The Google Search Console Signal Nobody Checks Until It's Too Late

Ask any SEO what they check in Search Console weekly and you'll hear performance reports and coverage errors. Ask what they check in the crawl stats report and you'll mostly get a blank look. That report, buried under settings, is the one that would've given us three weeks' advance warning on a problem that instead cost us three weeks of lost indexing.

Crawl stats shows you average response time, total crawl requests, and a breakdown by response code over time — it's the closest thing Search Console gives you to a leading indicator instead of a lagging one. Rankings and coverage reports tell you what already happened. Crawl stats can tell you something is starting to go wrong before it shows up anywhere else.

What it actually caught, eventually

A server migration quietly doubled average response time for Googlebot, well within what human users would ever notice, but crawl stats showed total crawl requests dropping steadily over three weeks as Googlebot throttled itself in response to the slower responses. By the time it showed up as a coverage issue, new content had already started indexing noticeably slower.

Rankings tell you what happened. Crawl stats tells you what's about to.

What we changed

• Added crawl stats to the weekly SEO health check instead of the quarterly one
• Set an internal alert threshold on average response time, tied to server monitoring, not just SEO reporting
• Treat any sustained drop in total crawl requests as a server-side investigation trigger, not just an SEO one

This isn't a glamorous fix and it won't show up in a client-facing report as a win. But catching a response-time regression in week one instead of week four is the difference between a quiet server ticket and a month of degraded indexing that's much harder to explain after the fact.