Mixpanel Funnel Analysis

Guide the user through building and analyzing conversion funnels using saved Mixpanel funnels or custom event sequences.

Pre-flight Check

Verify credentials are configured:

!$(mp auth test 2>&1 || echo "No credentials configured")

If credentials aren't configured, suggest running /mp-auth first.

Funnel Analysis Options

Present two approaches:

1. Saved Funnel - Use existing funnel from Mixpanel UI (faster)
2. Custom Funnel - Build funnel from local event data (more flexible)

Ask which approach the user prefers.

---

Option 1: Saved Funnel Analysis (Live API)

1. List Available Funnels

!$(mp inspect funnels --format table)

Show the user:

• Funnel ID
• Funnel name
• Number of steps

Tip: Use /mp-inspect to discover available events and cohorts for analysis context.

2. Select Funnel

If $1 is provided, use that as funnel ID. Otherwise, ask user to choose from the list.

3. Configure Analysis Parameters

Date Range (required):

• From date (YYYY-MM-DD)
• To date (YYYY-MM-DD)
• Validate range is reasonable (suggest ≤ 90 days for performance)

Time Unit (optional):

• day (default)
• week
• month

Segmentation (optional):

• Property to segment by (e.g., country, plan, source)
• Shows conversion rates broken down by this property

4. Execute Funnel Query

Basic funnel:

mp query funnel <funnel-id> \
  --from <from-date> \
  --to <to-date> \
  --format table

With segmentation:

mp query funnel <funnel-id> \
  --from <from-date> \
  --to <to-date> \
  --unit <day|week|month> \
  --on <property> \
  --format table

5. Interpret Results

Show the user:

• Overall conversion rate: Final step / First step
• Step-by-step drop-off: Where users are leaving
• Conversion rates by segment (if segmented)

Key insights to highlight:

• Which step has the biggest drop-off?
• What's the overall conversion rate?
• How do segments compare?

---

Option 2: Custom Funnel from Local Data

1. Check Available Tables

!$(mp inspect tables --format table)

If no tables exist, suggest running /mp-fetch first.

2. Define Funnel Steps

Ask the user for the event sequence (3-5 events recommended):

• Step 1: Event name (e.g., "View Product")
• Step 2: Event name (e.g., "Add to Cart")
• Step 3: Event name (e.g., "Checkout")
• Step 4: Event name (e.g., "Purchase")

3. Configure Analysis Parameters

Table: Which table to analyze Time window: Maximum time between steps (e.g., 24 hours, 7 days) Date range: Filter events by date

4. Build SQL Funnel Query

Use a window function approach to detect sequences:

WITH user_events AS (
  SELECT
    distinct_id,
    event_name,
    event_time,
    ROW_NUMBER() OVER (PARTITION BY distinct_id ORDER BY event_time) as event_seq
  FROM <table>
  WHERE event_name IN ('Step1', 'Step2', 'Step3', 'Step4')
    AND event_time >= '<from-date>'
    AND event_time <= '<to-date>'
),
funnel_progression AS (
  SELECT
    distinct_id,
    MAX(CASE WHEN event_name = 'Step1' THEN 1 ELSE 0 END) as completed_step1,
    MAX(CASE WHEN event_name = 'Step2' THEN 1 ELSE 0 END) as completed_step2,
    MAX(CASE WHEN event_name = 'Step3' THEN 1 ELSE 0 END) as completed_step3,
    MAX(CASE WHEN event_name = 'Step4' THEN 1 ELSE 0 END) as completed_step4
  FROM user_events
  GROUP BY distinct_id
)
SELECT
  SUM(completed_step1) as step1_users,
  SUM(completed_step2) as step2_users,
  SUM(completed_step3) as step3_users,
  SUM(completed_step4) as step4_users,
  ROUND(100.0 * SUM(completed_step2) / NULLIF(SUM(completed_step1), 0), 2) as step1_to_step2_rate,
  ROUND(100.0 * SUM(completed_step3) / NULLIF(SUM(completed_step2), 0), 2) as step2_to_step3_rate,
  ROUND(100.0 * SUM(completed_step4) / NULLIF(SUM(completed_step3), 0), 2) as step3_to_step4_rate,
  ROUND(100.0 * SUM(completed_step4) / NULLIF(SUM(completed_step1), 0), 2) as overall_conversion
FROM funnel_progression

Execute via:

mp query sql "<query>" --format table

5. Visualize Results

Offer to create a Python visualization:

import pandas as pd
import matplotlib.pyplot as plt

# Data from query results
steps = ['Step1', 'Step2', 'Step3', 'Step4']
users = [<step1_users>, <step2_users>, <step3_users>, <step4_users>]

# Create funnel chart
fig, ax = plt.subplots(figsize=(10, 6))
ax.barh(steps, users, color=['#4CAF50', '#8BC34A', '#CDDC39', '#FFEB3B'])
ax.set_xlabel('Number of Users')
ax.set_title('Conversion Funnel')

# Add conversion rates
for i, (step, count) in enumerate(zip(steps, users)):
    ax.text(count, i, f' {count:,} users', va='center')

plt.tight_layout()
plt.savefig('funnel.png')
print("Funnel visualization saved to funnel.png")

---

Advanced Analysis: Segmented Funnels

For local data, offer segmented funnel analysis:

Segment by property (e.g., country, plan, source):

WITH user_events AS (
  SELECT
    distinct_id,
    event_name,
    event_time,
    properties->>'$.country' as country
  FROM <table>
  WHERE event_name IN ('Step1', 'Step2', 'Step3', 'Step4')
),
funnel_by_segment AS (
  SELECT
    country,
    MAX(CASE WHEN event_name = 'Step1' THEN 1 ELSE 0 END) as completed_step1,
    MAX(CASE WHEN event_name = 'Step2' THEN 1 ELSE 0 END) as completed_step2,
    MAX(CASE WHEN event_name = 'Step3' THEN 1 ELSE 0 END) as completed_step3,
    MAX(CASE WHEN event_name = 'Step4' THEN 1 ELSE 0 END) as completed_step4
  FROM user_events
  GROUP BY distinct_id, country
)
SELECT
  country,
  SUM(completed_step1) as step1_users,
  ROUND(100.0 * SUM(completed_step4) / NULLIF(SUM(completed_step1), 0), 2) as conversion_rate
FROM funnel_by_segment
GROUP BY country
ORDER BY step1_users DESC

---

Common Funnel Patterns

E-commerce Funnel:

1. Product View → Add to Cart → Checkout → Purchase

SaaS Signup Funnel:

1. Landing Page → Sign Up → Onboarding → First Action

Content Engagement Funnel:

1. Page View → Scroll Depth → Click CTA → Convert

Mobile App Funnel:

1. App Open → Feature Discovery → Feature Use → Retention

---

Insights and Recommendations

After showing results, provide actionable insights:

High Drop-off Detection

If any step shows >50% drop-off:

⚠️ High drop-off detected at Step X → Step Y (Z% drop)

Recommendations:

• Investigate UX issues at Step X
• Add tracking for abandonment reasons
• A/B test simplified flow
• Check for technical errors

Segment Performance Comparison

If segmented:

📊 Segment insights:

• Best performing: [Segment A] (X% conversion)
• Worst performing: [Segment B] (Y% conversion)
• Delta: Z percentage points

Recommendations:

• Study what makes [Segment A] successful
• Optimize experience for [Segment B]
• Consider separate funnels per segment

Time-Based Trends

If using time units:

📈 Trend analysis:

• Conversion improving/declining over time
• Seasonal patterns detected

Recommendations:

• Investigate what changed during improvements
• Prepare for seasonal variations

---

Next Steps

After analysis, suggest:

1. Deep dive on specific step:

   /mp-query sql
   # Analyze users who dropped at specific step
   

2. Retention analysis:

   /mp-retention
   # Analyze retention for users who completed funnel
   

3. Export for presentation:

   mp query funnel <id> --from <date> --to <date> --format csv > funnel_results.csv
   

4. Create custom report:

   /mp-report funnel
   # Generate comprehensive funnel report
   

---

Troubleshooting

"Funnel not found":

• Check funnel ID with mp inspect funnels
• Verify access permissions in Mixpanel

"No events found":

• Verify event names exactly match (case-sensitive)
• Check date range includes relevant data
• Run mp inspect events to see available events

"Low sample size":

• Expand date range
• Check if events are being tracked correctly
• Verify filtering isn't too restrictive

"Segmentation returns null":

• Property might not exist on all events
• Check property name with mp inspect properties <event>
• Use defined(properties["prop"]) in filters