Builds conversion funnels from raw event data with drop-off attribution, segment comparison, and significance testing, and delivers a filled funnel table with a diagnosed bottleneck and next action. Use when someone asks "where are users dropping off", "why did checkout conversion fall this week", "build a signup-to-activation funnel", or "is mobile converting worse than desktop". Do NOT use for rewriting page copy or layout to lift a known weak step - use landing-page-cro instead; for decomposing active-user growth into new, retained, resurrected, and churned - use growth-accounting instead; for choosing which events to instrument and which product metrics to track - use product-analytics instead; for designing and reading a controlled experiment - use ab-test-analyzer instead.
Click to play with sound.
---
name: Funnel Analysis
description: Builds conversion funnels from raw event data with drop-off attribution, segment comparison, and significance testing, and delivers a filled funnel table with a diagnosed bottleneck and next action. Use when someone asks "where are users dropping off", "why did checkout conversion fall this week", "build a signup-to-activation funnel", or "is mobile converting worse than desktop". Do NOT use for rewriting page copy or layout to lift a known weak step - use landing-page-cro instead; for decomposing active-user growth into new, retained, resurrected, and churned - use growth-accounting instead; for choosing which events to instrument and which product metrics to track - use product-analytics instead; for designing and reading a controlled experiment - use ab-test-analyzer instead.
---
# Funnel Analysis
A funnel tells you where money leaks out of a multi-step flow, but only if the definition is precise and the differences you report are statistically real. The costly mistake this skill prevents is shipping a "mobile checkout is broken" diagnosis built on an ambiguous funnel definition and a 40-user sample - teams reorganize roadmaps around noise. The output is a filled funnel table, one named bottleneck step, and the evidence that the gap is real.
## Operating procedure
Follow the steps in order: the funnel definition (Step 1) determines every count downstream, and significance testing (Step 5) must come before any diagnosis leaves the room.
### Step 1: Gather inputs and pin the definition
Collect these before touching SQL. If the user cannot answer, apply the default and label it a guess.
1. Ordered steps, as event names (e.g. view -> add_to_cart -> checkout -> purchase). Write them down verbatim; ambiguous definitions produce misleading funnels.
2. Conversion window: how long after step 1 a user has to finish. Default: the 90th percentile of observed time-to-convert, rounded up to a clean unit. Typical defaults if no data yet: 1 day for e-commerce checkout, 7 days for signup-to-activation, 14-30 days for B2B trial-to-paid.
3. Ordered vs unordered: must steps happen in sequence, or just all occur within the window? Default: ordered.
4. Counting unit: unique users (default) or sessions. Never mix.
5. Segments to compare: device, channel, plan, acquisition cohort. Cap at the 2-3 the team can act on.
### Step 2: Compute the funnel in SQL