When Google organic traffic drops suddenly, first confirm that the data is comparable, then check live pages and Search Console notices. The percentage decline, curve shape, brand position or duration cannot alone establish a manual action. A page opening on your own computer also does not rule out problems affecting other regions or crawlers.
Preserve a baseline before changing the website
Record when the issue was first noticed, the property and affected queries or directories, complete comparison periods, and recent publishing, migration, CDN, plugin or tracking changes. Export the relevant data before making several unrelated changes to titles, URLs, robots directives and content.
Google's traffic-drop guide covers performance data, technical issues, demand and search changes. A lower site-wide total needs investigation rather than an immediate domain-wide diagnosis.
1. Check reporting scope and measurement changes
- Keep the property, search type, country and device filters consistent, and compare complete periods with similar weekday composition.
- Check recent-data processing notices; an incomplete day is not comparable with a full day.
- Separate Search Console clicks from analytics visits. If Search Console is relatively stable while analytics falls, investigate collection, consent, tags and filtering.
- Consider seasonality, campaign endings and brand demand. Use a comparable prior-year period when available.
Privacy filtering and table-row limits can make query-table totals differ from chart totals. Do not diagnose lost traffic from that difference alone. For a stale report date, see the Search Console data-update workflow.
2. Identify which segment changed
| Observed pattern | Useful starting checks |
|---|---|
| Many directories fall together | Shared templates, access, migration, indexing configuration, demand and documented search updates |
| One directory or page group falls | Its release changes, content, internal links, redirects and canonicals |
| Similar impressions but fewer clicks | CTR, query mix and current result presentation |
| Mobile falls while desktop is steadier | Compare the same queries, then inspect mobile content, redirects, resources and usability |
| Only a third-party estimate falls | The tool's coverage, region, sampling and updates |
These are investigative directions, not automatic diagnoses. Small samples make average position and CTR volatile. Inspect actual queries and pages rather than assuming one aggregate represents every keyword.
3. Check access, indexing and explicit notices
Test the homepage, important business pages, affected sections and unaffected controls. Check HTTP responses, final redirects, rendered content, robots directives, canonicals and necessary resources. Align errors with server and CDN logs. Verify crawler identity before interpreting a request as genuine Googlebot traffic.
Open the Manual Actions report and Security Issues report. Follow any specific notice. If there is no notice, continue technical and content investigation; the traffic curve cannot substitute for the reported status.
Use URL Inspection to distinguish recorded indexing information from a current live test. A site: query can provide public-search clues but is not an exact index count or penalty certificate.
4. Compare content and concurrent search changes
Compare old and new versions of affected pages. Look for missing answers, a title that no longer matches the original task, factual problems or overlap with another page. Check whether useful editorial links and category entry points were removed. These questions are more actionable than assuming every decline means the article needs more words.
Add the Google Search Status Dashboard to the timeline, while treating coincident dates as clues rather than proof. Core-update changes do not follow a universally smooth curve, and manual actions do not have one diagnostic cliff shape. Follow Google's core-update analysis guidance when relevant.
Choose the repair based on evidence
- For a confirmed access or directive error, back up the affected configuration, fix it and verify both origin and public responses.
- If a rewrite removed essential answers, restore useful information and add reliable evidence and conditions.
- For an explicit manual-action or security notice, address the affected scope and use the corresponding review process.
- If the cause remains uncertain, preserve comparison groups and change records while gathering evidence instead of rewriting the entire site speculatively.
Record successful deployment, crawl accessibility, indexing changes, search performance and qualified inquiries separately. A technical repair does not establish recovered rankings. For the limits of penalty checks, see how to verify whether a site has a manual action.