Sitemaps: Best Practice

[Fubra]

What should and what shouldn't go in the sitemap?

In particular, pages like subscribe to our newsletter/ unsubscribe to our newsletter? Is there really any benefit in highlighting those pages to the SEs?

Thanks for any advice/ anecdotes

[Fubra]

To further on from this, we have some parameter urls in our sitemap which make me uneasy. should url.com/blah.html?option=1 be in the sitemap? If so, what benefit is that giving us?

[effectdigital]

Pages that I like to call 'core' site URLs should go in your sitemap. Basically, unique (canonical) pages which are not highly duplicate, which Google would wish to rank

I would include core addresses

I wouldn't include uploaded documents, installers, archives, resources (images, JS modules, CSS sheets, SWF objects), pagination URLs or parameter based children of canonical pages (e.g: example.com/some-page is ok to rank, but not example.com/some-page?tab=tab3). Parameters are additional funky stuff added to URLs following "?" or "&".

There are exceptions to these rules, some sites use parameters to render their on-page content - even for canonical addresses. Those old architecture types are fast dying out, though. If you're on WordPress I would index categories, but not tags which are non-hierarchical and messy (they really clutter up your SERPs)

Try crawling your site using Screaming Frog. Export all the URLs (or a large sample of them) into an Excel file. Filter the file, see which types of addresses exist on your site and which technologies are being used. Feed Google the unique, high-value pages that you know it should be ranking

I have said not to feed pagination URLs to Google, that doesn't mean they should be completely de-indexed. I just think that XML sitemaps should be pretty lean and streamlined. You can allow things which aren't in your XML sitemap to have a chance of indexation, but if you have used something like a Meta no-index tag or a robots.txt edit to block access to a page - **do not **then feed it to Google in your XML. Try to keep **all **of your indexation modules in line with each other!

No page which points to another, separate address via a canonical tag (thus calling itself 'non-canonical') should be in your XML sitemap. No page that is blocked via Meta no-index or Robots.txt should be in your sitemap.XML either

If you end up with too many pages, think about creating a sitemap XML index instead, which links through to other, separate sitemap files

Hope that helps!