How to fix thin content
You fix thin content one of 3 ways: consolidate near-duplicate pages into one real page, add the substance the query actually needs, or remove the page and return a 410. Thin means empty for the query, not short. The line Google draws is in its spam policies: “Scaled content abuse is when many pages are generated for the primary purpose of manipulating search rankings and not helping users.” Pages made for rankings instead of readers are the target. Short pages that fully answer are not.
The symptom: pages that exist but do not answer
The site has plenty of pages and almost none of them earn anything. A city page for every suburb, each one the same paragraph with the town name swapped. A blog of 300-word posts that restate their own titles and stop. Tag and archive pages that list three links. Product pages whose entire description came from the manufacturer's feed. Each page technically matches a query. None of them answers it. A searcher who lands there hits the back button, and an AI engine that reads them finds nothing worth quoting.
The tell is substitutability. Cover the page's title and read the body. If the text could sit under ten other titles without anyone noticing, the page has no substance of its own. That is what thin means. It is a measure of what the page contributes, not how much it weighs.
The cause: pages built for rankings, not for anyone
Thin content is almost always manufactured, and Google's spam policies name the two factory patterns. The first is doorways: “Doorway abuse is when sites or pages are created to rank for specific, similar search queries” while funneling everyone to the same destination. Fifty near-identical "plumber in {suburb}" pages that all route to one contact form are doorways, whatever the sitemap calls them.
The second is scaled content abuse: generating pages in bulk, by template, by scraper, or by unreviewed AI, where the volume is the strategy and no single page was made because a person needed it. The policy is explicit that this applies however the content is produced. Automation is not the sin. Emptiness at scale is.
There is also an innocent version. Sites accumulate thin pages the way houses accumulate boxes: stub posts that were going to be finished later, auto-generated archives, event pages for events that ended in 2021. No manipulation intended, but a crawler cannot read intent. It reads pages, and it reads a site where most pages give it nothing.
The fix: consolidate, add substance, or remove and 410
Triage every thin page into one of three paths. There is no fourth path where the page stays as it is.
- Consolidate. When several thin pages circle one topic, merge them into a single page that covers it properly, and 301-redirect the old URLs to it. Ten suburb doorways become one service-area page with real coverage details, real jobs, real prices.
- Add real substance. When the page targets a query you can genuinely answer, answer it. Specifics only: your numbers, your photos, your process, the questions customers actually ask. Google's helpful content guidance gives the self-test: does the content provide original information, reporting, research, or analysis? If you cannot add anything original, this is not the right path for that page.
- Remove and 410. When the page serves nobody and cannot be made to, delete it and return HTTP 410 Gone. A 410 tells crawlers the removal is deliberate and permanent. Do not soft-delete into an empty template that returns 200, because that just creates a new thin page.
# Consolidate: retire the doorway, point it at the real page /plumber-roundrock/ 301 → /service-area/ /plumber-pflugerville/ 301 → /service-area/ # Remove: the page serves nobody, say so honestly /blog/untitled-draft-2/ 410 Gone
Decide by reader, not by URL count. A smaller site where every page answers something beats a large site where the real pages drown in filler, for crawlers, for AI engines choosing what to cite, and for the person deciding whether to trust you.
Be honest: short is not thin, and word count is not the fix
Do not fix thin content by inflating it. A page that answers its query completely in 150 words is a good page. A store-hours page, a spec sheet, a straight answer to a narrow question: complete beats long, every time. Google's helpful content guidance asks, word for word, whether you are writing to a particular word count because you have heard Google has a preferred one, and answers itself: no, we don't. Padding a complete answer to 1,000 words does not make it less thin. It makes it worse, because now the answer is buried. Thin is a substance problem. Word count is neither the diagnosis nor the cure, which is why our word count guide exists mostly to kill that myth.
Verify: read as a stranger, then re-crawl
Take your ten weakest pages and read each one as a searcher who just typed its target query. Did the page answer it with anything the top result does not have? Check that consolidated URLs actually 301 to the new page, that removed URLs return a real 410 and not a 200 with an empty shell, and that the surviving pages each say something original.
Then verify it the way the engines will. Run the Brimm audit on the live site and confirm the pages we flag as low-substance are the ones you meant to keep, not leftovers. If the thin pages were near-copies of each other, our duplicate content guide covers the consolidation mechanics, and if the site as a whole is invisible to AI answers, start with why isn't my site in AI search.