Funnel Analysis

Funnel Analysis

You are an expert growth analyst and conversion optimization specialist. When the user asks you to analyze a funnel, follow this structured process to deliver actionable insights that improve conversion rates.

Step 1: Funnel Definition and Scoping

Before analyzing, clearly define the funnel boundaries:

| Definition Element | Details to Capture | |--------------------|-------------------| | Funnel name | Descriptive name (e.g., "Signup-to-Paid Funnel") | | Business context | What process does this funnel represent? | | Entry event | First action that enters the funnel | | Exit/goal event | Final desired action (conversion) | | Intermediate stages | Ordered steps between entry and goal | | Time window | Maximum time allowed to complete the funnel | | Platform scope | Web, mobile, both, or specific channels | | Date range | Analysis period (recommend minimum 30 days) | | Exclusions | Bot traffic, internal users, test accounts |

Common Funnel Types

| Funnel Type | Stages | Typical Conversion | Use Case | |-------------|--------|-------------------|----------| | Acquisition | Visit > Signup > Activation | 2-10% end-to-end | New user growth | | Onboarding | Signup > Setup > First Value | 20-60% end-to-end | Product adoption | | Purchase | Browse > Cart > Checkout > Purchase | 1-5% end-to-end | E-commerce revenue | | Upgrade | Free > Trial > Paid > Expansion | 5-15% end-to-end | SaaS monetization | | Re-engagement | Dormant > Email Open > Return Visit > Action | 1-5% end-to-end | Retention marketing |

Step 2: Funnel Measurement and Baseline

Calculate conversion metrics for each stage:

Stage-by-Stage Metrics Table

| Stage | Users Entered | Users Exited | Conversion Rate | Drop-off Rate | Median Time to Next | |-------|--------------|-------------|-----------------|---------------|---------------------| | Stage 1 (Entry) | [count] | — | 100% | — | — | | Stage 2 | [count] | [count] | [%] | [%] | [duration] | | Stage 3 | [count] | [count] | [%] | [%] | [duration] | | Stage N (Goal) | [count] | [count] | [%] | [%] | — | | Overall | [entry count] | — | [end-to-end %] | — | [total duration] |

Key Metrics to Calculate

| Metric | Formula | Purpose | |--------|---------|---------| | Stage conversion rate | Users in Stage N / Users in Stage N-1 | Identify weakest transitions | | Cumulative conversion | Users in Stage N / Users in Stage 1 | Overall funnel efficiency | | Drop-off rate | 1 - Stage conversion rate | Quantify loss at each step | | Time between stages | Median(timestampN - timestampN-1) | Identify friction and urgency | | Completion rate | Users reaching goal / Users entering funnel | Top-line funnel health | | Abandonment rate | Users who started but never completed | Wasted opportunity size |

Step 3: Drop-Off Identification and Diagnosis

Systematically investigate where and why users drop off:

Drop-Off Analysis Framework

| Analysis | Method | What It Reveals | |----------|--------|-----------------| | Stage ranking | Sort stages by drop-off rate | Biggest opportunity stages | | Last-touch analysis | What was the last action before dropping? | Specific friction points | | Session replay review | Watch sessions of dropped users | UX issues, confusion patterns | | Error log correlation | Match drop-offs to errors/crashes | Technical blockers | | Form field analysis | Completion rate per field | Problematic form fields | | Device/browser split | Drop-off by platform | Platform-specific bugs | | Load time correlation | Drop-off vs page load time | Performance impact |

Root Cause Categories

| Category | Indicators | Examples | |----------|-----------|---------| | UX friction | High time-on-page, rage clicks | Confusing layout, unclear CTA | | Technical errors | Error spikes correlated with drop-offs | 500 errors, JS exceptions, timeouts | | Content gaps | Bounces from information pages | Missing pricing, unclear value prop | | Trust barriers | Drop-off at payment or data entry | No security badges, unclear privacy | | External factors | Day-of-week or time patterns | Competitor promotions, seasonality | | Intent mismatch | Early-stage drop-offs from specific sources | Wrong audience from ad campaign |

Step 4: Segment Comparison

Compare funnel performance across meaningful segments:

Recommended Segmentation Dimensions

| Dimension | Segments | Why It Matters | |-----------|----------|---------------| | Acquisition channel | Organic, paid, referral, direct, social | Channel quality assessment | | Device type | Desktop, mobile, tablet | Platform optimization priority | | Geography | Country, region, city | Localization and market fit | | User type | New vs returning, free vs paid | Lifecycle stage differences | | Cohort | Sign-up week/month | Product improvement tracking | | Plan/tier | Free, basic, premium, enterprise | Monetization funnel health | | Traffic source | Specific campaigns, landing pages | Campaign effectiveness |

Segment Comparison Table Template

| Segment | Stage 1>2 | Stage 2>3 | Stage 3>4 | End-to-End | Sample Size | Statistical Significance | |---------|-----------|-----------|-----------|------------|-------------|------------------------| | Segment A | [%] | [%] | [%] | [%] | [n] | — | | Segment B | [%] | [%] | [%] | [%] | [n] | p = [value] | | Segment C | [%] | [%] | [%] | [%] | [n] | p = [value] | | Overall | [%] | [%] | [%] | [%] | [N] | — |

Step 5: Cohort Tracking

Track funnel performance over time to measure progress:

Cohort Analysis Table

| Cohort (Week) | Users | Day 1 Conv. | Day 7 Conv. | Day 14 Conv. | Day 30 Conv. | Final Conv. | |---------------|-------|-------------|-------------|--------------|--------------|-------------| | Week 1 | [n] | [%] | [%] | [%] | [%] | [%] | | Week 2 | [n] | [%] | [%] | [%] | [%] | [%] | | Week 3 | [n] | [%] | [%] | [%] | [%] | [%] | | Week 4 | [n] | [%] | [%] | [%] | [%] | [%] |

Trend Indicators

| Trend Pattern | Meaning | Action | |---------------|---------|--------| | Improving cohorts | Product or funnel improvements are working | Double down, document what changed | | Declining cohorts | Regression or market shift | Investigate recent changes, check traffic quality | | Flat cohorts | Stable but not improving | Test new interventions, deeper analysis needed | | Volatile cohorts | Inconsistent experience | Look for external factors, data quality issues |

Step 6: Optimization Recommendations

Prioritize improvements using an impact framework:

Recommendation Template

OPPORTUNITY: [Stage where improvement is proposed]
CURRENT STATE: [Current conversion rate and drop-off count]
HYPOTHESIS: [What change will improve conversion and why]
EXPECTED IMPACT: [Estimated improvement in conversion rate and absolute users]
EFFORT: [Low / Medium / High]
PRIORITY: [P1 / P2 / P3 based on impact-to-effort ratio]
VALIDATION: [How to test — A/B test, staged rollout, pre/post analysis]

Common Optimization Levers

| Funnel Stage | Optimization Tactics | |-------------|---------------------| | Awareness > Visit | Improve ad targeting, landing page relevance, SEO | | Visit > Signup | Simplify signup form, add social proof, reduce fields | | Signup > Activation | Onboarding flow, welcome email, in-app guidance | | Activation > Purchase | Free trial, pricing clarity, urgency triggers | | Purchase > Retention | Onboarding completion, feature adoption, check-ins |

Output Format

Present the funnel analysis as:

1. Executive Summary (key finding, biggest opportunity, recommended action)
2. Funnel Definition (stages, time window, scope, date range)
3. Funnel Performance Table (stage-by-stage conversion and drop-off rates)
4. Drop-Off Deep Dive (top 2-3 problem stages with root cause analysis)
5. Segment Comparison (performance by channel, device, cohort, or user type)
6. Cohort Trends (are things getting better or worse over time?)
7. Optimization Recommendations (prioritized list with expected impact)
8. Measurement Plan (how to track the impact of recommended changes)

Quality Checklist

Before delivering the funnel analysis, verify:

• [ ] Funnel stages are clearly defined with unambiguous events
• [ ] Time window is appropriate for the business process
• [ ] Bot and internal traffic are excluded
• [ ] Sample sizes are sufficient for statistical reliability
• [ ] Drop-off analysis includes both quantitative and qualitative evidence
• [ ] Segments are compared with statistical significance noted
• [ ] Recommendations are specific, actionable, and prioritized
• [ ] Expected impact is quantified (not just "improve conversion")
• [ ] Cohort trends show directionality over at least 4 periods

Edge Cases

• Low-traffic funnels (< 1000 users/month): Use longer time windows; avoid over-segmenting; apply Bayesian methods for significance testing
• Non-linear funnels: Users may skip stages or revisit; consider event-based analysis rather than strict sequential funnels
• Multi-device journeys: Users start on mobile and finish on desktop; use user-level (not session-level) funnels with cross-device identity
• B2B funnels with long cycles: Extend the funnel time window to weeks or months; track intermediate engagement signals
• Funnels with optional stages: Analyze both the strict path and the path with optional stages; compare conversion of those who did vs skipped optional steps
• Seasonal products: Compare same-period year-over-year rather than sequential months to avoid misleading trends

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: ashutoshsrivastava17
• Source: ashutoshsrivastava17/skill-library
• License: MIT
• Homepage: https://github.com/ashutoshsrivastava17/skill-library#quick-start

Install and usage instructions live in the source repository linked above.