funnel-analysis

Funnel Analysis

You are analyzing a conversion funnel to identify where users drop off, why they drop off, and what actions would most improve overall conversion.

Framework: AARRR (Pirate Metrics — Acquisition, Activation, Retention, Revenue, Referral), Dave McClure's funnel stages.

Step 1 — Load Context

Read memory/user-profile.md for product stage and business model. Read context/company/analytics-baseline.md for existing conversion benchmarks.

Step 2 — Funnel Data Input

Ask the user to provide the funnel data. Accept in any format:

• Step names and conversion rates at each step
• Raw numbers (absolute users at each step)
• A description of the funnel steps and approximate drop-off

If the user describes the funnel without numbers, help them estimate or identify where to find the data.

Step 3 — Calculate Key Metrics

From the funnel data, calculate:

• Step-by-step conversion rate: Users at step N ÷ users at step N-1
• Overall funnel conversion: Users at the final step ÷ users at the first step
• Absolute drop-off: Users lost at each step (in absolute numbers, not just %)
• Revenue impact of each step: Absolute drop-off × revenue per converted user = value of fixing this step

Step 4 — Identify the Biggest Opportunity

Find the step with the highest combination of:

1. Large absolute drop-off (many users lost here)
2. High recovery value (if conversion improved by 10%, how much total revenue or engagement would that add?)
3. Actionability (is there a known reason for the drop? Is there a plausible fix?)

This is the "leaky bucket" — the one step that, if fixed, would most improve the entire funnel.

Step 5 — Diagnose Each Major Drop

For each step with > 20% drop-off:

What happens at this step? (What does the user have to do?) Why do users drop here? (Likely causes — look for patterns in: UX friction, value perception, trust signals, cognitive load, technical issues) What would you need to know to fix it? (Qualitative insight from interviews? Quantitative data from heatmaps or session recordings? An A/B test?)

Common funnel drop diagnoses:

Drop Location	Likely Cause	Diagnostic
Landing page → signup	Value prop unclear; sign-up friction	Session recording; copy test
Signup → first action	Onboarding too long; time to value too slow	Time to first action analysis; session replay
First action → second action	Not seeing value from first action	Interview: "what did you expect to happen?"
Free → paid	Paywall too early; wrong trigger; poor value communication	Behavioral analysis before upgrade prompt
Paid → renewal	Product not becoming habit; poor ongoing value	Cohort analysis; engagement depth

Step 6 — Prioritization Matrix

For each major drop:

• Volume of impact (absolute users × value per conversion)
• Effort to fix (quick UX fix vs. major engineering)
• Confidence that the fix will work (high evidence vs. hypothesis)

Recommend the top 1–2 improvements to try in the next sprint.

Step 7 — Output

Produce:

• Funnel visualization (text table with step names, conversion rates, absolute drop-off)
• Revenue/engagement impact per step (if baseline data is available)
• Diagnosis for each major drop with most likely cause
• Prioritized fix recommendations
• Recommended experiments (one per top-priority step)