Example Workflows¶
Real-world conversation examples demonstrating the MCP server's capabilities.
Schema Discovery¶
Understand what data exists in your Mixpanel project before querying.
Explore Events and Properties¶
You: What events are tracked in my Mixpanel project?
Claude: Uses list_events tool
Your project tracks 47 events. Here are the most common:
| Event | Description |
|---|---|
| Login | User login |
| Signup | New user registration |
| Purchase | Completed purchase |
| Page View | Page visited |
| Feature Used | Feature interaction |
You: What properties are on the Purchase event?
Claude: Uses list_properties tool with event="Purchase"
The Purchase event has 12 properties:
| Property | Type | Description |
|---|---|---|
| revenue | number | Purchase amount |
| currency | string | Currency code |
| product_id | string | Product identifier |
| category | string | Product category |
| payment_method | string | Payment type used |
| ... | ... | ... |
You: What payment methods do people use?
Claude: Uses list_property_values tool
Top payment methods: - credit_card (67%) - paypal (22%) - apple_pay (8%) - google_pay (3%)
Live Analytics¶
Query Mixpanel directly for real-time insights.
Time Series Analysis¶
You: How many signups happened each day last week?
Claude: Uses segmentation tool
Jan 15: 1,234 signups
Jan 16: 1,189 signups
Jan 17: 1,456 signups
Jan 18: 1,102 signups
Jan 19: 892 signups (weekend)
Jan 20: 847 signups (weekend)
Jan 21: 1,298 signups
Week-over-week: +3.2% vs previous week
Funnel Analysis¶
You: What's the conversion rate for my signup funnel?
Claude: Uses list_funnels then funnel tools
Signup Funnel (last 30 days):
| Step | Users | Conversion |
|---|---|---|
| 1. Landing Page | 45,234 | 100% |
| 2. Start Signup | 12,456 | 27.5% |
| 3. Complete Form | 8,234 | 66.1% |
| 4. Verify Email | 6,789 | 82.4% |
| 5. First Login | 5,123 | 75.5% |
Overall conversion: 11.3%
The biggest drop-off is between Landing Page and Start Signup (72.5% lost).
Retention Analysis¶
You: Show day-7 retention for users who signed up last month
Claude: Uses retention tool
Signup → Return Retention (January 2025 cohorts):
| Cohort | Day 0 | Day 1 | Day 7 | Day 14 | Day 30 |
|---|---|---|---|---|---|
| Jan 1-7 | 100% | 42% | 28% | 21% | 15% |
| Jan 8-14 | 100% | 45% | 31% | 24% | 18% |
| Jan 15-21 | 100% | 43% | 29% | 22% | — |
| Jan 22-28 | 100% | 44% | — | — | — |
Day-7 retention is averaging 29%, which is healthy for this product category.
Intelligent Analysis (v2)¶
AI-powered tools for deeper insights.
Diagnose Metric Drops¶
You: Why did signups drop on January 7th?
Claude: Uses diagnose_metric_drop tool
Analysis of Signup Drop on January 7th:
Signups dropped 47% (1,234 → 654) compared to the previous day.
Root Causes Identified:
- Geographic: US signups dropped 62% while other regions were stable
-
Timing coincides with a major US holiday
-
Traffic Source: Paid acquisition channels showed 78% decline
-
Suggests ad spend was paused or campaigns ended
-
Device: Mobile web signups dropped 71%
- Possible technical issue on mobile landing page
Recommendation: Check if paid campaigns were intentionally paused for the holiday. If not, investigate mobile landing page for issues.
Natural Language Queries¶
You: What features do our best users engage with?
Claude: Uses ask_mixpanel tool
Power User Feature Engagement (top 10% by event count):
Power users (>100 events/month) heavily engage with:
| Feature | Usage Rate | vs Average Users |
|---|---|---|
| Dashboard | 94% | +45% |
| Export | 78% | +62% |
| Collaboration | 71% | +89% |
| Advanced Filters | 67% | +134% |
| API Access | 45% | +312% |
Key Insight: Power users are 3x more likely to use the API. Consider promoting API access in onboarding for users who show power-user potential.
Funnel Optimization Report¶
You: Generate a funnel optimization report for my signup funnel
Claude: Uses funnel_optimization_report tool
Signup Funnel Optimization Report
Executive Summary: The signup funnel converts 11.3% of landing page visitors. There are two major optimization opportunities worth ~$2.4M annually.
Drop-off Analysis:
| Step | Drop-off | Impact | Priority |
|---|---|---|---|
| Landing → Start | 72.5% | High | 1 |
| Form → Verify | 17.6% | Medium | 3 |
| Verify → Login | 24.5% | Medium | 2 |
Recommendations:
- Landing Page CTA (Priority 1)
- Current conversion: 27.5%
- Benchmark: 35-40%
-
Action: A/B test more prominent signup CTA
-
Email Verification (Priority 2)
- 24.5% never verify email
-
Action: Implement delayed verification, allow limited access before verify
-
Form Optimization (Priority 3)
- Consider progressive profiling
- Remove non-essential fields
Interactive Workflows (v2)¶
Guided analysis with user confirmation.
Guided Analysis¶
You: Help me analyze my data
Claude: Uses guided_analysis tool
I'll help you analyze your Mixpanel data. To get started:
What's your analysis goal?
- Understand user behavior patterns
- Investigate a metric change
- Compare user segments
- Explore product usage
- Something else (describe)
You: I want to understand why conversions dropped last week
Claude: Let me investigate the conversion drop.
Step 1: Confirm scope - Metric: Signup-to-Purchase conversion - Period: Last 7 days vs previous 7 days - Proceed with this analysis? (Yes/No)
You: Yes
Claude: Executes analysis queries
Findings:
Conversion dropped from 4.2% to 3.1% (26% decline).
Contributing factors:
- Traffic quality changed
- Organic: Conversion stable at 5.1%
-
Paid: Dropped from 3.8% to 2.1%
-
New user segment
- Influx of users from a new ad campaign
- This cohort has 40% lower conversion baseline
Recommendation: The conversion drop is due to a new paid acquisition channel bringing lower-intent users. This is expected for new channels. Consider: - Adjusting targeting for the new channel - Creating a specific onboarding flow for these users
Local Analysis¶
Work with data stored in DuckDB.
Fetch and Query¶
You: Fetch events from January 1-7 and show me the top 10 users by event count
Claude: Uses fetch_events tool
Fetched 156,234 events to table jan_events.
Uses sql tool
Top 10 Users by Event Count:
| User | Events | First Seen | Last Seen |
|---|---|---|---|
| user_8234 | 2,456 | Jan 1 | Jan 7 |
| user_1892 | 1,923 | Jan 1 | Jan 7 |
| user_4521 | 1,845 | Jan 2 | Jan 7 |
| user_7234 | 1,712 | Jan 1 | Jan 6 |
| ... | ... | ... | ... |
These users are significantly more active than average (avg: 23 events/week).
Complex SQL Analysis¶
You: Find users who purchased but never logged in again
Claude: Uses sql tool
SELECT
p.distinct_id,
p.properties->>'$.email' as email,
MAX(e.event_time) as last_purchase,
COUNT(*) as purchase_count
FROM jan_events e
JOIN users p ON e.distinct_id = p.distinct_id
WHERE e.event_name = 'Purchase'
AND NOT EXISTS (
SELECT 1 FROM jan_events e2
WHERE e2.distinct_id = e.distinct_id
AND e2.event_name = 'Login'
AND e2.event_time > e.event_time
)
GROUP BY 1, 2
ORDER BY last_purchase DESC
Found 234 users who purchased but never returned. These are candidates for a win-back campaign.
Product Health Dashboard¶
You: Show me a product health dashboard for the last month
Claude: Uses product_health_dashboard tool
Product Health Dashboard (January 2025)
Acquisition¶
- New signups: 12,456 (+8% MoM)
- Top channels: Organic (45%), Paid (32%), Referral (23%)
Activation¶
- First value event rate: 67%
- Median time to activation: 4.2 hours
- Activation by channel varies: Organic 78%, Paid 52%
Retention¶
- Day 1: 45%
- Day 7: 29%
- Day 30: 18%
- Trending up 2% from last month
Revenue¶
- Total revenue: $234,567
- ARPU: $18.82
- Transactions: 12,456
Referral¶
- Referral rate: 12%
- Viral coefficient: 0.31
- Top referrers driving 45% of referral signups
Health Score: 7.2/10 (Healthy)
Primary growth lever: Improve paid channel activation (currently 52% vs 78% for organic).