Thin content: the threshold nobody publishes
What “thin” actually means
Every SEO forum has a thread asking how many words a page needs to avoid being “thin content.” People want a number. 300 words. 500. 1,000. Pick one, hit it, move on.
That number doesn’t exist, and it never has. Thin content isn’t a word count problem. It’s a value problem. A page is thin when it doesn’t give a visitor anything they couldn’t get faster somewhere else, including from a search results page that already answered the question in the snippet.
I’ve published pages under 200 words that rank and hold, because they answer one specific question directly and there’s nothing else to say. I’ve also built pages past 2,000 words that are pure padding: reworded intros, a bulleted list that repeats the H2, a conclusion that restates the title. Length told me nothing about which one was doing its job. Only intent match and follow-through did.
The word count myth won’t die
The reason word count sticks around as a proxy is that it’s easy to measure and hard to argue with. “Add more words” is an instruction you can hand to a junior writer or an AI tool with zero judgment involved. “Add more value” requires someone to actually know the topic.
I’ve watched teams pad pages to hit an internal minimum, and the padding is always the same shape: a paragraph of throat-clearing before the actual answer, a “things to consider” section that lists the obvious, and a closing paragraph that says nothing new. None of that helps the page. It just makes the useful part harder to find, and it tells anyone reading closely that the writer didn’t have enough to say and filled space instead.
If you’re running a content operation at any scale, the temptation to set a floor is real because it’s the only lever that scales without editorial judgment on every single piece. I get it. But the floor produces exactly the kind of filler that makes a page thin in the first place. You end up optimizing for the metric instead of the outcome it was supposed to represent.
Where thin pages actually come from
In my own portfolio, thin pages show up from three sources, almost never from “the writer didn’t try hard enough.”
Programmatic pages that ran out of unique inputs. Location pages, “best X for Y” pages, product variant pages. The template is good, but past a certain number of instances there isn’t enough real difference between one page and the next. The 40th city page and the 41st city page say the same thing with the city name swapped.
Old pages that never got updated after the topic moved on. A page written for a tool or a rule that changed. The words are all still there, the page just stopped being useful, and nobody went back to fix it because it wasn’t causing visible pain.
Pages built to catch a keyword, not to serve a reader. Someone pulled a keyword list, saw a term with volume, and wrote a page around the term instead of around a real question a real person has. These read like they exist for the search engine, because they do.
None of these are word count failures. They’re all failures of “does this page do a job that a person needed done.”
Why one thin page can cost you more than one page
The mistake I made early on was treating each thin page as its own isolated risk. Worst case, it just doesn’t rank, right? No harm beyond wasted effort.
That’s not how it plays out in practice. Quality signals aggregate at the site or section level, not purely per URL. A site that’s mostly solid but has a large chunk of low-value, near-duplicate, or filler pages doesn’t just fail to get credit for the good pages faster, it can make the whole domain look less trustworthy for topics it’s actually strong on. I’ve seen sections of a site under-perform relative to their content quality specifically because they sat next to a pile of thin pages that dragged the average down.
Crawl budget is the other cost people underrate. If a crawler spends its visit budget on 400 near-duplicate location pages, it has less attention left for the handful of pages you actually care about getting picked up quickly. On smaller or newer domains this matters more than people expect, because the crawl budget itself is smaller to begin with.
How I audit for thin content
I don’t run a word count report. I pull three things instead:
Organic sessions per page over a real window, usually 90 days minimum so seasonal noise doesn’t skew it. Pages with zero or near-zero sessions and zero backlinks are candidates, not verdicts.
Content overlap. For programmatic sections, I’ll spot check pairs of pages side by side. If I can swap two H1s and 80% of the body still reads fine, that’s a duplication problem dressed up as two pages.
Time on page against the reading time of the actual answer. If the real answer is one sentence and the page takes four paragraphs to get there, that’s the padding pattern, regardless of how long the page runs.
None of this requires guessing at what a crawler or a ranking system is “thinking.” I’m reading the page the way a visitor would and asking whether it did the job it promised in the title tag.
Fixing thin pages: consolidate, expand, or cut
Once I’ve flagged a page, there are only three real outcomes, and picking the wrong one wastes more time than doing nothing.
Consolidate. If five pages all answer variations of the same question with none of them doing it well, merge them into one page that actually covers the topic and redirect the others. This is the fix for most programmatic overreach. It’s also usually the highest-leverage fix, because you’re not adding new work, you’re combining existing content into something that clears the bar.
Expand with something real. If the page targets a genuine, distinct query and there’s a real reason to keep it separate, go add the thing that was missing: original data, a worked example, a screenshot of the actual process, a caveat that only shows up if you’ve done the thing yourself. Not more words about the same three points.
Cut it. If a page has no traffic, no backlinks, and no realistic path to serving a distinct query, delete it and let it 404 or redirect it to a genuinely related page. Keeping it around because it might rank someday is how sites end up with thousands of dead weight URLs that nobody remembers writing.
I default to consolidate first, because it’s the option that actually turns several weak pages into one page with a real shot, instead of just removing exposure.
The threshold nobody publishes (because there isn’t one)
If you came here for a number, I’m not going to invent one to be helpful. Nobody outside the search engines has verified access to a specific length or density threshold, and anyone selling you one is selling a guess dressed up as a rule. What’s testable, and what I actually go check, is whether a page answers its own title better than the next five results do. That’s a judgment call, not a formula, and it’s the only version of “thin content” that’s held up across every site I’ve run.
If you want help finding which of your pages are actually earning their spot instead of just occupying one, that’s the kind of audit we walk through on The SEO Desk.