New URL matching?

No need for a hyphen

[DerekH]

I've just searched for blue widgets and the Google SERPS have offered me

www.redwidgets.com - the widgets part of the URL showing a match.

Is this as new as the other coats of paint? I've always used a hyphen in my URLs until now!
DerekH

[your_store]

I want to say that they have been parsing kw's from url's at least since Florida, but it could go back further.

[DerekH]

Before that. But it's always been of the form www.kw1-kw2-kw3.com

This is something I haven't seen before - highlighting part of a URL that has no separator characters.

DerekH

[Mohamed_E]

I wonder whether the highlighting is simply part of the display algorithm, and not part of the search algorithm.

Please remember that therapist is not exacly the same as the-rapist.

[tedster]

I first noticed this with the Florida update - I believe it's a result of the new stemming engine that Google activated at that time.

[allanp73]

I tested to see if it allows people to search based on parts of a URL for example mydomain.com
A search for "mydom" does not find this site. However, a search for "my domain" does. It seems to recognize the words only by separable distinguishable words. Whether it does any ranking bonus to sites with my-domain compared to mydomain is still debatable.

[steveb]

The point of recognition is application.

Google has been able to recognize the words for at least a few months now... they are just showing off that they can now. :)

[Mohamed_E]

I just did a couple of tests on some files that seem to show Google ignoring parts of a compound word.

I have a file called someplace2002.html (one word), and the year (2002) appears nowhere in the file itself. A site search for someplace 2002 fails to return that page.

I also have a page called kw1_kw2_2001.html (underscores) and again the year appears nowhere on that page. Once again a site search for kw1 kw2 2001 did not return that page.

All pages which contain hyphens in the URL were optimized for the keywords that occur in their names, so nothing to test here.

[your_store]

I believe it's a result of the new stemming engine that Google activated at that time.

They are definitely parsing stemmed words from non-hyphenated domains. I'm seeing it a ton on singular / plural searches.

Random note, but GG said in a thread a while back that having widgets in the directory name could help for a search on widget. I believe he called it a signal.

Found the thread.. Msg #1
[webmasterworld.com...]

[caryl]

a search for allinurl:keyword keyword in Google seems to show that Google gives priority to keyword-keyword or keyword keyword Not keywordkeyword.

[MikeBeverley]

Mohamed_E - I would say that it is mainly just display matching.

Google do have the ability to parse keywords from keyword1keyword2 domain names due to their recent stemming/latent semantic indexing implementations but I really don't think they will be daft enough to actually use it.

Why would you allow domain names to influence results? None of the company's I know actually use their keywords as their company name (and therefore their domain name). Paypal.com is top for credit cards, ebay.com is top for auctions, Google realised domain names were misleading months ago and stopped using them as a weighted part of the algorithm.

The only reason people are seeing some boost from domain names is probably because the domain name is mentioned in any links to you and sometimes they even use the domain name as the text for the link.

[polarmate]

Didn't GG mention that there was no algo change just a change in UI? Which would indicate that the bolding of keywords in URLs is just window-dressing?

[steveb]

No, it would indicate they have been doing it for awhile.