« Appropriate Discoverability | Main | Over-Communicate »

Anchor text not limited to the anchor

I tried a search for 'google third age' and this came up:

Now that's an odd snippet. It's not from my document, it's from Stanley Wong's reference to my post:

I just read Rick Skrenta’s great blog post,
Winner-Take-All: Google and the Third Age of Computing

Rick is right on the money with a lot of his observations, especially the fact that Google has built their huge lead on the backs of the Search and Advertising dominance.

Why did that snippet come up instead of one from my document? I think it's because 'third age' doesn't actually appear in the body of my blog post. But why not just show the beginning of the post? Maybe the little blockquote table with IBM and Microsoft and the dates threw it off? Hmmm.

If I search for 'google insect overlords' I get the appropriate snippets from the body of my post:

So it looks like Google hasn't just added the text within the anchor href to the target document index material, it has gone quite a bit outside of the anchor to pull in surrounding text, and added that as well.

index material for a page =
    text on page + anchors to page + relevant text surrounding anchors

This could boost relevance in cases where anchors aren't optimally formed from the point of view of the search engine, but sufficient confidence can be gained that the nearby text is relevant to the link. e.g.: "The text for the U.S. Constitution can be found here." Clearly 'here' is not very valuable anchor text, but "U.S. Constitution" is tantalizingly close...

This isn't just for snippet purposes, words from that extended snippet/anchor work on the query side. "rick's great blog post" turns up my post, even though 'rick' doesn't appear in the link proper or anywhere on my site (I go by 'Rich', not 'Rick').

Maybe Google has been doing this for a while but I've never noticed it before. It could certainly lend a whole new angle to Googlebombing, if you can not only spike the searches for someone, but actually write their site's snippets for some queries. :-)

Comments (3)


I'm pretty sure the snippet DID come from your post - by the time G crawled it, the post apparently already had Stanley's trackback included after your post. I believe that's where the snippet came from - not Stanley's actual post on his domain.

(Also, it's pretty easy to set Movable Type up to offer custom meta descriptions for each post, which I'd be more than happy to show you. If you'd had a meta desc on your post and it included the three terms from your query, I'm nearly certain it would have pulled that copy as snippet text.)

As for terms with near-proximity to anchor text having some influence on what the target page ranks for, I think you're on the money.


Ack...well that's a much more direct explanation than the theory I came up with.

Yep, as Erik said. But to expand it more, Google generally wants to show a description that contains the reference in the document that someone searched on.

So someone search for [google third age] and gets your document. Google says "where's the first reference to those words." As you note, you don't say them in your blog post. Instead, the text comes up in that trackback post.

It's still a bit odd in that this is the second body reference, since you've got the entire "Listed below are links to weblogs that reference Winner-Take-All: Google and the Third Age of Computing" line above it. But the snippets are dynamic and may also be looking for the first reference in a "paragraph" like context -- thus skipping both your headline and your trackback header.

And yep, as Erik said, if you'd had a description tag using those three words, chances greatly increased your description tag would get used. Of course, if someone searches for something else, like:


Then you get a completely difference snippet (and your description tag might not show up in that case if all the words aren't in it)

Post a comment

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)


This page contains a single entry from the blog posted on January 11, 2007 11:46 PM.

The previous post in this blog was Appropriate Discoverability.

The next post in this blog is Over-Communicate.

Many more can be found on the main index page or by looking through the archives.

Powered by
Movable Type 3.33