Hreflang on paginated and faceted pages: a crawl-budget case

Hreflang on paginated and faceted pages: a crawl-budget case

Hypothesis: applying full hreflang to paginated and faceted URLs was diluting crawl budget without ranking benefit. Methodology: one large catalog, ~2.4M crawlable URLs after facet explosion, 6 markets. We stripped hreflang from paginated (?page=2+) and faceted combinations, keeping it only on canonical category and product pages (~180k URLs). Log-file analysis over 10 weeks.

What the data suggests:
— Pre-change, ~34% of Googlebot hits landed on faceted URLs carrying hreflang they didn't need; those pages weren't indexed anyway.
— After restricting hreflang (and tightening canonicals) to the ~180k 'real' pages, Googlebot hits on canonical product pages rose 41% — the same crawl budget redirected to pages that matter.
— Median time-to-reindex for updated product pages dropped from 9 days to 4.
— Ranking impact was indirect and modest: localized product pages, now crawled fresher, gained an average 0.6 positions across markets.

Caveats: we changed hreflang scope and canonical handling together, so we can't cleanly attribute the crawl shift to hreflang alone — they're confounded. The 0.6-position gain is small and could be noise on individual pages; it's the aggregate that's suggestive.

Conclusion: hreflang belongs on indexable canonicals only. On a giant faceted catalog, scoping it tightly was a crawl-efficiency win more than a ranking one.