Thin content is a page that gives the person reading it too little useful, original, or trustworthy information to actually do what it promised. It has nothing to do with word count on its own. A short page can do its job well. A long page can still leave the reader empty handed, and that is the page Google treats as thin.
If you have ever heard "thin content seo" thrown around as a reason a page stopped ranking, it is worth slowing down and getting specific about what that actually means, because the term gets used loosely. Google does not publish a formula that scores every page for thinness. It does use the phrase "thin content with little or no added value" as the name of one specific manual action, and it describes several spam practices, like thin affiliation, where original value is missing. But most pages that quietly lose visibility were never hit with a manual action at all. They just failed to be useful enough for what they were trying to do, and Google's systems noticed.
This piece walks through what google thin content actually refers to, the signals that separate a page that works from one that doesn't, and what to check before you assume padding is the answer. It is not about how long an article should be. It is about what a page owes the person who landed on it.
What thin content actually means
Start with the plain version: a page is thin when it does not give its intended visitor enough of what they came for. Comprehensiveness is the goal, and comprehensiveness means answering what that one visitor reasonably needs, with enough original substance behind it to be worth trusting. It does not mean covering every tangential question or hitting some word count. A page answering "what are your store hours" can be four sentences and complete. A 3,000 word "ultimate guide" that repeats things everyone already knows, without ever answering the reader's actual decision, is the thin one, even though it is much longer.
That distinction matters because a lot of teams still treat length as a proxy for quality. It is an easy number to hit, so it is tempting to chase it. But Google has been direct about this: it has no preferred word count. There is no published minimum that makes a page safe, and no evidence that padding a short page with more paragraphs fixes the underlying problem. The fix is not more words. It is more of what the page was supposed to deliver in the first place.
It also helps to separate a few things that get lumped together under "thin content." There is genuinely low quality content, which is a page that just fails to deliver. There are specific spam policy violations, like scraped material or doorway pages. There is a named manual action Google can apply to a site. There is duplication, where a page loses to its own near-duplicates. And there is a page that ranks poorly for reasons that have nothing to do with quality at all. These are different problems with different fixes, and calling all of them "thin content" makes it harder to diagnose which one you actually have.
The quality signals that actually matter
Google's people-first content guidance does not hand out a numeric checklist. Instead it poses a set of questions meant to help a publisher self-assess, and a few of them get right at what makes a page thin:
- Does the page provide original information, reporting, or analysis, rather than just repeating what is already out there?
- Does it give a complete, substantial account of the topic it claims to cover?
- Does it offer insight beyond the obvious, something a reader could not have guessed on their own?
- When it draws on other sources, does it add real value on top of them, instead of paraphrasing?
- Would the reader leave satisfied, in a way that makes the page worth sharing or recommending?
- Is the information presented in a way that feels trustworthy, with visible sourcing and evidence that whoever wrote it actually knows the subject?
You can translate those questions into things you would actually notice reading a page. A headline promises an answer the body never delivers. A section paraphrases the same three sources everyone else used, without adding a fresh angle. A claim that matters to the reader's decision is left unsupported. A location or product name gets swapped into an otherwise identical template. The reader finishes the page and has to go back to the search results to actually answer their question. None of these are fixed by better formatting or an extra paragraph, because the gap they describe is about substance, not structure.
Google's Search Quality Evaluator Guidelines, last updated in September 2025, draw a useful line between pages rated Low, which fail to achieve their purpose because something important is missing, and pages rated Lowest, which can include material that is actively spammy or untrustworthy. These are illustrations of how Google thinks about quality, not a public copy of its ranking algorithm, and an individual rater's judgment on one page does not directly move that page in search results. Still, the distinction is a good gut check: is this page missing something important, or is it actively bad.
E-E-A-T comes up constantly in this conversation, and it is worth being precise about it too. Google has said E-E-A-T is not one specific ranking factor. It is a set of qualities, experience, expertise, authoritativeness, and trust, that show up across many signals, and how much of it a page needs depends on the topic. A recipe blog and a page about a medical treatment do not carry the same bar. Slapping an author bio on a thin page does not turn it into a trustworthy one. The bio only matters if the substance behind it holds up.
Patterns that actually make a page thin
A few recurring patterns are worth naming, because they are more specific and more useful than the general label:
An empty answer dressed up as an article opens by restating the question, then offers broad statements without ever landing on a usable answer. The reader finishes and still does not know what to do.
Repackaged sources reword material that is already available elsewhere, without adding reporting, a worked example, or any interpretation of your own. It reads like a summary of someone else's page, because that is what it is.
Copied merchant descriptions show up on affiliate pages that reproduce a seller's product copy or reviews verbatim. Google calls this thin affiliation specifically. Affiliate links themselves are not the problem. A page with genuine testing, original ratings, or a real comparison behind its affiliate links is doing something different than one that just repeats the merchant's own language.
Slightly altered duplicates swap a city name, a product name, or a phrase across dozens of otherwise identical pages. Depending on how it is implemented, this can raise duplication or doorway concerns, though not every templated page automatically violates a policy.
Mass produced pages assemble content from search results or other sites, sometimes stuffed with keywords that don't read as coherent to an actual person. Google's scaled content abuse policy targets pages made in volume primarily to manipulate rankings, regardless of whether AI or a person produced them.
A doorway page targets one of many near-identical queries and mostly funnels the visitor somewhere else, to a destination that actually serves what they came for. The page itself never does the job.
None of this means every short page, every templated page, or every piece of AI assisted writing is automatically thin. The use of AI by itself does not establish that content is low value. Google's stated concern is content made in bulk primarily to game search rankings, not the tool used to produce it. The test is still the same one running through all of these patterns: does this specific page add something real for the person reading it, or does it just occupy space where an answer should be.
Why a thin page loses visibility, or never shows up at all
It helps to keep three separate outcomes apart, because they get collapsed into "my page got deindexed" way too often.
The first is poor ranking. A page can be fully indexed and still get almost no visibility, because Google's automated ranking systems are built to prioritize content it judges genuinely helpful. That is a ranking decision, not proof of a violation.
The second is never being indexed at all. Google can crawl a URL and choose not to add it to the index, and it says this can happen when there is insufficient value or insufficient user demand for that content, separate from any technical issue. A page that was never indexed was never technically "deindexed," even though it feels the same to the person watching it in Search Console.
The third is an actual removal or omission tied to a spam policy or a manual action. Google's spam policies can lead to lower ranking or complete omission from results, and the Manual Actions report includes a specifically named action called "Thin content with little or no added value." A manual action can cover a handful of pages or an entire site. This is the only one of the three that comes with a formal flag you can check directly, and it is worth confirming before assuming your problem is this one.
Google's own Search Essentials page states plainly that meeting its guidelines does not guarantee crawling, indexing, or serving. Submitting a sitemap or requesting indexing is a request, not an entitlement. That is a useful thing to sit with before reaching for the word "deindexed" as a catch-all, since it implies something was pulled out that was already there, when a lot of the time the page simply never made it in.
What Search Console can and cannot tell you
The Manual Actions report is the one place that gives you a direct answer. If Google has issued the thin content manual action, or any other manual action, it will be listed there with the affected pages. If nothing shows up, don't treat a ranking or indexing problem as a manual penalty, because it isn't one.
The Page Indexing report shows other statuses that get misread constantly. "Crawled, currently not indexed" means Google visited the page and chose not to index it yet, and Google explicitly says this status does not mean the page is judged thin. It might be indexed later without you doing anything. "Discovered, currently not indexed" means Google found the URL but has not crawled it at all, often just to avoid overloading the site with requests, so there is no content judgment happening here whatsoever.
A page marked as a duplicate without a user selected canonical means Google picked a different URL to serve as the canonical version, which can be entirely expected behavior rather than a quality signal. A soft 404 means Google believes a page that returns a successful response is actually a not found page in disguise, and that one is worth checking with fresh eyes, since it points at something structural rather than something you'd catch by proofreading.
The URL Inspection tool's live test checks the current state of a URL directly, including what Google can fetch and how the page renders. A successful live test confirms Google can access and read the page. It does not guarantee indexing, and requesting indexing through that tool does not guarantee it either. None of these signals hand you a thin content verdict on their own. They tell you what happened technically, and you still have to bring the judgment about whether the page is actually doing its job.
Making a page genuinely more complete
The fix is not a target length. It is a page purpose test you can run against any page you're worried about: what question or decision brought this visitor here, and what does the page need to supply for them to leave satisfied? Which parts of it just repeat what is already sitting elsewhere online? What can you actually add, whether that's firsthand experience, an original example, a clearer explanation, or evidence a reader can check for themselves?
From there the fixes are specific rather than generic. Answer the question the headline promises, directly, instead of circling it. Add the qualification or context that was missing. Fix anything factually wrong. Separate what you're contributing yourself from what you're drawing from other sources, so a reader can tell the difference. Replace copied descriptions with your own evaluation. Cut the filler a template left behind. Make the expertise behind the page visible when the topic calls for it, rather than assumed.
If you find several pages on your own site saying basically the same thing, that's a duplication and canonical problem, not a case for adding more words to each one. And if a page is genuinely obsolete, the right move is removing or replacing it, not padding it out to keep it alive.
Once a confirmed manual action is involved, the fix needs to cover every affected page, not a sample of them, and any other listed manual actions need resolving too. Only after that is done does Request Review in the Manual Actions report make sense, with a clear account of the issue and what changed. That review path exists specifically for manual actions. It is not the move for an ordinary "crawled, not indexed" page that has no manual action attached at all.
There is no documented threshold where a page crosses over into "comprehensive enough," no published minimum word count, and no guaranteed timeline for reindexing after you fix something. The honest goal is a page that is genuinely more useful to the person reading it, checked against what Search Console is actually telling you, not a promise that traffic comes back on schedule.
Teams producing content at real volume run into this pattern often: writing at pace and never circling back to check whether older pages still earn their place is how a backlog of thin pages builds up quietly. Content Studio inside DeepSmith is built to research and structure each article against a purpose before it's written, and Content Map shows where your published pages sit against the coverage your topic actually needs, so a gap in substance is something you catch on purpose rather than something you find months later in Search Console.
The question worth returning to, on any page you're unsure about, is simple: did it actually finish the job it promised the person who landed on it? If the answer is no, that's the problem to fix, regardless of how long or short the page already is.
