Thin content is a page that gives a searcher little or nothing they could not get elsewhere: a near-empty template, a paragraph of generic text, a copy of another page with a few words swapped. Google crawls these pages and often decides not to index them, which shows up in Search Console as "Crawled - currently not indexed" or "Soft 404". Length is not the test. Value is.
When a large share of a site is thin, the effect goes beyond those pages, because Google evaluates quality at site level as well.
What counts as thin
Google's spam policies and quality guidance describe several patterns. In practice you will meet these:
| Type | Example | Typical Search Console status |
|---|---|---|
| Empty or near-empty templates | Category with zero products, profile page with a name only | Soft 404 |
| Boilerplate-heavy pages | 40 words of unique text inside 800 words of shared template | Crawled - currently not indexed |
| Doorway-style pages | "Plumber in [city]" repeated for 500 cities with the same text | Crawled - currently not indexed, or indexed and never ranking |
| Auto-generated archives | Tag pages with one post, date archives, author archives on a one-author blog | Crawled - currently not indexed, duplicates |
| Thin affiliate pages | Manufacturer descriptions and a buy button, nothing added | Crawled - currently not indexed |
| Scraped or syndicated copies | Content that exists word for word elsewhere | Duplicate, or not indexed |
| Scaled generic articles | Hundreds of posts that paraphrase the top results | Indexed at first, then dropped |
| Placeholder pages | "Coming soon", lorem ipsum, test posts | Soft 404 or not indexed |
Short pages that are not thin: a contact page, a glossary definition that answers the query, a product page with specific specs and original photos, a changelog entry. They serve a clear purpose completely.
How thin content affects indexing
Google's documentation on crawling and indexing is open about the fact that not everything crawled gets indexed and that content quality is a main factor. Three mechanisms are at work.
The per-page decision. After crawling, Google decides whether the page adds anything to the index. Thin pages lose that decision. The Page indexing report files them under "Crawled - currently not indexed", and Google's help text adds that there is no need to resubmit. Resubmitting the same page gets you the same decision.
Crawl demand. Google's crawl budget documentation lists low quality and duplicate content among the things that waste crawling and says that if Google spends time on URLs like these, it may decide the rest of the site is not worth looking at as often. On large sites this is how thin pages delay the indexing of good ones, which pile up in "Discovered - currently not indexed".
Site-level quality. Google's helpful content guidance says that removing unhelpful content can help the rankings of your other content. The reverse also holds.
Finding thin pages on your site
From Search Console
- Indexing > Pages > Crawled - currently not indexed. Export the examples. Group them by URL pattern. If 90 percent are
/tag/,/author/or?filter=URLs, you have found a template problem, not a writing problem. - Soft 404. Same exercise. These are pages Google considers empty.
- Performance > Pages, 12 month range. Export, then compare with the full list of URLs from your sitemap (the Sitemap URL Extractor gets you that list). Pages in the sitemap with zero impressions in a year are candidates.
From a crawl
Crawl the site with any desktop crawler and look at:
- Word count of the main content area, not the full page. Sort ascending. This finds empty templates quickly.
- Near-duplicate detection. Most crawlers can flag pages above a similarity threshold.
- Duplicate titles and H1s. Often a sign of pages produced from the same template with little unique data.
For a spot check of a single page, paste its main text into the Keyword Density Checker to get word count and see whether the most frequent phrases are topic words or template words.
By reading them
Numbers only shortlist. Open twenty pages from each suspicious group and ask the questions Google publishes for self-assessment:
- Does the page provide original information, reporting, research or analysis?
- Would you bookmark it or send it to a friend?
- After reading it, does someone need to search again to get what they came for?
- Was it written for people, or to have a page for a keyword?
If you wince, Google's systems probably do too.
Improve, merge, noindex or remove
Every thin page gets one of four treatments.
| Situation | Treatment | How |
|---|---|---|
| Topic matters, page is weak | Improve | Rewrite with real substance, then request indexing once |
| Several weak pages on one topic | Merge | Combine into one strong page, 301 the others to it |
| Useful on the site, worthless in search | Noindex | <meta name="robots" content="noindex">, remove from sitemap |
| No traffic, no links, no purpose | Remove | Return 404 or 410, remove internal links and sitemap entry |
Before deleting anything, check three things: organic clicks in the last 12 months, external links (Search Console > Links > Top linked pages) and conversions. A page with good links should be redirected to the closest relevant page, not deleted. Do not redirect everything to the homepage, since Google treats irrelevant redirects as soft 404s.
One caution on noindex: it keeps the page out of the index but Google still crawls it, though less often over time. For crawl waste on a very large scale, such as filter combinations, robots.txt is the better instrument. The differences are explained in noindex vs robots.txt vs canonical.
What "improve" means in practice
Adding 500 words of filler to reach a number does not change Google's decision. Add what the page lacks:
Product pages: your own description, real specifications, original photos, sizing or compatibility information, answers to the questions customers email you, genuine reviews.
Category pages: enough products to justify the page, a short useful introduction, filters, links to subcategories and buying guides. An empty category should return 404 or be noindexed until it has stock.
Location pages: information that is truly local: address, staff, opening hours, photos, prices for that branch, local projects, reviews of that branch. If you cannot say anything specific about a city, you do not need a page for it.
Articles: first-hand experience, original data or screenshots, specific steps with real menu names and code, a clear answer at the top. Cut the introduction.
Programmatic pages: each page needs unique data that answers a real query. If the only thing that changes between pages is the keyword in the H1, that is the doorway pattern. See programmatic SEO indexing at scale.
Template fixes that remove thin pages at the source
On most sites, the bulk of thin URLs comes from the CMS, not from writers.
- WordPress: noindex or disable tag archives with few posts, date archives, author archives on single-author sites, and attachment pages. See WordPress indexing problems.
- Ecommerce: handle filters and sorts deliberately. See faceted navigation and indexing.
- Internal search results: block crawling or noindex.
- Pagination: keep it crawlable, but do not expect page 37 to be indexed, and do not worry when it is not.
- User profiles and empty listings: only make them indexable once they pass a content threshold, and keep the rest out of the sitemap.
After the cleanup
- Update the XML sitemap so that it lists only the pages you kept and improved, with a truthful
lastmodon changed pages. - Request indexing for a few of the most important improved pages. Leave the rest to normal recrawling.
- Record the counts for "Crawled - currently not indexed" and total indexed pages, and check again every two weeks.
Set expectations properly. Individual improved pages can be re-evaluated within days or weeks of a recrawl. Site-level reassessment takes longer, and Google says improvements may take months to be reflected. The indexed page count may go down after pruning. That is the intended result: fewer indexed URLs, a larger share of which deserve their place.
If you track status with TurboIndex, the per-URL history from the URL Inspection API shows which improved pages moved from "Crawled - currently not indexed" to indexed, which tells you what kind of improvement Google responds to on your site.
What not to do
- Do not pad pages to a word count. There is no such threshold.
- Do not spin or paraphrase existing content into "unique" versions.
- Do not hide thin content behind tabs or accordions and expect a different result. The content is the same.
- Do not block thin pages in robots.txt and also noindex them. Google cannot see a noindex on a page it may not crawl.
- Do not resubmit unchanged pages through the inspection tool or any API. The decision was about the content.
- Do not mass delete without checking links and traffic. Pruning by word count alone removes pages that were doing fine.
Frequently asked questions
- Is there a minimum word count for Google to index a page?
- No. Google has said repeatedly that word count is not a ranking or indexing factor. A 150 word page that fully answers a question can be indexed and rank. A 2,000 word page that says nothing new can be skipped. Thin means low value, not short.
- Does thin content on some pages hurt the rest of my site?
- It can. Google's guidance on helpful content says that a site with a lot of unhelpful content may see its other content perform less well. A few weak pages are harmless. Thousands of them are a site quality problem.
- Should I noindex or delete thin pages?
- Delete (404 or 410) pages nobody needs. Redirect pages that have a clearly better equivalent. Use noindex for pages that are useful to visitors on the site but have no value in search, such as internal search results or thin tag archives.
- Is AI-generated content thin content?
- Not automatically. Google's position is that it rewards helpful content regardless of how it is produced. Generating many pages with little value for users, by any method, falls under its scaled content abuse spam policy.