Mistake: splitting large sitemaps by hash instead of by type

Mistake: splitting large sitemaps by hash instead of by type

When you cross 50k URLs and shard by ID modulo or random hash, every file mixes blog posts, products, and tag pages. You lose the ability to see which content type has indexing problems in the Coverage report.

Fix checklist:
— ✅ Split by content type first (posts.xml, products.xml, categories.xml), then by number if a type exceeds 50k
— ✅ Each shard stays under 50,000 URLs and 50MB
— Stable file naming — posts-1.xml, posts-2.xml — so a URL stays in the same file across regenerations
— This lets GSC show indexed-rate per type, turning the sitemap into a diagnostic

Definition of done: open the index; every child sitemap maps to one content type. Submit each separately in GSC and confirm you can read per-type indexed counts.