Should XML sitemaps include *all* pages or just the deeper ones?

[timwills]

Hi guys,

Ok this is a bit of a sitemap 101 question but I cant find a definitive answer:

When we're running out XML sitemaps for google to chew on (we're talking ecommerce and directory sites with many pages inside sub-categories here) is there any point in mentioning the homepage or even the second level pages? We know google  is crawling and indexing those and we're thinking we should trim the fat and just send a map of the bottom level pages.

What do you think?

[RyanKent]

Your sitemap should include every page of your site that you wish to be indexed.

The idea is that if your site does not provide crawlable navigation, Google can use your sitemap to crawl your site. There are some sites that use flash and when a crawler lands on a page there is absolutely no where for the crawler to go.

If your site navigation is solid then a sitemap doesn't offer any value to Google other then an indicator of when content is updated or added.

[theideapeople]

Tim,

We always index ALL pages...the help tip on Google XML also suggests including all pages of your site in the XML sitemap.

[timwills]

Well yes, that's kinda my point. We do have a sensible, crawlable navigation so there will be no problems there, so then the sitemap really becomes an indicator of what needs to be crawled (new and updated pages), but then the same question stands...

With other sites we've managed with thousands of pages we've found it detrimental to give Google hundreds of pages to crawl on a sitemap that we don't feel are important. We're pretty sure (and SEOmoz staff have supported this) that domain authority and the number of pages you can get into the index are closely related.

[RyanKent]

It is correct that DA, PA, depth of pages, etc. are all factors in determining which pages get indexed. If your site offers good navigation, reasonable backlinks, anchor text, etc then you can get close to all pages indexed even on a very large site.

Your site map should naturally include a date on every link which indicates when content was added or changed. Even if you submit a 10k list of links, Google can evaluate the dates on each link and determine which content has been added or modified since your site was last crawled.