You have a /category/ai-tools/ page. It is auto-generated: an <h1> at the top, 20 article cards below, a breadcrumb, a footer. Google never indexes it, or if it does, it ranks for nothing. Open the Page indexing report in Search Console and the URL sits under Crawled - currently not indexed or Discovered - currently not indexed. Users who land there bounce in a few seconds because there is no orientation, just a wall of cards with no signal of which to read first.
Fastest fix: a category page that is only a list is navigation, not content, and Google indexes content. Add three things above the card grid, in this order: a 200-400 word editorial intro, a curated “Start here” with one-line editor notes, and a short opinionated take on the topic. Then put each paginated page on a self-referencing canonical (not a canonical pointing back to page 1), make sure 5 or more child articles link back to the category, and only then click Request Indexing in the URL Inspection tool. That sequence turns “site UI” into a topic page worth ranking. The rest of this guide is the longer version with diagnosis and verification.
Which bucket are you in
Run these four checks before you change anything. They map directly to the fix you need.
| Check | What you see | Most likely cause |
|---|---|---|
| View source of the category page | Everything between <h1> and <footer> is article cards | No editorial layer (cause 1 or 2) |
| Look at the slug | /category/posts/, /category/tutorials/, /category/guides/ | Generic label, no SEO target (cause 3) |
| Inspect the URL in Search Console | Status is Crawled - currently not indexed or Soft 404 | Thin content; Google saw it and declined |
grep -r "/category/<slug>/" src/content | Zero hits from child articles | Orphaned category (cause 5) |
View source, search for robots | <meta name="robots" content="noindex"> | Self-inflicted noindex (cause 6) |
Discovered - currently not indexed means Google knows the URL exists but has not bothered to crawl it; that is usually weak internal linking. Crawled - currently not indexed means Google crawled it and decided it was not worth keeping; that is a content-quality problem. Soft 404 means the page is so thin Google treats it like a missing page. All three are fixed below.
Common causes
Ordered by hit rate, highest first.
1. Template auto-generates category pages with no intro field
Your Astro / Next / Hugo template builds a list page from frontmatter. There is no field for an editorial intro and no place to write one. The category page is mechanically generated and editorially silent.
How to spot it: View source of any category page. If everything between <h1> and <footer> is article cards, no editorial layer exists.
2. Editors fill the categories but never write the category page itself
Each category has 30 or more articles inside, but the category page itself has never been “written” by anyone. It exists, but no one composed it.
How to spot it: Check git log on the file that renders the category page. If only the template was edited and there are no per-category content edits, no one curated this page.
3. Category labels are generic instead of specific
/category/tutorials/ or /category/posts/ have no SEO target. “Tutorials about what?” Specific category names like /category/openai-api-tutorials/ have ranking potential; generic ones do not.
How to spot it: List your category slugs. If they are abstract types (posts, guides, stuff), they are not SEO-targeted. If they are topics (openai-api, claude-best-practices), they could rank.
4. Pagination dilutes the category’s main page
/category/ai-tools/ has 30 articles spread across pages 1 to 3. The trap is canonicalizing page 2 and page 3 back to page 1: that tells Google the deeper pages are duplicates, so it drops them and the articles that only appear on later pages become invisible. As of June 2026, Google’s pagination guidance is the opposite of the old rel="prev"/rel="next" approach (which Google stopped using in 2019 and now ignores): treat every paginated page as a normal indexable page with a self-referencing canonical, never noindex or robots-block the deeper pages, and link the pages to each other with plain anchor links so the crawler can walk the series.
How to spot it: View source of /category/ai-tools/page/2/. If its <link rel="canonical"> points to page 1 instead of itself, your pagination is hiding content.
5. No internal links into the category page from related articles
Articles in the category do not link back to the category. The category page floats unsupported in the link graph, so PageRank-like signals never accumulate, and Google often leaves it under Discovered - currently not indexed because nothing tells it the page matters.
How to spot it: For five random articles in a category, grep for the category URL. If none reference it, the category is orphaned by its own children.
6. Category page is noindex “for SEO hygiene”
Someone read advice to noindex category and tag pages to avoid duplicate content and to save crawl budget. You followed it. Now Google cannot even consider ranking the category. That hygiene guidance is sound for genuinely thin, overlapping tag archives, but it is exactly wrong for a page you have actually written editorial content for.
How to spot it: <meta name="robots" content="noindex"> on the category page. Remove it once you have added a real intro and curation.
Shortest path to fix
Ordered by ROI. Steps 1 to 3 turn a navigation page into a hub article.
Step 1: Write a 200-400 word editorial intro for each category
In your CMS or frontmatter, add an intro field:
---
slug: openai-api-tutorials
title: "OpenAI API tutorials"
intro: |
We've shipped 28 OpenAI API tutorials since the API launched.
This page is your map: where to start if you're new, what to read
if you're optimizing prompt-cache hit rates, and how to avoid the
five most common billing surprises.
[200-400 words explaining: scope of the category, who the reader is,
why this topic matters, what makes our coverage different.]
---Have the template render the intro before the card grid. Aim for 300 words of specific, original prose; generic boilerplate (“Welcome to our tutorials section”) does not move a page out of Crawled - currently not indexed.
Step 2: Add a curated “Start here” with editor notes
Do not just list 30 cards in publish-date order. Curate:
## Start here (5 essentials)
1. **[API quickstart in 10 minutes](/articles/...)** — read this first if you've never made an API call.
2. **[Prompt-caching for production](/articles/...)** — most teams underestimate the cost savings; this is where 80% of optimization comes from.
3. **[Cost monitoring + alerts](/articles/...)** — set this up *before* you hit production, not after.
[2 more]The editor’s note, the one-line take, is the key signal: “this writer has an opinion.” Generic lists do not have that.
Step 3: Add a “topic take” — your point of view
What is your distinct angle on this category? Two paragraphs:
## Our take on OpenAI API
Most teams treat the API as "ChatGPT with code." We disagree.
The interesting work isn't in completions — it's in prompt-cache
strategy, batch API for non-interactive jobs, and structured outputs
for replacing webhook-parsing code.
If your usage looks like "send a prompt, get a response, repeat,"
you're paying 3x what you should and missing the patterns that make
this API actually transformative.This is what makes the page rank for “openai api” beyond just navigation. Opinions are content, and an original opinion is the single hardest thing for a machine-generated page to fake.
Step 4: Fix pagination so deeper pages stay indexable
If the category paginates, do not canonicalize page 2 and beyond to page 1. Give each page a self-referencing canonical and keep it indexable:
<!-- on /category/ai-tools/page/2/ -->
<link rel="canonical" href="https://example.com/category/ai-tools/page/2/" />Add numbered links to every page in the series so Googlebot can crawl the whole set, and do not put <meta name="robots" content="noindex"> or a robots.txt block on the deeper pages: if Google cannot crawl page 2, it cannot reach the articles that only live there. rel="prev" and rel="next" are no longer used by Google, so do not rely on them; plain anchor links to each page are what matters now.
Step 5: Link related categories to each other
At the bottom of the page:
## Related categories
- [Claude API tutorials](/category/claude-api/) — same patterns, different provider
- [Prompt engineering](/category/prompt-engineering/) — the prompts that go into these APIs
- [LLM cost optimization](/category/llm-costs/) — cross-provider cost workCategory-to-category links build a topical graph Google can read.
Step 6: Verify backlinks from child articles
Each article in the category should link back to the category. Add to the article template:
<CategoryLink slug="openai-api-tutorials">
More OpenAI API tutorials →
</CategoryLink>Or in body content where it is natural:
This is one of our [OpenAI API tutorials](/category/openai-api-tutorials/) — see the rest for the broader topic.Five articles linking back to the category is enough to pull it out of Discovered - currently not indexed. Zero leaves it orphaned.
Step 7: Remove noindex and request a recrawl
Once the intro and curation are live, remove any <meta name="robots" content="noindex"> from the category template. Then in Search Console, open the URL Inspection tool (the search bar at the top), paste the full category URL, click Test Live URL to confirm the page returns URL is available to Google with no noindex, and click Request Indexing. As of June 2026, Search Console caps requests at roughly 10 to 12 URLs per day per property, and the button greys out for 24 hours once you hit the limit, so prioritize your strongest categories. Without a request, Google may take weeks to revisit on its own.
How to confirm it is fixed
- Live check: In URL Inspection, click Test Live URL. You want
URL is available to Google,Indexing allowed? Yes, and the rendered HTML should show your intro and “Start here” text, not just cards. - Index status: A few days after requesting indexing, re-inspect the URL. The status should move from
Crawled - currently not indexedtoURL is on Google. - Coverage report: In the Page indexing report, the category URL should leave the “Not indexed” group. Filter by the
/category/path to track all of them at once. - Live SERP: Search
site:example.com/category/ai-tools/. If it returns the page, it is indexed. If the snippet shows your intro text, your editorial layer is what Google is reading.
If a category is still not indexed after two weeks and a recrawl, the content is the problem, not the technical setup. Expand the intro, sharpen the take, and add more inbound links rather than re-clicking Request Indexing, which does nothing without real changes.
Prevention
- Treat every category page as a mini-article: 300-plus word intro, curated “Start here,” topical take.
- Add a CI check that fails if a category page has no
introfield or the intro is under 200 words. - Use specific topic slugs, not generic types (
openai-api, nottutorials). - Make the article template link back to its category by default, and aim for five or more inbound links per category page.
- Give paginated pages self-referencing canonicals; never canonicalize page 2-plus to page 1.
- Do not
noindexcategory pages that have real editorial content; that defeats the purpose. Reservenoindexfor genuinely thin tag archives. - Quarterly, review which categories rank. Weak ones need either better content or merging into a stronger category.
FAQ
Should category pages be indexed or noindexed?
Index the ones you have actually written: a real intro, curated picks, and a point of view. noindex only the thin, overlapping archives (most tag pages, date archives, internal search results) that add nothing a real page does not already cover. A blanket “noindex all category pages” rule is outdated and costs you rankable hub pages.
Why is my category page “Crawled - currently not indexed”?
Google crawled it and decided it was not worth keeping, which is almost always a content-quality signal: the page is just a card list with no original prose. Add the editorial intro, take, and curation from Steps 1 to 3, then request indexing. Compare with Discovered - currently not indexed, which means Google has not even crawled it yet and is usually fixed by adding internal links (Step 6).
Do I still need rel=prev and rel=next for pagination?
No. Google stopped using those signals years ago and ignores them as of June 2026. They are not harmful, but they do nothing. What works now is a self-referencing canonical on each paginated page plus plain anchor links between the pages.
How long until Google re-indexes after I fix the page?
After clicking Request Indexing, most URLs are recrawled within hours to a few days, but reaching URL is on Google can take one to two weeks. Do not spam Request Indexing; the daily cap is about 10 to 12 URLs per property, and repeated requests on an unchanged page change nothing.
My category page is indexed but ranks for nothing. What now?
Indexed but not ranking means the page has no clear keyword target or no authority. Check the slug is a real topic (Step 3 cause), make the intro and <title> target one specific phrase, and build inbound links from child articles and related categories (Steps 5 and 6). Ranking follows topical relevance plus internal links, not just being indexable.
Related
Tags: #Content ops #Site quality #Site audit #Troubleshooting #Category page