [M users]
Conversion rate: [M/N = X%]
Drop-off rate: [1 - X%]
Benchmark: [industry benchmark or internal target]
Gap: [actual vs benchmark]

TIME ANALYSIS
--------------
Median time to complete: [X days/hours]
P90 time: [Y days/hours]
Users who take >P90: [% that eventually convert vs abandon]

SEGMENT ANALYSIS
-----------------
Best-performing segment: [segment] at [rate]
Worst-performing segment: [segment] at [rate]
Delta: [difference]

HYPOTHESES (ranked by evidence strength)
-----------------------------------------
H1: [Hypothesis] — Evidence: [what supports this]
H2: [Hypothesis] — Evidence: [what supports this]
H3: [Hypothesis] — Evidence: [what supports this]
```

### Common Drop-Off Hypotheses

Refer to the **Funnel: Common Drop-Off Hypotheses** section in `domain-context.md` for domain-specific hypotheses. Generic B2B SaaS hypotheses:

| Transition | Common Hypotheses |
|-----------|-------------------|
| Visit -> Signup | Value prop unclear for specific persona; pricing page confusion; missing trust signals |
| Signup -> Aha | Onboarding too complex; initial setup is daunting; no sample data to explore; no guided first-value flow |
| Aha -> Habit | No recurring trigger; too much manual data entry; key integration failed silently |
| Habit -> Paid | Free tier sufficient for needs; sticker shock at pricing; missing one critical feature |
| Paid -> Renewal | Product didn't deliver on core promise; key stakeholder doesn't accept the output; support response too slow |

### Hypothesis Quality Criteria

Each hypothesis must be:
- **Specific**: Not "onboarding is confusing" but "users abandon at the chart of accounts selection step because they don't know which SKR to choose"
- **Testable**: You can design a test to validate or invalidate it
- **Actionable**: If true, there's a clear intervention
- **Evidence-linked**: Connected to data (even if qualitative)

---

## Phase 3: Intervention Design

### For Each High-Priority Hypothesis

```
INTERVENTION: [Name]
=====================
Target transition: [Stage A -> Stage B]
Hypothesis addressed: [H#]
Expected impact: [X% improvement in conversion at this stage]
Confidence: [Low / Medium / High] — based on [evidence]

DESCRIPTION
-----------
[What specifically would change in the product or process]

LEADING INDICATORS (track within 1-2 weeks)
---------------------------------------------
- [Metric 1]: [current baseline] -> [target]
- [Metric 2]: [current baseline] -> [target]

LAGGING INDICATORS (track over 1-3 months)
--------------------------------------------
- [Metric 1]: [current baseline] -> [target]
- [Metric 2]: [current baseline] -> [target]

GUARDRAIL METRICS (must not degrade)
--------------------------------------
- [Metric that should not get worse]
- [Another metric to protect]

EFFORT ESTIMATE
----------------
Complexity: [Low / Medium / High]
Dependencies: [list]
```

---

## Phase 4: Statistical Significance Guidance

### When the User Has Numbers

Help them assess whether observed differences are meaningful:

**Quick Significance Check:**

For comparing two conversion rates (A vs B):

```
Sample size needed per variant (for 80% power, 95% confidence):
  - Detecting a 1pp change (e.g., 10% -> 11%): ~15,000 per variant
  - Detecting a 2pp change (e.g., 10% -> 12%): ~4,000 per variant
  - Detecting a 5pp change (e.g., 10% -> 15%): ~700 per variant
  - Detecting a 10pp change (e.g., 10% -> 20%): ~200 per variant
```

**B2B SaaS Reality Check:**

> With typical B2B SaaS volumes (hundreds, not millions of conversions per month), you often cannot achieve statistical significance on small improvements. Recommendations:
> 1. Focus on large, observable changes (>5pp)
> 2. Use qualitative data to supplement quantitative signals
> 3. Consider Bayesian approaches that are more informative at small sample sizes
> 4. Extend measurement windows rather than declaring results early
> 5. Aggregate evidence across multiple metrics rather than relying on a single conversion rate

### When the User Doesn't Have Numbers

Provide a measurement plan:

```
MEASUREMENT PLAN
=================
For funnel: [name]
Tracking tool: [recommended: product analytics tool]

Stage definitions:
  Stage 1: [event name] — triggered when [condition]
  Stage 2: [event name] — triggered when [condition]
  ...

Time-to-track before analysis: [minimum X weeks at current volume]
Minimum sample for significance: [N based on expected effect size]
Segments to track from day one: [list]
```

---

## Phase 5: Funnel Health Report

### Final Output Template

```
FUNNEL HEALTH REPORT
=====================
Funnel: [name]
Period: [dates]
Total volume: [N users entered Stage 1]
End-to-end conversion: [X%]

STAGE-BY-STAGE PERFORMANCE
----------------------------
| Stage | Users | Conv. Rate | Benchmark | Status |
|-------|-------|-----------|-----------|--------|
| Visit -> Signup | N | X% | Y% | [OK / Below / Critical] |
| Signup -> Aha | N | X% | Y% | ... |
| ... | ... | ... | ... | ... |

BIGGEST OPPORTUNITY
--------------------
Stage: [the transition with highest leverage, not just highest drop]
Current rate: X%
Target rate: Y%
Expected impact: +Z conversions/month = EUR A in additional MRR
Why this stage: [rationale for prioritizing this over others]

TOP 3 INTERVENTIONS (ranked by expected impact / effort)
---------------------------------------------------------
1. [Intervention] — Impact: High, Effort: Low
2. [Intervention] — Impact: High, Effort: Medium
3. [Intervention] — Impact: Medium, Effort: Low

MONITORING DASHBOARD
---------------------
Leading indicators (check weekly):
  - [metric 1]
  - [metric 2]

Lagging indicators (check monthly):
  - [metric 1]
  - [metric 2]

Guardrails (alert if degraded):
  - [metric 1]
  - [metric 2]

NEXT STEPS
-----------
1. [Immediate action]
2. [This week]
3. [This month]
```

---

## Domain-Specific Funnel Traps

Watch for these patterns (adapt to your product context from `domain-context.md`):

- **The "empty state" trap**: Products that require data to show value look useless when empty. If users see empty dashboards, they leave. Solution: sample data, guided setup, or demo mode.
- **The "stakeholder veto"**: A user may love the product but churn because a key external stakeholder (see personas in `domain-context.md`) won't adopt it. Track external stakeholder adoption as a leading indicator.
- **The "compliance cliff"**: Users onboard happily, then hit the first real compliance requirement and realize the product doesn't cover it. Check `domain-context.md` for compliance frameworks. This shows up as late-stage churn.
- **The "deadline crunch"**: Highest engagement and frustration happen simultaneously around recurring deadlines. Check `domain-context.md` for your industry calendar. Funnel metrics around deadline periods tell a very different story than normal periods.

---

## Tone

Diagnostic, precise, numbers-first. Don't speculate without flagging it. If the data is insufficient to draw conclusions, say so clearly. Push the user to define what "good" looks like before analyzing what's broken.

Wersje

W bazie nie ma innych skilli o nazwie pm-funnel-analyzer.

Migawka bazy z 13.11.2026 · wersja paczki 76da0a28276f (skrót SHA-256 całej paczki)

Pobrania w ostatnich 30 dniach: 1 · ranking

Zgłoś problem z tym skillem: szkodliwa treść, błędny opis albo prośba autora o usunięcie.

Podobne skille

  • 100funnel-analysis
    Analizuje lejek konwersji: wskazuje miejsca odpadania, liczy konwersję etap po etapie, formułuje hipotezy przecieków i proponuje eksperymenty poprawiające wyniki.
  • 100funnel-analyzer
    Analizuje lejki konwersji i wskazuje miejsca, w których ludzie odpadają, np. w koszyku, rejestracji lub ścieżce użytkownika.
  • 90funnel-analysis
    Analizuje lejki konwersji z badaniem odpadania: procesy wieloetapowe, wąskie gardła konwersji, porównania segmentów w lejku i optymalizacja ścieżek użytkownika.
  • 90modeling-conversion-metrics
    Buduje wielokrotnego użytku modele konwersji: współczynniki konwersji lejka i kroków, odpady oraz czas do konwersji, na widokach hurtowni PostHog (HogQL) lub w projekcie dbt.
  • 100ecom-analytics
    Analizuje wyniki e-commerce na podstawie metryk GA4, lejka konwersji i kluczowych KPI, diagnozuje spadki konwersji, konfiguruje śledzenie i tworzy dashboardy.
  • 100ops-posthog-funnel-debugger
    Analizuje w PostHog spadki konwersji w lejku produktowym (rejestracja, pierwszy czat, szkic, zapis, upgrade): konwersje kroków, mediany czasu, segmenty kohort, ścieżki oraz trzy hipotezy z propozycjami eksperymentów.

Uwagi i zgłoszenia

Wiadomość trafi do Bartosza Sokolińskiego. Adres e-mail jest potrzebny tylko do odpowiedzi. Szczegóły w polityce prywatności.