More Google tips from Matt Cutts
- Data center IP addresses can point to more than one data center
- Google redesigned its supplemental index earlier this year to use a different architecture from the main index
- The site: command only presents an estimate of what Google has indexed
- Another software infrastructure update that affects crawling for the main index is being gradually rolled out
In the comments section for his post, Matt elaborates on some other issues:
- To Google, there is no distinction between "internal" and "external" backlinks
- Finding pages in the supplemental index probably only indicates a lack of PageRank and links for many people
- The main index is smaller than the supplemental index
- If you see differing estimates with the site: command, the lower numbers are more likely correct
In Handling noindex meta tags Matt describes a case study he did. He found that if you include noindex in your robots meta tag, both Ask and Google will exclude the page from their indexes completely. MSN will index the URL but nothing else. Yahoo! ignores the instruction.
In Scoble visiting the Plex Matt notes that a URL with "cns!" followed by a long string of numbers, letters, and characters may look like a session ID to Google and other search engines. He also reiterates the helpfulness of repeating keywords from the title element in the page's URL, a point he has made in previous discussions. In the comments, Matt elaborates by saying, "including the keyword in the url just gives another chance for that keyword to match the user’s query in some way".
I've incorporated some of Matt's comments in these posts and others into new sections on my (new) SEO Consulting site. Look at the SEO Fundamentals, SEO Theories, and Search Engine Facts pages.