Thin Content

Thin content is a page that fails to satisfy the intent it attracts — it exists, it targets a query, and it leaves the visitor with nothing they came for. The stakes stopped being theoretical years ago: sites carrying a large share of thin pages get suppressed at the quality-classification level, meaning your good pages pay for your bad ones.

Thin is about substance, not word count

The laziest definition — "under 300 words" — is wrong in both directions. A 90-word page answering "what's the SMTP port" completely is not thin. A 2,000-word page generated from a template, saying nothing a hundred sibling pages don't also say, is thin as hell at any length. Thin means: no unique information, no evidence of effort or experience, nothing that justifies this URL existing separately from the pages around it.

How thin content shows up in your data

You usually meet it in Search Console before you meet it in a content review. The classic fingerprint: a growing pile of URLs under Page Indexing → Crawled – currently not indexed, concentrated in one template. Google fetched the pages, evaluated them, and declined to index — that's a quality verdict delivered at scale. Confirm the pattern with a crawl:

# Screaming Frog export: internal HTML, sorted by word count ascending
URL                                        Word Count   Indexability
/blog/tag/widgets-2/                       87           Indexable
/blog/tag/blue-widgets/                    91           Indexable
/locations/springfield-widget-repair/      143          Indexable
/locations/shelbyville-widget-repair/      143          Indexable
/locations/ogdenville-widget-repair/       143          Indexable

Identical word counts across a location template are the tell: same page, different city token. That's simultaneously thin and near-duplicate — the two problems travel together, and at volume they mature into index bloat.

Thin vs. genuinely-useful-short: the triage table

Deleting everything under a word threshold is the lazy fix, and it's how sites throw away pages that were one afternoon of work from ranking. Triage instead:

Notice the pattern: the default fix is enrich. Deletion is reserved for pages that never had a reason to exist. A thin page targeting a real query is an asset with the value not yet added — the URL, its age, and any links it collected are things a fresh page starts without.

How to check it on your own site

1. Crawl and sort by word count in Screaming Frog — not to judge by length, but because identical or near-identical low counts expose templated thinness instantly.
2. Cross-reference GSC Page Indexing: export "Crawled – currently not indexed" and look for template concentration. One URL pattern dominating the list is your target.
3. Pull 12 months of GSC performance per URL. Pages with impressions but near-zero clicks, or long-flat zero lines, go on the triage list.
4. Read ten pages from the worst template yourself and ask the only question that matters: if I searched this page's target query, would this satisfy me? Data finds candidates; judgment makes the call.
5. Log the verdicts — enrich / merge / leave / remove — with owners and dates, so next quarter's audit measures progress instead of rediscovering the list. For calibration on how much content queries actually demand, see the thin content study.

Common audit mistakes

• Word-count worship. Flagging every sub-300-word page for deletion torches useful-short pages and misses long thin ones. Fix: word count is a sorting key, not a verdict.
• Mass-deleting as the "content pruning" flex. Pruning has its place, but delete-first audits destroy URLs whose only sin was being unfinished. Fix: enrich or merge first; delete what genuinely has no query, no links, no future.
• Padding instead of enriching. Inflating a 200-word page to 1,000 words of restated fluff makes it worse — now it's thin and exhausting. Fix: add information (data, examples, images, answers), not sentences.
• Fixing pages one by one when the template is the problem. If 3,000 location pages are thin, the fix is a template redesign plus a data source, not 3,000 tickets. Fix: audit at the pattern level.
• Expecting instant recovery. Quality reassessment takes recrawls and time — removal from the index alone can take weeks, as the deindexing timeline case study documents. Fix: set expectations in months, then let the changes cook.

FAQ

How many words does a page need to avoid being thin?

There is no number, and anyone selling you one is selling a checklist. Match the depth the query deserves: some intents are satisfied in 80 words, others need 3,000. Thin is a satisfaction gap, not a length gap.

Does thin content still get penalized?

Manual actions for thin content exist but are rare. The common mechanism now is quieter: pages don't get indexed, and sitewide quality signals drag everything down. No penalty notice arrives — traffic just erodes.

Should I noindex thin pages instead of fixing them?

Noindex is a holding pattern, not a fix — reasonable for accidental archive sprawl, wrong for pages with real query targets. A noindexed page earns nothing and its inbound links go to waste. Decide what the page should be, then either make it that or consolidate it.

Is AI-generated content automatically thin?

Not automatically — but AI output published without added information, verification, or a point of view is the current era's dominant thin-content factory. Judge it by the same standard: does it contain anything a searcher couldn't get from the ten pages already ranking?

Can thin pages hurt pages that aren't thin?

Yes, and that's the whole reason to care. Site-level quality classification means a large thin section suppresses the strong pages' performance too. Cleaning up the crap is defense for your best work.