Alerts and anomaly detection
Purpose: Get notified when metrics move.
Who this page is for
| Audience | Why it matters to you |
|---|---|
| Analysts | Monitor KPIs |
| Managers | Stay informed |
Alerts and Anomaly Detection
Intelligent Alerts
Set a rule:
- "Alert me if Visitors drops > 20% vs last week"
- "Alert me if Revenue > 50K"
- "Alert me if Bounce Rate changes > 15%"
Configure:
- Metric
- Threshold
- Frequency (daily, weekly)
- Notification method (email, Slack)
Use: Early warning system for problems
Anomaly Detection
Adobe uses machine learning to spot unusual values:
Example:
- Normal pattern: 1000 visitors/day
- Today: 50 visitors (anomaly detected)
- Adobe highlights it in reports
How it works:
- Learns normal range (last 35 days)
- Detects deviations (upper/lower bounds)
- Shows confidence (95%, 99%)
Use: Spot problems you didn't know to look for
Anomaly use cases
"Traffic is down"
- Anomaly detection flags 50% drop vs normal
- Investigate immediately
- Find root cause (site outage, tracking bug, etc.)
"Revenue spike"
- Anomaly detection flags 200% increase
- Investigate: Is it real or a bug?
- Celebrate if real, fix if not
---
End of Analysis & Reporting