Start a search investigation by locating the failing stage. A traffic decline is an outcome, not a diagnosis. The framework below moves from whether a page can be found and processed to whether it meets a user’s need and produces a useful visit.
Discovery → Crawling → Rendering → Indexing → Canonicalization → Ranking eligibility → Query relevance → Search appearance → Traffic → Conversion
This is a diagnostic ordering, not a literal specification of a search engine’s processing pipeline. Canonicalization and indexing interact, and feedback can move an investigation back to an earlier stage.
1. Discovery: is the URL reachable through useful paths?
Is the exact intended URL linked from accessible pages? Does a sitemap list the canonical destination? Is the page orphaned, reachable only through a form or linked through a broken path? Inspect the links themselves before assuming the crawler knows the page.
2. Crawling: what response can be fetched?
Does the URL respond reliably? Are robots.txt rules or access controls relevant? Is the response a page, redirect or error? Compare current responses with dated logs, remembering that logs show requests rather than indexing outcomes.
3. Rendering: is the essential material available?
Does the content depend on JavaScript, login state or unavailable resources? Are links and headings present in the processed page? A successful view in your own browser is only one observation.
4. Indexing: what state is reported?
Is an indexing restriction intentional? What does the URL-level report say, and when was the page last processed? Distinguish a live test from indexed-state information. See Google’s URL Inspection documentation.
5. Canonicalization: which representative is selected?
Do the canonical declaration, redirects, internal links and sitemap agree? Does the selected canonical differ from the intended URL? If two pages are genuinely different, check whether their actual content communicates that difference.
6. Ranking eligibility: is there a known restriction?
Review applicable technical requirements, manual-action information and security reports. Do not infer a penalty from a lower chart alone. Eligibility permits participation; it does not promise a position or search feature.
7. Query relevance: does the page solve the task?
Which query group is affected? Does the page provide the information or action those readers need? Compare intended purpose with observed results using the page-type mapping method.
8. Search appearance: what can users see?
Check titles, snippets and relevant result features under recorded conditions. Appearance can vary by query and context. A configured description does not guarantee the same snippet for every search.
9. Traffic: which measure changed?
Separate impressions, clicks and analytics sessions. Segment comparable populations and check tracking before attributing a change to search behavior. Use the traffic diagnosis guide.
10. Conversion: did the visit support the intended action?
Define the action and verify its instrumentation. A conversion decline can reflect usability, availability, audience mix or measurement. The search result may have delivered the right visitor to a destination that no longer works.
Record a diagnosis that someone can verify
Save an example URL, the observed symptom, test date, expected behavior, evidence and next action. Compare affected and unaffected pages. Fix a confirmed mechanism, retest that mechanism and then monitor the broader outcome.
For the technical distinctions, read Crawled, Rendered, Indexed. For the strength of the conclusion, use SEO Research Methods.