funnel-diagnostics

name	funnel-diagnostics
description	Funnel diagnostics engine — bottleneck detection, cross-data correlation, AARRR benchmarks by vertical

Funnel Diagnostics Engine

Diagnose conversion bottlenecks by cross-referencing data from PostHog, Stripe, Meta Ads, Google Ads, and email platforms. This skill provides the inference rules that power automated funnel analysis.

The AARRR Framework

Acquisition → Activation → Revenue → Retention → Referral

Each stage has:
  - A defining event (what counts as completing this stage)
  - A conversion rate to the next stage
  - Benchmarks by vertical
  - Diagnostic patterns when rates are below benchmark

Stage Definitions

Stage	Defining Event	What It Measures
Acquisition	User visits site (unique visitor)	Can you attract people?
Activation	User completes key action (signup + first value moment)	Do they "get it"?
Revenue	User pays (first payment)	Will they pay?
Retention	User returns (active in week 2+)	Do they stay?
Referral	User invites others (share, invite, review)	Do they spread?

Diagnostic Process

Step 1: Map the Funnel

Identify the specific events for each AARRR stage in the target product:

Example (Coding Bootcamp):
  Acquisition: $pageview on /landing (UTM-tagged)
  Activation: onboarding_completed
  Revenue: payment_completed (Stripe checkout.session.completed)
  Retention: course_lesson_completed in week 2+
  Referral: referral_link_shared or invite_sent

Step 2: Pull Conversion Rates

For each stage transition, calculate the conversion rate over the last 30 days:

Acquisition → Activation:  visitors → activated / visitors
Activation → Revenue:      activated → paid / activated
Revenue → Retention:       paid → retained_week4 / paid
Retention → Referral:      retained → referred / retained

Step 3: Compare to Benchmarks

See the AARRR benchmarks rule file for industry-specific benchmarks.

Step 4: Identify the Bottleneck

The stage with the largest gap between actual and benchmark performance is the bottleneck. Focus all optimization effort here.

Step 5: Apply Diagnostic Patterns

Each bottleneck has specific diagnostic patterns. See the bottleneck-patterns rule file.

Step 6: Separate Demand Blockers from Plumbing Leaks

A stage can leak because the OFFER has a demand blocker, not because the page is slow. Map the SDV B-tier demand blockers (B0/B1/B2) onto the AARRR stages and emit a ranked revenue-at-risk ledger. See the revenue-at-risk rule file. (F = fidelity tier of the forecast, B = blocker tier — separate from the AARRR stages.)

Data Sources for Diagnosis

Data Source	What It Provides	Key Signals
PostHog	User behavior, funnels, session recordings	Drop-off points, rage clicks, dead clicks
Stripe	Revenue, subscriptions, churn	MRR, trial conversion, payment failures
Meta Ads	Ad performance, audience quality	CPA, CTR, CPM, frequency
Google Ads	Search intent, keyword performance	Quality Score, impression share, CPC
Email Provider	Engagement, sequence performance	Open rate, click rate, unsubscribe rate
Google Search Console	Organic visibility	CTR, position, impressions

Cross-Data Correlation

The Correlation Matrix

When diagnosing a bottleneck, correlate data across sources to identify root cause:

High CPA + Low CVR = Landing page problem (traffic quality is fine, page doesn't convert)
High CPA + High CVR = Targeting problem (converting users are expensive to reach)
Low CPA + Low CVR = Misaligned offer (cheap traffic but wrong audience)
Low CPA + High CVR = Working well (scale this)

High Open Rate + Low Click Rate = Email content problem (subject line works, body doesn't)
Low Open Rate + Any Click Rate = Deliverability or subject line problem
High Click Rate + Low CVR = Landing page problem (email works, page doesn't)

Severity Classification

Severity	Criteria	Action
Critical	Conversion rate <50% of benchmark AND >$1000/month at stake	Fix immediately, pause spend if needed
High	Conversion rate 50-75% of benchmark	Fix this week
Medium	Conversion rate 75-90% of benchmark	Fix this month
Low	Conversion rate 90-100% of benchmark	Optimize when bandwidth allows
None	At or above benchmark	Monitor, don't touch

Reporting Template

Weekly Funnel Health Report

FUNNEL HEALTH: [DATE RANGE]

Stage          | Actual | Benchmark | Status    | Trend
Acquisition    | 2,400  | --        | --        | +12% WoW
Acq → Act      | 18%    | 20-30%    | Medium    | Flat
Act → Revenue  | 3.2%   | 5-10%     | Critical  | -0.8% WoW
Rev → Retain   | 72%    | 70-80%    | Healthy   | +2% WoW
Retain → Refer | 4%     | 5-15%     | Medium    | Flat

BOTTLENECK: Activation → Revenue
ROOT CAUSE: [diagnosis from patterns]
PRESCRIPTION: [from prescription rules]