Automated Ad Performance Insights & Notifications

Overview

A production-ready, enterprise-level automation system that automatically analyzes Google Ads performance data, provides AI-powered insights, and sends intelligent notifications based on performance trends. Built with comprehensive error handling and real-world scenario management.

Key Features

Core Functionality

• Automatic Triggering: Responds to new data in Google Sheets
• AI-Powered Analysis: Uses Google Gemini AI for intelligent performance insights
• Smart Notifications: Conditional Slack alerts based on performance status
• Data Validation: Comprehensive input validation with error handling
• Historical Context: Week-over-week comparison or industry benchmarks

Robustness & Reliability

• What-If Logic: Handles all edge cases gracefully
• Retry Mechanisms: Automatic retry on AI failures
• Fallback Systems: Manual review channels for unexpected scenarios
• Error Monitoring: Comprehensive error tracking and alerting
• Data Quality Checks: Validates input data before processing

Performance Analysis

• Week-over-Week Comparison: When previous week data is available
• Industry Benchmarks: When historical data is missing
• Structured Insights: Consistent JSON output format
• Actionable Recommendations: Specific optimization suggestions

Architecture

Phase 1: Data Validation & Quality Check

• Validates incoming Google Sheets data
• Checks for missing fields, invalid types, and #REF! errors
• Routes invalid data to error notification channel

Phase 2: Context Preparation & AI Analysis

• Fetches historical data for context
• Prepares data for AI analysis
• Handles missing previous week data gracefully

Phase 3: AI Performance Analysis

• Leverages Google Gemini AI for intelligent analysis
• Provides structured insights and recommendations
• Supports both comparison and benchmark analysis

Phase 4: AI Output Validation & What-If Logic

• Validates AI output structure and content
• Implements retry logic for failed AI responses
• Routes invalid outputs to manual review

Phase 5: Sheet Writeback & Conditional Notifications

• Updates Google Sheets with AI insights
• Sends appropriate Slack notifications based on performance
• Handles all notification scenarios

🔧 Technical Specifications

Platform

• Built on: n8n (Node.js-based workflow automation)
• AI Model: Google Gemini 2.0 Flash
• Integrations: Google Sheets, Slack, Google Gemini API

Data Requirements

Required Google Sheets Columns:

• Week (numeric)
• Campaign Name (text)
• Status (text)
• Budget (€) (numeric)
• CTR (%) (numeric)
• Clicks (numeric)
• CPC (€) (numeric)
• Cost (€) (numeric)
• Conversions (numeric)
• Impression Share (%) (numeric)

Output Format

AI Analysis Output:

{
  "interpretation": "Performance analysis summary (max 200 words)",
  "recommendations": ["Actionable recommendation 1", "Actionable recommendation 2"]
}

📱 Notification System

Slack Channels

1. Stable Performance (#for-stable-good-updates)

   • Triggered when performance is stable/improved
   • Green circle emoji, positive messaging

2. Action Required (#marketing-action-items)

   • Triggered when optimization is needed
   • Red circle emoji, actionable recommendations

3. Error Handling (#workflow-errors)

   • Data validation failures
   • AI analysis failures
   • System-level errors

Message Examples

Stable Performance:

🟢 Weekly Ad Performance Update - Stable Performance

Campaign: Campaign 1
Week: 1
Analysis Context: week-over-week

AI Summary: Week 27 shows stable performance with CTR at 2.5%...

✅ All systems go! Performance is stable this week.

Action Required:

🔴 Action Required: Weekly Ad Performance Alert!

Campaign: Campaign 1
Week: 1
Analysis Context: industry-benchmarks

AI Summary: CTR at 1.2% is below industry average...

Recommendations:
• 1. Optimize ad copy to improve click-through rates
• 2. Review keyword targeting for better relevance

🚨 Let's review this ASAP and take action.

🚀 What-If Scenarios Handled

Data Quality Issues

• ✅ Missing required fields → Error notification
• ✅ Invalid data types → Error notification
• ✅ #REF! errors → Error notification
• ✅ Empty values → Error notification

AI Analysis Issues

• ✅ Invalid JSON output → Retry logic (2 attempts)
• ✅ Missing required fields → Fallback to manual review
• ✅ Malformed structure → Fallback to manual review
• ✅ API failures → Fallback to manual review
• ✅ Ambiguous output → Manual review channel

Context Issues

• ✅ Missing previous week data → Industry benchmarks
• ✅ Multiple campaigns → Campaign-specific matching
• ✅ Data gaps → Graceful degradation

System Issues

• ✅ Workflow errors → Global error handler
• ✅ Node failures → Comprehensive error logging
• ✅ Network issues → Retry mechanisms

Performance Metrics

Industry Benchmarks Used

• CTR: Good >2.0%, Average 1.5-2.0%, Poor <1.5%
• CPC: Good <€2.0, Average €2.0-3.0, Poor >€3.0
• Conversion Rate: Good >3%, Average 2-3%, Poor <2%
• Impression Share: Good >80%, Average 50-80%, Poor <50%

Analysis Contexts

1. Week-over-Week: When previous week data is available
2. Industry Benchmarks: When historical data is missing
3. Isolated Analysis: For first-time campaigns

Access Control

• Google Sheets integration with OAuth2
• Slack workspace-specific channels
• API key management through n8n credentials

📋 Setup Instructions

Prerequisites

1. n8n instance (self-hosted or cloud)
2. Google Sheets with performance data
3. Slack workspace with designated channels
4. Google Gemini API access

Configuration Steps

1. Import Workflow: Import the JSON file into n8n
2. Configure Credentials: Set up Google Sheets and Slack credentials
3. Update Sheet IDs: Replace Google Sheet document IDs
4. Update Channel IDs: Replace Slack channel IDs
5. Test Workflow: Add test data to Google Sheets
6. Activate Workflow: Enable the workflow for production

Required Permissions

• Google Sheets: Read/Write access to performance data sheet
• Slack: Post messages to designated channels
• Google Gemini: API access for AI analysis

Testing

Test Scenarios

1. Valid Data Flow: Add valid performance data
2. Missing Previous Week: Test with first week data
3. Invalid Data: Test with missing/invalid fields
4. AI Failure: Test with malformed AI output
5. System Error: Test error handling mechanisms

Expected Behaviors

• ✅ Valid data → AI analysis → Appropriate notification
• ✅ Invalid data → Error notification → No processing
• ✅ AI failure → Retry → Fallback notification
• ✅ System error → Global error notification

Support & Maintenance

Monitoring

• Check Slack error channels regularly
• Monitor workflow execution logs
• Review AI analysis quality periodically
• Track notification accuracy

Troubleshooting

• Data Issues: Check Google Sheets format and permissions
• AI Issues: Verify Gemini API credentials and quotas
• Notification Issues: Check Slack channel permissions
• Workflow Issues: Review n8n execution logs

Updates & Maintenance

• Regular workflow performance reviews
• AI prompt optimization based on feedback
• Benchmark updates as industry standards change
• Error handling improvements based on real-world usage

Business Value

Immediate Benefits

• Automated Insights: No manual analysis required
• Proactive Alerts: Early warning for performance issues
• Consistent Reporting: Standardized analysis format
• Time Savings: Automated weekly performance reviews

Long-term Value

• Performance Optimization: Data-driven improvement recommendations
• Cost Efficiency: Reduced manual analysis overhead
• Scalability: Handles multiple campaigns and accounts
• Reliability: Robust error handling ensures continuous operation

---

Built with ❤️ for automated performance optimization