“Thin content” is one of the most misused phrases in SEO. Every few months a client forwards me a screenshot from an AI tool flagging their 400-word service page as “thin” and asks if they need to pad it out to 1,500. The answer is almost always no, and the question itself shows how badly Google’s actual position has been mangled in translation.
Google has never defined thin content by word count. It has defined it, repeatedly and consistently since 2011, as pages with little or no added value to the user. That distinction matters because the fix for genuinely thin content is not “write more”, it is to add value, consolidate, or remove the page entirely. Padding a page that already answers the query is one of the fastest ways to make it worse.
This post walks through what Google means by thin content, how the policy has evolved from the original Panda update through the March 2024 core update, the specific patterns the search team flags, and a practical process for auditing and fixing thin pages on your own site.
What Google Actually Means by Thin Content
Google’s own definition, going back to its original quality guidelines and reiterated in the Search Essentials documentation, treats thin content as a quality problem rather than a length problem. The phrasing Google has used most consistently is “little or no added value.” A page can be 200 words and substantial; a page can be 3,000 words and thin.
The categories Google has explicitly named as thin, in the manual actions documentation and the spam policies, are narrower and more specific than most “thin content” warnings suggest:
- Automatically generated content without editorial value
- Thin affiliate pages, product pages that copy the manufacturer description and add nothing
- Scraped or rewritten content with light spinning
- Doorway pages, multiple near-duplicate pages targeting slight keyword variations
That list is from the manual actions page, and it is the closest thing to an official taxonomy. Notice that “short” is not on it. A concise page that fully satisfies the searcher’s intent is not thin under any version of Google’s policy, it is simply efficient.
A Brief Policy History: Panda to March 2024
Panda (February 2011)
The Panda update was Google’s first large-scale attempt to demote low-quality content sitewide. It introduced the idea that a single page of thin content could pull down rankings for an entire domain, content farms like Demand Media’s properties were the obvious targets, but plenty of legitimate sites got caught for things like duplicate product descriptions and shallow tag archives. Panda became part of the core algorithm in January 2016.
Helpful Content System (August 2022)
The Helpful Content Update launched in August 2022 and was rebranded as the Helpful Content System later that year. It introduced a sitewide classifier that demoted domains producing content “primarily for search engines rather than people.” Crucially, the HCS was sitewide, a portion of unhelpful content could drag down the rest of the domain, which made it functionally similar to Panda but with an explicit focus on content written for ranking rather than readers.
March 2024 Core Update and the Scaled Content Abuse Policy
The March 2024 core update did two notable things. First, Google folded the Helpful Content System into the core ranking signals, there is no separate HCS update anymore, but the underlying signals still apply. Second, Google introduced an explicit “scaled content abuse” spam policy targeting the production of pages at scale, regardless of whether they were generated by humans, AI, or some mix, when the primary purpose is manipulating rankings.
That policy update is the most significant shift since Panda. It removes the “but a human edited it” defense for low-value content at scale and explicitly puts programmatic and AI-assisted content into the same category as automatically generated spam if the output is not genuinely useful.
The Patterns Google Actually Flags
Mapping the official policies to real-world content, these are the patterns that get sites in trouble. Most thin content audits should be looking for these specifically, not for pages under some arbitrary word count.
Auto-generated and scaled content without editorial value
The scaled content abuse policy covers the obvious cases, AI-generated articles published with minimal review, Markov chain spam, content spun from competitor pages, but also a lot of borderline practices: city-page templates where only the location name changes, “best X in Y” pages where the X and Y are swapped into a template, and FAQ sections generated from keyword lists.
Doorway pages
Doorway pages are sets of similar pages built to rank for slight keyword variations that all funnel users to the same destination. A plumber with 47 near-identical pages for “plumber Toronto,” “plumber North York,” “plumber Etobicoke”, each saying essentially the same thing with the city swapped, is the textbook case. Local landing pages are not inherently doorways; they become doorways when they exist only for the keyword and offer nothing location-specific.
Thin affiliate pages
Affiliate content gets flagged when the page adds nothing the user could not get from the merchant, same product description, same specs, same images, with an affiliate link. The fix is original review content, comparisons, hands-on testing, photos you took yourself, or data the manufacturer does not publish.
Scraped and lightly rewritten content
Republishing or rewriting other people’s content with synonym swaps is the classic version. The modern version is feeding an article into an AI tool with “rewrite this” and publishing the output. Same problem, same policy.
Programmatic pages with no differentiated information
Programmatic SEO is not inherently thin, Zapier’s integration pages and G2’s review pages are programmatic and useful because each page has genuinely differentiated data. Programmatic pages become thin when the underlying data is sparse and the template pads it out with generic filler. If 95% of the page is the same across 10,000 URLs, those URLs are functionally duplicates.
Concise vs. Thin: The Distinction That Matters
This is where most internal SEO audits go wrong. A 200-word page that directly answers “what time does the LCBO close on Sundays” is not thin, it is exactly what the searcher wants. Padding it with the history of liquor regulation in Ontario makes the page worse, not better.
Conversely, a 2,400-word “ultimate guide” that restates what every other article on the topic says, with no original research, no specific examples, no opinion, and no expertise behind it, is thin by Google’s definition even though no audit tool will flag it on length.
The honest test for thin content is not word count. It is whether the page offers something the searcher cannot easily get from the top 10 results already ranking, original data, firsthand experience, a clearer explanation, a perspective from someone who has actually done the thing. If the answer is no, length will not save it.
How to Identify Thin Content on Your Own Site
A practical audit pulls from a few sources and looks for converging signals. No single metric identifies thin content on its own, a page with zero impressions might be brand new, and a page with high bounce rate might just be answering the query quickly.
1. Pull the GSC Pages report
In Search Console, open Performance, switch to the Pages tab, and set the date range to the last 12 months. Sort by impressions ascending. Pages with fewer than 50 impressions over a full year on an established site are candidates, either Google does not consider them relevant for any meaningful query, or no one is searching for what they cover.
2. Cross-reference with the Indexing report
Pages flagged “Crawled – currently not indexed” or “Discovered – currently not indexed” are often Google’s own thin-content judgment in action. The system has decided the page is not worth keeping in the index. Treat that list as a starting point for review.
3. Layer in engagement data
Pull GA4 engagement rate and average engagement time per page. A page that ranks but has a 15% engagement rate and a sub-10-second engagement time is sending strong signals that it is not satisfying searchers, that is a thin-content symptom even if the word count looks healthy.
4. Spot the templates
Crawl the site (Screaming Frog, Sitebulb, or Ahrefs Site Audit) and look for clusters of pages with near-identical title patterns, similar word counts, and matching internal link structures. Those clusters are usually programmatic or doorway-style pages, and the audit should evaluate them as a group.
How to Fix Thin Content
There are three real options. Pick one per page, do not try to fix everything by adding paragraphs.
- Consolidate. If you have five overlapping pages on the same topic, merge them into one strong page and 301 the others to it. This is the most common fix and almost always the right one when the underlying topic deserves coverage but the execution is fragmented.
- Rewrite with actual value. Keep the URL, but replace the content with something the page is currently failing to provide, original data, a worked example, screenshots, a specific recommendation. This is the right move for pages where the topic and URL are valuable but the content is not.
- Delete and 410. For pages with no traffic, no backlinks, no internal links of any consequence, and no topical value, return a 410 Gone. A 410 is a stronger signal than 404, it tells Google the page is intentionally removed, and Google drops it from the index faster.
Avoid the fourth option that gets recommended a lot online, noindexing thin pages and leaving them up. Noindex works for things like internal search results or filtered category pages that have a real user purpose. For genuinely thin pages, noindex just hides the problem from search while leaving the bad content discoverable. Better to fix or remove.
The Sitewide Risk
The most underrated thing about Google’s thin-content signals, and the part that has been true since Panda and got stronger with the Helpful Content System, is that they operate at the site level. A pile of thin pages does not just fail to rank on its own; it can suppress the ranking of your strong pages on the same domain.
This is why “let’s leave the old blog posts up, they aren’t hurting anything” is wrong. They might be. If you have 600 published posts and 400 of them are thin, the 200 good ones are competing with a sitewide quality signal that is being dragged down by the rest. Sites recovering from HCS-era demotions have almost always done significant pruning as part of the recovery.
The Practical Takeaway
Thin content is not a word count problem and never has been. It is a value problem, pages that do not give the searcher something worth the click. The patterns Google has flagged for 15 years are remarkably consistent: scaled production without editorial care, near-duplicates targeting keyword variants, affiliate or programmatic pages with nothing original, and content written to rank rather than to be read.
If you are auditing your own site, skip the word count column entirely for a first pass. Look at impressions, indexing status, engagement, and whether each page offers something the top results don’t. Then consolidate, rewrite, or delete, and accept that for a site carrying a lot of legacy content, deleting is often the highest-leverage move you can make.