funnel-analysis

name	funnel-analysis
description	Analyse conversion through a multi-step process, identify where users drop off, and diagnose why. Activate when conversion is low, when comparing conversion across segments, or when optimising a user journey.

When to use

• Conversion is low and the team needs to know where users are dropping off
• Comparing conversion rates across channels, devices, or user segments
• A product change was made and you need to measure its effect on funnel completion
• Building a business case for investing in a specific funnel step

Process

1. Define funnel steps and timeframe — list ordered steps (e.g. Visit → Signup → Activate → Purchase); define completion window (e.g. must complete within 7 days). Ambiguous definitions here will invalidate the analysis.
2. Build user-level dataset — extract first occurrence of each event per user within the completion window using scripts/funnel_analyzer.py
3. Calculate conversion metrics — compute absolute counts, step-to-step conversion rates, and overall funnel conversion rate
4. Analyse time-to-convert — calculate median and distribution of time between steps; identify whether slow converters have lower overall completion
5. Segment the funnel — break conversion by available dimensions (channel, device, geography, user segment); identify highest-impact differences
6. Prioritise recommendations — rank drop-off points by (users lost × revenue impact per user); fill assets/funnel_report_template.md

Inputs the skill needs

• Required: event log data with user ID, event name, and timestamp
• Required: ordered list of funnel steps
• Required: completion window (e.g. 30 days from first step)
• Optional: segmentation dimensions
• Optional: estimated revenue value of a completed conversion

Output

• scripts/funnel_analyzer.py — computes step-by-step conversions and drop-offs from event logs
• references/funnel_design_guide.md — funnel definition patterns and common measurement pitfalls
• assets/funnel_report_template.md (filled) — funnel table, segment comparison, and ranked recommendations