Evidence that keyword in URL works

Check it out

[SlyOldDog]

I was just playing around and I typed the word "go" into Google.

I noticed a lot of Japanese and Korean sites come up. Of course they didn't have the word "go" on the page (they are in foreign character sets). They had it only in their URL.

There was always speculation that keyword in URL helped because people would link to your site with the keyword in the anchor text, but that's the first hard evidence I've seen that keyword-in-url has a direct affect on ranking.

[IanTurner]

Not surprising since 99% of the top Go players in the world are Japanese or Korean. Any of the top european players will be linking to the best Japanese sites almost by default.

[rfgdxm1]

Good example. In specific, my guess is that this is due to linking with just the URL, and no actual anchor text. I've seen lots of similar evidence, but this is one of the clearest examples.

[rfgdxm1]

IanTurner, few of the top 100 sites seem to have anything to do with that game.

[IanTurner]

Agreed but it does look like .go.jp is the equivalent of .gov for US sites.

So it is likely they will have considerable page rank and inbounds using go in the link text.

(Go - the game is just my personal experience from my contacts in Japan)

[cabbie]

This is hard to refute for those doubting thomas's and Se gurus who say it has no effect whatsoever.
My own observations with my own sites always made me believe this was the case but it was hard to prove in scientific terms to the doubters.

[PatrickDeese]

If you view the cache for a couple of the Japanese sites, you'll see this in the Google "comment area".

These terms only appear in links pointing to this page: go

Just a few dozen links like

<a href="www.example.go.jp">www.example.go.jp</a>

would probably do it.

[cabbie]

sorry misread the OP

[heini]

Unless you can prove the sites don't have incoming links using go in anchortext, we are still where we started from, nothing proven whatsoever.

just a few dozen links like

<a href="www.example.go.jp">www.example.go.jp</a>

would probably do it.

Exactly.

[ciml]

I agree with heini. Proof is very difficult with large-scale queries. Instead, it helps to think of sites that have some pages with the domain name in and some pages without the domain name in. My view:

word1word2.com can be found for a search of word1word2 or +com (this is largely irrelevant to rankings as the URL weighting is tiny)

word1word2.com gets a boost for a search of word1word2 or +com if people link to the site using the domain name as anchor text (this is often quite relevant to peoples' rankings)

word1word2.com cannot be found for a search of word1 or word2 because of the domain name.

word1word2.com gets a boost for a search of word1word2 if people link to the site using the domain name's keywords as anchor text, but split up (this is often quite relevant to peoples' rankings)

[dirkz]

Yes, no proof. Are you all sure nobody here has already done the experiment by getting a domain veryweirdkeyword.com, putting just one sentence on it and then linking it with a "click here" or "Google Test"?

[ciml]

dirkz, I can assure you that such experiments are routine. It's how people find the answers to questions like these.

[SlyOldDog]

I have one, but there is one backlink with the URL included, so I cannot prove it. Does anyone have a domain like widgets123 which is really new and has very few backlinks? Perhaps you could try it by searching for widgets123 and seeing if the URL is somewhere in the SERPs?