Traffic Anomaly Detection: How to Monitor and Analyze Traffic Changes

Traffic Anomaly Detection: How to Monitor and Analyze Traffic Changes

By Michael Thompson

March 16, 2025 at 10:36 PM

Traffic Alerts effectively monitor and analyze significant changes in your website's traffic patterns, helping you understand performance across different marketing channels.

Traffic alerts use advanced anomaly detection through machine learning to identify statistically significant traffic changes by comparing current traffic to the previous 28 days of data. A significant change occurs when traffic falls outside 99% of the normal distribution.

alt text

alt text

How to View and Analyze Anomalies:

  1. Open Analytics panel and select Traffic
  2. Choose desired date range from dropdown menu
  3. Click alert icons to view detailed information

Anomaly details include:

  • Event date
  • Traffic volume during anomaly
  • Top affected pages (up to 2) and their percentage changes
  • Main contributing traffic channels
  • For spikes: Most viewed pages during peak traffic
  • For drops: Pages with largest decreases

Using Anomaly Data:

  • Identify successful marketing channels
  • Understand traffic patterns
  • Adjust strategies based on performance
  • Test different approaches for underperforming channels

Best Practices:

  • Monitor regularly for pattern changes
  • Analyze contributing factors to spikes/drops
  • Implement data-driven marketing adjustments
  • Test different strategies for improvement

Provide Feedback:

  1. Locate "Was this helpful?" at bottom of anomaly details
  2. Click thumbs up/down
  3. Add optional detailed comments
  4. Submit feedback to help improve the tool's accuracy

Note: Site must have at least 28 days of traffic data for anomaly detection to function effectively. Results and improvement strategies will vary based on individual content, goals, and target audience.

Traffic alerts serve as a valuable tool for understanding website performance and making informed marketing decisions based on actual data and traffic patterns.

Related Articles

Previous Articles