A structured way to separate urgent search problems from worthwhile improvements and low-value distractions.
Triage starts with the symptom, not a shopping list
Folio detail
An SEO review becomes unhelpful when every observation is presented as equally urgent. Start by writing down the symptom in plain terms: the whole site has lost search visits, one service page cannot be found, enquiries have fallen while visits appear steady, or a new page is not indexed. Record when it began and what changed around that date. That creates a testable incident rather than a vague request to ‘improve SEO’. A site-wide fall after a migration deserves a different investigation from one page sitting below stronger competitors.
Rule out access and indexing failures first
Folio detail
Check whether Google can reach the affected URL, whether the URL returns a successful status, and whether robots instructions, a noindex directive, canonical choice or redirect sends a conflicting signal. Use URL Inspection and the Page indexing report in Search Console where access is available. These checks come first because polishing copy cannot help a page that Google cannot crawl or has reasonably chosen not to index. Keep evidence URL-specific: status, declared canonical, Google-selected canonical and last crawl are more useful than a screenshot of an overall ‘SEO score’.
Separate evidence from plausible theories
Folio detail
Label each finding as confirmed, supported or untested. ‘Page has a noindex directive’ is confirmed. ‘Important page has no internal links in the crawl’ is supported by observable structure. ‘Google dislikes the design’ is an untested theory unless evidence narrows it. This discipline prevents expensive work being commissioned on confidence alone. For each theory, name the cheapest discriminating check: inspect rendered text, compare affected page groups, review change history, or test whether branded and non-branded queries moved differently.
Then test relevance and presentation
Folio detail
If access is sound, compare the page with the searches it is meant to answer. Does the title identify the actual service or subject? Does the main copy answer the likely decision before wandering into company background? Are important details present in readable text rather than only inside an image? Search Console query and page data can reveal whether impressions exist but clicks are weak, or whether the page rarely appears at all. The former suggests presentation or intent mismatch; the latter points towards relevance, internal discovery, competition or indexing.
Turn the queue into reversible actions
Folio detail
Fix high-impact blockers with strong evidence first, then measure before broadening the change. Preserve the original title, copy or redirect map so an edit can be reversed. Assign one owner and a verification step to every action. A useful ticket says ‘remove accidental noindex from these three listed service URLs, request recrawl, and confirm indexability’, not ‘do technical SEO’. Lower-confidence content changes should usually be tested on a small, coherent page group rather than rolled across the site at once.
Practical matrix
SEO triage matrix
Place each finding in one row and prioritise evidence-backed blockers before speculative enhancements.
P1: site unavailableEvidence: affected URLs time out or return 5xx responses. Action: restore stable access, then verify representative URLs and server logs before any content work.
P1: indexing blockedEvidence: an important URL has noindex, robots blocking, an unintended redirect or a conflicting canonical. Action: correct the specific signal and inspect the live URL again.
P2: migration mismatchEvidence: valuable old URLs now return 404, redirect to irrelevant destinations or disappeared after a release. Action: reconcile old-to-new URL mapping and test each redirect chain.
P2: discoverability gapEvidence: an important page is absent from navigation, contextual links and sitemaps. Action: add relevant crawlable links from pages where a visitor would expect the next step.
P3: intent mismatchEvidence: the page is indexed but its purpose, title and main answer do not match the queries or decision being served. Action: reshape one page around that specific need.
P3: weak search snippetEvidence: meaningful impressions with poor clicks, while the title is vague, duplicated or misleading. Action: write an accurate, distinctive title and description; compare later periods cautiously.
P4: enhancementEvidence: no blocking defect; an improvement such as clearer examples, structured navigation or stronger supporting detail may help. Action: schedule after faults and define the expected user benefit.
Hold: unsupported theoryEvidence: opinion or tool score without a URL-level cause. Action: name a check that could disprove the theory before allocating implementation time.