How a sitemap index works
An ordinary XML sitemap lists page URLs. A sitemap index lists sitemaps instead: each entry points at a sitemap file and, optionally, records when that file last changed. You submit the index once and the search engine follows it down to everything underneath.
It exists because the sitemap protocol caps both the number of URLs a single file may hold and the size of that file. Beyond those ceilings you have no choice but to split, and the index is what keeps the pieces connected as one set. Most content management systems, WordPress included, switch to an index automatically the moment a site outgrows a single file, which is why many owners have one without ever deciding to.
Why a sitemap index matters
The real benefit is diagnosis, not ranking. Search Console reports indexing results for each submitted sitemap separately. If every URL sits in one file, all you learn is that some pages are not indexed. If products, category pages, blog posts and location pages each have their own sitemap under an index, you learn which type is failing, and that is a completely different conversation with a developer.
It also makes maintenance sensible. A sitemap that holds only blog posts can be rebuilt when a post is published without touching the rest, and a whole section can be pulled out of the index the day it stops being something you want crawled.
Where sitemap indexes go wrong
Nesting is the first trap. An index may list sitemaps, but it may not list other indexes, and a nested structure will simply be rejected rather than followed.
The second is assuming the split itself is an improvement. Splitting a bad list into several files gives you several bad files. A sitemap should list canonical URLs that return a normal response and are genuinely meant to be indexed. Redirected URLs, noindexed pages, thin tag archives and dead links waste attention wherever they are listed, and they make the indexing report harder to read rather than easier.
The third is the last-modified date. If the system stamps today’s date on every entry each time it regenerates, the signal carries no information and gets discounted. It should change when the content actually changes.
Location matters too. A sitemap is expected to sit on the same host as the URLs it lists, and a file placed inside a subfolder can only cover URLs at or below that folder, so a sitemap dumped in a random directory may cover nothing at all.
How to get it right
Split by content type before you split by size, because that is the split that answers questions later. Submit only the index file in Search Console and Bing Webmaster Tools and let the engines find the rest, then reference the index in robots.txt so any other crawler finds it too.
After that, treat the per-sitemap indexing figures as a monthly health check. A section where indexed pages fall well behind submitted pages is telling you something specific about that template, that content type or that part of the navigation, and chasing it there is far quicker than auditing the whole site. On large catalogues this is usually where technical SEO work starts paying for itself.