Skip to main content

Feedback SDKs

Track user interactions and feedback signals to improve recommendation accuracy and business outcomes.

Overview​

Feedback tracking is essential for:

  • Improving Recommendations - Learn from user behavior to enhance future suggestions
  • Measuring Performance - Track conversion rates and engagement metrics
  • Business Intelligence - Understand user preferences and patterns
  • A/B Testing - Compare different recommendation strategies

Installation​

Python SDK​

pip install dodo-decisions

JavaScript/Node.js SDK​

npm install @dodo/sdk

Python SDK​

Initialization​

import dodo

# Initialize SDK with your access token
client = dodo.Client(api_key="your_access_token")

Tracking User Interactions​

Product Views​

# Track when user views a recommended product
client.track({
    "user_id": "user123",
    "decision_id": "dec_123456",
    "context": {
        "item_id": "warm-sweater-123",
        "page": "product_detail",
        "event": "product_view",
        "timestamp": "2025-10-24T10:30:00Z"
    }
})

Click Events​

# Track clicks on recommendations
client.track({
    "user_id": "user123",
    "decision_id": "dec_123456", 
    "context": {
        "item_id": "running-shoes-456",
        "page": "homepage",
        "event": "recommendation_click",
        "position": 1,  # Position in recommendation list
        "timestamp": "2025-10-24T10:35:00Z"
    }
})

Purchases/Conversions​

# Track successful conversions
client.track({
    "user_id": "user123",
    "decision_id": "dec_123456",
    "context": {
        "item_id": "warm-sweater-123",
        "page": "checkout",
        "event": "purchase",
        "value": 89.99,
        "currency": "USD",
        "timestamp": "2025-10-24T11:00:00Z"
    }
})

Negative Feedback​

# Track when user dismisses or ignores recommendations
client.track({
    "user_id": "user123",
    "decision_id": "dec_123456",
    "context": {
        "item_id": "unwanted-item-789",
        "page": "search_results",
        "event": "recommendation_dismiss",
        "reason": "not_interested",
        "timestamp": "2025-10-24T10:40:00Z"
    }
})

Batch Tracking​

# Track multiple events at once
events = [
    {
        "user_id": "user123",
        "decision_id": "dec_123456",
        "context": {"event": "view", "item_id": "item1"}
    },
    {
        "user_id": "user123", 
        "decision_id": "dec_123456",
        "context": {"event": "click", "item_id": "item2"}
    }
]

client.track_batch(events)

JavaScript SDK​

Browser Implementation​

import { DodoClient } from '@dodo/sdk';

// Initialize SDK
const dodo = new DodoClient('your-api-key');

// Track product views
dodo.track('view', {
  user_id: 'user123',
  decision_id: 'dec_123456',
  context: {
    item_id: 'warm-sweater-123',
    page: 'product_detail',
    event: 'product_view',
    position: 1
  }
});

// Track clicks
dodo.track('click', {
  user_id: 'user123',
  context: {
    item_id: 'running-shoes-456',
    page: 'homepage',
    event: 'recommendation_click',
    position: 2
  }
});

// Track purchases
dodo.track('purchase', {
  user_id: 'user123',
  decision_id: 'dec_123456',
  context: {
    item_id: 'warm-sweater-123',
    page: 'checkout',
    event: 'purchase',
    value: 89.99,
    currency: 'USD'
  }
});

Node.js Backend Implementation​

const { DodoClient } = require('@dodo/sdk');

const dodo = new DodoClient('your-api-key');

// Track server-side events
async function trackConversion(userId, decisionId, itemId, value) {
  try {
    await dodo.track('purchase', {
      user_id: userId,
      decision_id: decisionId,
      context: {
        item_id: itemId,
        event: 'purchase',
        value: value,
        currency: 'USD',
        timestamp: new Date().toISOString()
      }
    });
    console.log('Conversion tracked successfully');
  } catch (error) {
    console.error('Error tracking conversion:', error);
  }
}

Event Types​

Standard Events​

  • view - User views a recommended item
  • click - User clicks on a recommendation
  • purchase - User completes a purchase/conversion
  • add_to_cart - Item added to shopping cart
  • wishlist - Item added to wishlist
  • share - Recommendation shared with others
  • dismiss - User ignores/dismisses recommendation

Custom Events​

You can define custom events specific to your business:

# Media website example
client.track({
    "user_id": "user123",
    "decision_id": "dec_123456",
    "context": {
        "article_id": "ai-breakthrough-2025",
        "page": "article_page",
        "event": "article_read_complete",
        "read_time": 245,  # seconds
        "scroll_depth": 0.95
    }
})

# Travel website example
client.track({
    "user_id": "user123",
    "decision_id": "dec_123456", 
    "context": {
        "hotel_id": "luxury-resort-789",
        "page": "booking_page",
        "event": "booking_initiated",
        "stay_duration": 7,
        "room_type": "suite"
    }
})

Best Practices​

1. Always Include Decision ID​

# ✅ Good - Includes decision_id
client.track({
    "user_id": "user123",
    "decision_id": "dec_123456",  # Links feedback to specific recommendation
    "context": {"event": "click", "item_id": "item123"}
})

# ❌ Bad - Missing decision_id
client.track({
    "user_id": "user123",
    "context": {"event": "click", "item_id": "item123"}
})

2. Use Consistent User IDs​

Ensure the same user_id is used across all tracking for a given user.

3. Include Timestamps​

import datetime

client.track({
    "user_id": "user123",
    "decision_id": "dec_123456",
    "context": {
        "event": "purchase",
        "timestamp": datetime.datetime.utcnow().isoformat() + "Z"
    }
})

4. Handle Errors Gracefully​

try:
    client.track(event_data)
except Exception as e:
    # Log error but don't break user experience
    logger.error(f"Failed to track event: {e}")

Direct API Integration​

If you prefer not to use SDKs, you can track feedback directly via API:

curl -X POST "https://api.trydodo.xyz/api/feedback/track?project_id=${PROJECT_ID}" \
  -H "Authorization: Bearer ${ACCESS_TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{
    "project_id": "your_project_id",
    "decision_id": "dec_123456",
    "feedback": "positive",
    "created_at": "2025-10-24T00:00:00Z"
  }'

Data Privacy​

  • No PII - Avoid sending personally identifiable information
  • Data Minimization - Send only necessary data for feedback tracking
  • User Consent - Ensure proper consent mechanisms are in place
  • GDPR Compliance - Follow applicable data protection regulations

Troubleshooting​

Common Issues​

  1. Missing Decision ID

    • Ensure you're capturing and passing the decision_id from recommendation responses

  2. Invalid API Key

    • Verify your access token is valid and not expired
    • Use the refresh token mechanism if needed

  3. Network Issues

    • Implement retry logic for failed requests
    • Consider batching events to reduce API calls

Debug Mode​

Enable debug logging to troubleshoot issues:

import logging
logging.basicConfig(level=logging.DEBUG)

client = dodo.Client(api_key="your_token", debug=True)

Next Steps​