Thin International Content

Element Code: HR-028

What "thin international content" actually means

This is not about hreflang syntax being broken. Your annotations can be perfectly valid, reciprocal, and pointing at the right URLs, and you can still have this problem. It shows up when the pages hreflang connects together are not equivalent in substance: the English version has a 2,000-word guide with examples, an FAQ and internal links, and the German or Japanese variant linked to it via hreflang has three paragraphs, a stub product description, or a machine translation that dropped half the sections during a rushed rollout.

It is extremely common on sites that expanded into new markets fast: launch day ships full parent-market content, international variants get a placeholder translation "to get something live," and the placeholder becomes permanent because nobody schedules the follow-up work. I have seen this on ecommerce category pages, SaaS feature pages, and blog content alike, usually traced back to the same root cause: localization treated as a translation task instead of a content task.

Why this matters beyond "it looks bad"

Google has been explicit for years that hreflang is a hint for which URL to serve to which user, not a signal that overrides quality evaluation per page. Each URL in an hreflang cluster is still assessed on its own merits for its own market and language. A thin variant does not inherit the ranking strength of its well-developed sibling. It gets evaluated as what it is: thin content, in that market, competing against local competitors who may have full, well-developed pages.

There is a second-order effect I have watched play out repeatedly: when a meaningful share of a site's international pages are consistently thin relative to the primary market, it can shape how Google's systems assess the overall quality and effort behind that domain's international expansion, which affects crawl prioritization for those sections. Thin pages get crawled less often, which means content updates, price changes, and fixes on those pages take longer to get picked up. It compounds.

There is also a straightforward user problem underneath the SEO problem: a visitor in a target market who lands on an obviously thinner, worse experience than the "real" version of the site trusts the brand less in that market. That shows up in conversion rate long before it shows up in a ranking report.

How to detect it

• Word count comparison: pull word counts for every URL in each hreflang cluster (Screaming Frog's crawl export includes this natively) and compare variant to primary. A gap of 40 percent or more is worth a manual look; anything under roughly 100-150 words on a page that should carry real content is a near-certain stub.
• Manual spot-check by native or fluent speaker: word count alone will not catch bad machine translation that hits a similar length but reads badly or misses local context, idioms, currency, or units. Have someone who actually reads the language review a sample.
• Search Console, filtered by country: look at impressions and average position for the international variant's target market. Consistently weak performance relative to how the primary version performs in its market is a signal worth investigating alongside the content itself.
• Sitebulb or Screaming Frog's hreflang report: both tools can map out clusters and flag inconsistencies; pair that map with a content-length or template audit to catch which clusters are unbalanced.
• Log file analysis: if a market's pages are being crawled noticeably less often than equivalent parent-market pages, thin or stale content on that variant is a common cause worth checking first.

How to fix it

1. Inventory every hreflang cluster and record word count, last-updated date, and whether the content was professionally localized, machine-translated, or hand-written for that market.
2. Prioritize by traffic potential: markets with real search demand and revenue potential get full content parity first, not the alphabetically first language code.
3. Localize rather than translate: adapt examples, currency, units, regulations, and cultural references to the target market instead of running a literal translation of the source page.
4. Match structural depth, not just word count: if the primary page has an FAQ, comparison table, or internal links to related content, the variant needs equivalent structure, not just equivalent length padded with filler.
5. Where a full local team is not realistic yet, invest in professional translation plus a local reviewer at minimum. Raw machine translation with no review is the single most common cause of this issue.
6. Re-run the word count and quality audit quarterly as source content gets updated, since parity that existed at launch drifts every time the primary page gets refreshed and the variant does not.

What good looks like

Every URL in an hreflang cluster carries content that is substantively equivalent in depth and usefulness to the primary version, adapted for its market rather than dumped through a translator once and forgotten. Word counts across the cluster sit in a reasonable range of each other, structural elements like FAQs and tables are present across variants, and each market's pages get refreshed on a real cadence rather than frozen at launch-day state.

FAQ