Fitness Studio Member Retention Engine (Fork) (Fork) (Fork)

How It Works

1. 1
   Data Extraction

   Connect to Mindbody API and pull previous 7 days of check-in data, class bookings, cancellations, and member profile updates. Retrieve historical baseline data for comparison analysis.

   Mindbody API v6
2. 2
   Pattern Analysis

   Send attendance data to Claude with structured prompts analyzing frequency changes, booking-to-attendance ratios, and engagement trends. Calculate risk scores based on deviation from member's historical patterns.

   Claude API
3. 3
   Risk Classification

   Apply business rules to Claude's analysis results, categorizing members as low, medium, or high churn risk based on attendance decline percentage, days since last visit, and engagement score changes.

   Custom scoring algorithm
4. 4
   Alert Generation

   Format at-risk member data into structured JSON payloads containing member ID, risk level, specific concerns, and recommended intervention strategies for downstream processing.

   JSON formatting
5. 5
   Trigger Activation

   Send formatted at-risk member data to Win-Back Campaigner through internal API call or message queue, ensuring reliable delivery and preventing duplicate processing.

   HTTP POST or Redis queue

Implementation

# Churn Risk Scanner Implementation

## File Structure
```
churn_scanner/
├── main.py                 # Main execution script
├── data_extractor.py      # Mindbody API client
├── risk_analyzer.py       # Claude integration
├── config/
│   ├── settings.py        # Configuration management
│   └── prompts.py         # Claude analysis prompts
├── utils/
│   ├── logging.py         # Structured logging
│   └── cache.py           # Data caching utilities
└── tests/
    ├── test_extractor.py
    └── test_analyzer.py
```

## Environment Variables
```bash
MINDBODY_API_KEY=your_api_key
MINDBODY_SITE_ID=your_site_id
CLAUDE_API_KEY=your_claude_key
WINBACK_WEBHOOK_URL=http://localhost:8080/trigger
LOG_LEVEL=INFO
CACHE_DURATION=3600
```

## Key Functions

### main.py
```python
def weekly_scan():
    extractor = MindbodyExtractor()
    analyzer = RiskAnalyzer()
    
    # Pull weekly data
    attendance_data = extractor.get_weekly_attendance()
    member_profiles = extractor.get_active_members()
    
    # Analyze patterns
    risk_results = analyzer.analyze_churn_risk(attendance_data, member_profiles)
    
    # Process at-risk members
    for member in risk_results.high_risk_members:
        trigger_winback_campaign(member)
    
    log_results(risk_results)
```

## Cron Setup
```bash
# Add to crontab
0 7 * * 1 /usr/bin/python3 /opt/churn_scanner/main.py >> /var/log/churn_scanner.log 2>&1
```

## Data Processing Pipeline
1. Initialize API clients with retry logic
2. Extract attendance data with date range validation
3. Cache member baseline data for comparison
4. Send structured prompts to Claude with attendance patterns
5. Parse Claude responses into risk scores
6. Apply business rule filters
7. Trigger downstream actions for flagged members

Data Flow

Prerequisites

• -Mindbody API credentials with read access to member and attendance data
• -Claude API key with sufficient monthly token allowance
• -Redis or similar caching system for baseline data storage
• -Cron job scheduling permissions on target server

Error Handling