First, confirm the status and whether the URL deserves the index
Open the page indexing report in Search Console (GSCGoogle's search performance and indexing tool.), click the "Crawled - currently not indexed" row and export the sample. Sort the examples by template before looking at any single URL: a status that clusters on tag archives, filtered listings or thin location pages is a template decision, while a scatter of strong articles points somewhere else. Note the "last crawled" date on each example, because a URL fetched a few days ago is in a different position from one that has sat in the list for months.
Then take two or three representative URLs through URL inspection and run the live test. Read the rendered HTML and the screenshot, not the source file: if the main text, product details or listings are missing from the render, the problem is rendering, not quality. Finally, ask the uncomfortable question for each template: should this URL be indexed at all? A sizeable share of these exports are pages nobody would miss, and the correct fix for those is to stop producing them rather than to coax Google into keeping them.
Diagnosis table
Diagnosis
| Symptom | Check | Likely cause | Fix |
|---|---|---|---|
| Real pages with little original substance sit in the export | Compare the page with similar URLs that are indexed; read it as a searcher would | Thin or low-value content Google chose not to keep | Rewrite with first-hand substance, merge into a stronger page, or retire it |
| The excluded URL closely mirrors another page on the site | Search the main phrase on the site; compare titles, headings and body copy of the pair | Near-duplicate pages competing with a stronger URL | Consolidate the pair and redirect, or differentiate the weaker one |
| Excluded pages are reachable only from the sitemap or deep pagination | Crawl inlink counts; click depth from the home page | Weak internal linking signalling low importance | Link the page from relevant, already-indexed pages with descriptive anchors |
| The live test render is missing the body text or listings | URL inspection live test: rendered HTML and screenshot | Main content only rendered by JavaScript that fails | Serve the primary content in the initial HTML, or fix the blocked or failing script |
| The export is dominated by filter, sort or tracking parameters | Group the sample by URL pattern | Faceted or parameter URLs that should not be indexed | Stop linking or generating them, or consolidate them on the clean URL |
| New URLs crawled only days ago, nothing else wrong | "Last crawled" date in the sample; publication date | Recently crawled and still being evaluated | Leave it, strengthen links, and re-check the URL later before acting |
Fix, verify and prevent each cause
Low-value content
Google's guidance on helpful, people-first content describes what it wants to keep: pages that show original information, real experience and a clear purpose for a reader. A page that rephrases what ranks elsewhere, or a template filled with a single paragraph of generic copy, gives the index nothing new.
Fix: decide per page whether it earns a rewrite. If it does, add what only your business can say: evidence, examples, specifics a practitioner would check. If it does not, fold it into a stronger page or remove it.
Verify: request indexing for a rewritten sample and watch whether those URLs move into the indexed list over the following weeks, while the retired ones drop out.
Prevent: set a content standard for each template before it ships, using the reviewer questions in the E-E-A-T and content quality guide.
A stronger near-copy exists
When two pages answer the same need, Google tends to keep one and set the other aside. This is different from URL variants of one page; here two genuinely separate pages say much the same thing.
Fix: pick the page with the better links and engagement, merge the useful parts of the weaker one into it, and redirect the old URL. Where both must exist, rewrite the weaker page around a distinct question.
Verify: after recrawl, the retained page is indexed and the redirected one leaves the report.
Prevent: check the keyword map before commissioning any page, the workflow described in keyword cannibalisation. If the duplicates are URL variants rather than separate articles, the sibling page on duplicates without a user-selected canonical is the better route.
Few internal links point to it
A URL that the site itself links from almost nowhere is a weak candidate: internal links are one of the ways a site tells Google which pages matter. Pages reachable only through the XML sitemap, or found at the bottom of long paginated archives, show up here often.
Fix: add contextual links from indexed pages on the same topic, and from hubs or category pages, using anchors that describe the destination.
Verify: a fresh crawl shows the page at a shallower depth with more inlinks, and URL inspection reports a newer crawl.
Prevent: make linking part of publishing, so no page goes live without links in and out, as the internal linking guide sets out.
Rendering leaves the page empty
If the content that makes a page worth keeping arrives only after JavaScript runs, a failed request, a blocked script or a timeout leaves Google judging a shell. Google's JavaScript basics documentation describes rendering as a separate stage after the fetch.
Fix: move the primary content into the server response, or repair what fails: unblock the script path in robots.txt, fix the API error, remove the dependency on user interaction.
Verify: the live test screenshot and rendered HTML now show the full content.
Prevent: add a rendered-HTML check to releases that touch templates, following the pre-release list in JavaScript SEO.
Filter and parameter URLs
Faceted navigation and tracking parameters can produce far more URLs than the site has real pages. Google crawls a sample, finds near-identical listings, and keeps few of them. Here the status is often correct and the problem is that the URLs exist.
Fix: decide which filter combinations deserve their own page, and stop generating crawlable links to the rest, or point them at the clean URL with a Canonical TagTells search engines the preferred URL..
Verify: the share of parameter URLs in the export falls over the following crawl cycles.
Prevent: run every new filter through the decision framework in the faceted navigation guide before it is launched.
Still being evaluated
Google does not promise to index every page it crawls, and a newly fetched URL can sit in this list before a decision. Acting too early, especially by repeatedly requesting indexing, wastes effort.
Fix: do nothing drastic; make sure the page is linked and complete.
Verify: re-inspect after a reasonable interval; if the status has not moved and nothing else in the table applies, treat it as a quality case.
Prevent: keep new pages out of the sitemap until they are finished, and read the indexing pipeline in crawling and indexing so the team knows which stage a URL is in.
Prevention: publishing only pages that deserve the index
The durable cure for this status is upstream. Every template should have an owner who can answer, before launch, why a searcher would choose this page over the ones already ranking. Thin templates should not ship with a promise to improve them later. New sections should be linked from the navigation or a hub on day one. Faceted and parameter URLs should be designed, not discovered in a report. Review the export monthly by template rather than by URL, and treat a rising count on one template as a product decision to revisit. Keep the indexed count and the count of pages you want indexed side by side in reporting; the gap between them is the real measure. When the status appears on pages that never received a crawl at all, that is the separate problem described in Discovered, currently not indexed.
Tag, author and date archives are among the commonest generated pages in this export, and the index rules for tags, archives and utility pages set a policy for each template before they ship.
A recipe site with a growing exclusion list
A recipe publisher sees 4,200 URLs in the crawled-but-not-indexed export. Grouping the sample by pattern shows 2,900 are ingredient tag pages with three recipes each, 800 are print-view URLs, and 500 are recipes. The tag and print URLs are thin or duplicate by design, so the team removes print links, sets the print view to point at the recipe, and noindexes tags with fewer than ten recipes. Of the 500 recipes, a live test shows 120 missing their method steps, which load from a script blocked in robots.txt. Unblocking it fixes those. The remaining recipes are rewritten or merged over the next quarter, and the export falls to about 600 URLs.
Listings, vacancies and events whose pages are stripped when they end often appear in this export as empty shells; the guide to expired listings, events and offers sets a rule for each state before pages expire.
Common mistakes
Requesting indexing again and again for the same weak page. Adding noindex or robots rules to "fix" the status, which hides the symptom without improving anything. Reading single URLs instead of grouping by template. Treating the status as a penalty rather than an editorial decision.
