B2B SaaS · product-adoption software

Diagnosing 30,000 Not Indexed Pages

Challenge: Search Console reported close to 30,000 pages as not indexed against a sitemap of only 777 URLs, and it had been escalated as an emergency.Results: 7,127 of 9,200 were pagination parameters already blocked and decaying on their own. Every inspected content page was indexed. No action needed.Timeline: 16 months of Search Console data reviewed

The challenge

The client opened Search Console, saw close to 30,000 pages reported as not indexed, and escalated it as an emergency. The number is genuinely alarming on a site with a few hundred real pages, and the instinct is to start fixing.

  • Around 30,000 not-indexed URLs against a sitemap of only 777.
  • The obvious response was the wrong one. Requesting indexing, rewriting content or loosening robots rules would have cost weeks and achieved nothing.
  • Nobody had established what those URLs actually were. The count was being treated as a diagnosis rather than as a symptom needing one.

Approach

Enumerate from the source of truth, not from the alarming number. We pulled every distinct URL Google had surfaced across 16 months rather than reasoning from the Search Console summary tile.

Classify the whole set by URL pattern. This is where the answer was: the overwhelming majority shared a query-string signature.

Test the real pages independently. Every inspected content page came back "Submitted and indexed", which is the control that proves the problem is not where the client feared it was.

Track the decay curve rather than the snapshot. Monthly impressions for the suspect class were followed over the full period to show the trend was already resolving itself under an existing rule.

Say so plainly. The deliverable's conclusion was that no action was needed, with the evidence laid out so the client could verify it rather than take reassurance on trust.

Results

  • 7,127 of 9,200 surfaced URLs (77%) were pagination parameter combinations, not real pages, already covered by a live Disallow rule.
  • Monthly impressions for that class fell from 6,339 to 24 over 16 months, decaying on their own as Google dropped them. 24 clicks in total across the entire period.
  • 100% of inspected content pages returned "Submitted and indexed". The real site was fine.
  • A benign verdict delivered with the evidence attached, and an unrelated genuine issue flagged separately for the client to decide on.

Key insight

A frightening number in Search Console is a question, not a finding. Classifying the URLs before acting turned a 30,000-page emergency into a known, self-resolving artefact of one pagination rule, and saved the client from a remediation programme aimed at pages that were never meant to be indexed. The useful deliverable was the evidence that nothing was wrong, plus the one genuine issue that would otherwise have stayed buried under the noise.