Daily anomaly scan: spot drops or spikes worth investigating
Compare yesterday to the trailing 14-day average. Flag any metric off by more than 30% with a probable cause.
The prompt
Copy this and paste it into Claude, Cursor, or any AI client connected to your GenieSeo MCP URL.
For my main Search Console property, compare YESTERDAY vs the trailing 14-day average for: clicks, impressions, average position, CTR.
For any metric whose delta exceeds ±30%:
1. Identify the top 5 queries and top 5 pages that drove the swing.
2. Note whether the swing is concentrated (1–2 URLs) or distributed (across the site).
3. Suggest the most likely cause — algorithm update, indexing event, content edit, seasonality, ranking-position change, sudden CTR change.
4. Recommend ONE first action to take today.
If no metric exceeds the threshold, output a one-line "all clear" summary with the actual deltas.
Format as a short Markdown report I could paste into Slack.When to use it
Run daily (or as a Monday catch-up). Catches deindexings, algorithm hits, and viral traffic in 24 hours instead of a week.
What you get back
- Comparison table: yesterday vs trailing 14-day average for clicks, impressions, average position, CTR.
- Per-anomaly investigation: which queries / pages drove the swing.
- Probable cause + first action to take.
How this runs
GenieSeo doesn't execute prompts — your AI client does. Make sure your MCP URL is connected to Claude, Cursor, Codex, or Antigravity. New here? Read the 60-second setup guide.
Automate it
Run this every morning with a Claude Code Routine. Hands-off — the report lands in your inbox or Slack on schedule.
Tags
- alerts
- monitoring
- daily