Hyphens in URLs

[chromate]

Anyone alse noticed that google is highlighting (bolding) keywords in urls, even when a hyphen is not being used?

Example: search for "tree" and look at the urls of the sites returned.

Is it reasonable to assume that we no longer need hyphens to separate words? Is there any benefit to using hyphens anymore from a SEO perspective?

Or has google always done this, and I just haven't noticed it? :)

[Kyle]

I do not believe Google has done this. I think this is new...

[Mike]

Yea I think it's new.

I like the new design as well :)

[chromate]

I think the new design will work out better for people using adwords at least.

Back to the hyphen thing... this is quite a major change if it is new. Chances are google have been doing it for ages anyway and it just hasn't been apparent in the SERPs results.

It would be nice to wave goodbye to hyphens in domain names and reap the rewards of more return traffic as well as search engine traffic. What do we think?

[Mike]

I agree about the return traffic, although I think having one hyphen is ok. More than that and it is one annoying, two spammy, and three harder to remember.

[Dan Morgan]

I saw a blog post regarding that. Looks good, although it does not seem to always pick out words properly as complete and independent words.

May be a visual thing but hopefully a sign of things to come.

Choosing between hyphen and not is tricky for me, although I had no choice with the last one and hypenated ones seem to be available more often.

[Nick]

Hey Dan,

I just came here to post about this, and I noticed your post.

Yeah, check out the 6th site down. Look at the url
here (http://www.google.com/search?q=sports+nutrition&sourceid=mozilla-search&start=0&start=0&ie=utf-8&oe=utf-8)

[flyingpylon]

Just out of curiosity, why does everyone think that just because Google is highlighting search terms in the displayed results, that they're actually able to pick those words out of URLs when there is nothing there to separate them?

I'm not saying they're not doing that, but it's pretty trivial to just highlight the terms in the displayed results. Seems like there could be an awful lot of false positives if they are truly attempting to pull keywords out of a string of characters.

Then again, I'll admit that they are a lot smarter than I am.

[chromate]

Originally posted by flyingpylon
Just out of curiosity, why does everyone think that just because Google is highlighting search terms in the displayed results, that they're actually able to pick those words out of URLs when there is nothing there to separate them?

Because Google would have to be able to pick the words out in order to highlight them in the first place? :)

Whether they use them for ranking in the SERPs is a different matter. I would THINK they do though purely from personal observations of the changes I've seen for the keywords I'm monitoring.

[flyingpylon]

I guess my point was simply that there could be two independent processes going on.

1. Google takes the search terms, checks its index, and returns what it thinks are relevant sites (just as it always has).

2. The text that displays those results goes through a simple search/replace where occurances of the search terms are bolded. Nothing magical here, just searching for occurances of one string within another.

[Best.Flash]

Originally posted by chromate
I would THINK they do though purely from personal observations of the changes I've seen for the keywords I'm monitoring.

Isn't the strength of the Google system over the years the fact that they don't rely on too many on site factors for ranking sites, inother words webmaster controlled factors like these: domain names, file naming, meta tags and excessive/repeditive use of keywords on a page.

[chromate]

Originally posted by Best.Flash
Isn't the strength of the Google system over the years the fact that they don't rely on too many on site factors for ranking sites, inother words webmaster controlled factors like these: domain names, file naming, meta tags and excessive/repeditive use of keywords on a page.

Google DOES use all of those factors when ranking a page, apart from meta tags. And I'm not aware of any limit to how many times keywords can be repeated, within reason (ie, what's written makes sense).

We already know that google does look for keywords in the URL, which is why it was desirable to use hyphens in the first place. If they could find keywords in a string of characters without significant overhead then, I don't see why they wouldn't do this aswell.

flyingpylon could well be right about the two separate processes. However, the fact that google chooses to bold the keywords in the URL at all is significant. It means that they consider it of importance to the user when they're trying to find a relevant result. If Google thinks that searchers would use the URL to work out how relevant a result is, then they would probably incorporate it into the ranking algs too. Especially considering we already know they do this with hyphenated phrases.

Just a few thoughts.

[Best.Flash]

Good, I enjoy a friendly argument :D

Originally posted by chromate
Google DOES use all of those factors when ranking a page
So just to clarify your saying Google DOES use file names when ranking pages? You know this to be a fact and not conjecture?

And I'm not aware of any limit to how many times keywords can be repeated, within reason (ie, what's written makes sense).
Are you saying here Google can discern whether a pages copy 'makes sense' or not?

We already know that google does look for keywords in the URL, which is why it was desirable to use hyphens in the first place.
The reason it was advisable to use hyphens in the first place was so Google can pick out the hyphenated keywords when someone links to a page on your site using the full path as the anchor text.

However, the fact that google chooses to bold the keywords in the URL at all is significant. It means that they consider it of importance to the user when they're trying to find a relevant result. If Google thinks that searchers would use the URL to work out how relevant a result is, then they would probably incorporate it into the ranking algs too.
The fact that Google highlights any search terms in the URL is important for the user when trying to find a relevant result but how do you know that's to be taken as a positive or proof their using it in the algo? - they could be saying, Look this page keyword stuffs their directory/file names you may not want to click on this page.

Do you agree Google success was built largely on the fact that they broke from the on site meta tag, keyword density lottery used by AltaVista etc. back in the days?

[Yoda]

Google is the Alan Greenspan of the Internet. They give a sigh and hoards of people drop what they're doing to analyze the situation.

[chromate]

Originally posted by Best.Flash
Good, I enjoy a friendly argument :D

Not arguing :) Because I'm not saying Google are doing it for a fact, it's way too early to tell. I'm just throwing some thoughts around as to why I think they would do it.


Originally posted by Best.Flash
So just to clarify your saying Google DOES use file names when ranking pages? You know this to be a fact and not conjecture?

Well, when it comes to google we barely know anything to be fact these days because fact requires solid proof and the web's not exactly a controlled environment :)

But from reading numerous articles (including Chris's article on this site about forming URLs) and from personal experience, I'm pretty certain that the URL is one of the factors taken into consideration when ranking pages.



Originally posted by Best.Flash
Are you saying here Google can discern whether a pages copy 'makes sense' or not?

Nope - but humans can and they can also file spam reports.


Originally posted by Best.Flash
The reason it was advisable to use hyphens in the first place was so Google can pick out the hyphenated keywords when someone links to a page on your site using the full path as the anchor text.


Well, this is part of it - yes. But that's not the sole factor.


Originally posted by Best.Flash
The fact that Google highlights any search terms in the URL is important for the user when trying to find a relevant result but how do you know that's to be taken as a positive or proof their using it in the algo? - they could be saying, Look this page keyword stuffs their directory/file names you may not want to click on this page.


I didn't say it was proof that they're using it in the ranking algs. I said I thought it seemed likely.

You could be right in saying that google bolds the search terms enable people to pick out the sites that they don't want to click on. However, I don't think that's likely. Google's job is to find the pages that people naturally DO want to click on and rank them accordingly.

If we assume that google does use hyphenated URLs as a factor in the ranking algs, then it would make sense for them to use non-hyphenated URLs too, overhead permitting. Why would they pick one exclusively, over the other? If Google considers keywords appearing in the URL as important, then they're important, whether the URL is hyphenated or not.


Originally posted by Best.Flash
Do you agree Google success was built largely on the fact that they broke from the on site meta tag, keyword density lottery used by AltaVista etc. back in the days?

Yep :) But then again, we also know that Google does place importance on stuff the webmaster can still control: Title, Heading tags etc etc. A page title is just as easy to change as a URL, and yet that's still a key factor too.

[Best.Flash]

I rechecked Chris's article about SE freindly URL's (was this the article you where referring to?) and there's no mention of keywords in directory/file names at all.

If Google considers keywords appearing in the URL as important, then they're important, whether the URL is hyphenated or not.


I don't go along with this logic. At best i feel its to be viewed as a SERP aid for users and at worst another attempt to add more FUD to SE marketing practises which Google has repeatedly shown is one of their aims post Florida.

b.t.w as luck would have it (and for what its worth) this issue came up in the latest edition of the High Rankings Advisor newsletter.


Q) I just noticed that Google search results now highlight the searched
criteria in the file/document name. Does this mean they are giving
more weight to this than in the past?


A) Nobody really knows what it means except for Google. However, it's my
opinion that it doesn't mean anything other than the fact that their
new highlighter sees all the search words on the page, even words that
are not parsed by a hyphen, etc.

It's doubtful that this has anything to do with Google's algorithm.
They recently changed their look, and this is just a manifestation of
that.

I am doubtful that Google would put any additional weight on file
names or domain names in their ranking algorithm. If anything, I
would imagine they'd be moving further and further away from giving
those fields any weight.

I also think that this new little twist is just another way to drive
SEOs and Webmasters crazy, and give people like Danny Sullivan and me
more questions to answer!

[chromate]

Originally posted by Best.Flash
I rechecked Chris's article about SE freindly URL's (was this the article you where referring to?) and there's no mention of keywords in directory/file names at all.

I wasn't actually referring to the search engine friendly URL article. I was referring to Creating Effective URLs. To quote the article: "Your URL is often weighted more than your page content so it is a premium location for inserting keywords."


Originally posted by Best.Flash
I don't go along with this logic. At best i feel its to be viewed as a SERP aid for users and at worst another attempt to add more FUD to SE marketing practises which Google has repeatedly shown is one of their aims post Florida.

Yeah, but the reason you probably don't agree with that bit of logic is probably because you don't think google are factoring in the URL at all, even when hyphenated. If I believed that to be the case, then I wouldn't agree with that logic either :)


If anything, I would imagine they'd be moving further and further away from giving those fields any weight.

Maybe *less* weight, yes. But cutting out *all* weight given to the URL is less likely. After all, what's easier to change, on page content or URLs? I would say on page content. Considering google do still give on page content some weight, I can't see why they would then de-value the URL.

Just as an example though, search for "southbeach diet". The current #1 site is the official site. People link to it using "South Beach Diet" or using the URL "southbeachdiet.com". Until this last update, that site was no where to be seen in the SERPs for "southbeach diet" and my site "southbeach-diet.info" was hovering between #1 and #2. I can't think of any other reason why the official site should suddenly appear at #1 other than if Google no longer looks for distinct words, but rather words within a string of characters.

I'm no way trying to pass that off as some sort of conclusive evidence! :) It's just an observation which seems to fit with what I expect is happening.

At the end of the day, we wont know until more time passes so it's probably fruitless discussing it. Just to reiterate, I'm not saying they ARE doing it, I'm just saying that I THINK that if it is in their capacity then they would be doing it. :)

[Best.Flash]

quote from Creating Effective URLs
Your URL is often weighted more than your page content

I realise algorithms change as do peoples interpretation but this was the Beasmeister's opinion about keywords in URL's a few months ago:


The big bonus from this is if someone links to you using your URL, which happens all the time (most external links in forums are done this way), you will have keywords in your anchor text. If your subpages get most of your incoming links then your domain doesn't matter as much because you can always put keywords in your file and directory names.

http://www.sitepoint.com/forums/showpost.php?p=892175&postcount=19


Yeah, but the reason you probably don't agree with that bit of logic is probably because you don't think google are factoring in the URL at all

I believe they are factoring it when used in conjunction with anchor text which is the main factor Google uses to rank sites.


Maybe *less* weight, yes. But cutting out *all* weight given to the URL is less likely. After all, what's easier to change, on page content or URLs?

I think its irrelevant whether its easier to change a page content or URLs. :) but if you are going to consider it visit any SEO forum and look at the lengths site owners will explore to get a higher ranking, changing the style of the URL is one of the less extreme measures people will go to.

Regarding your example the term South Beach Diet is effectively that sites company name and Google has always been good at finding official sites, eventually ;)


At the end of the day, we wont know until more time passes so it's probably fruitless discussing it. Just to reiterate, I'm not saying they ARE doing it, I'm just saying that I THINK that if it is in their capacity then they would be doing it.

If we'll ever know :)

[chromate]

Originally posted by Best.Flash
I realise algorithms change as do peoples interpretation but this was the Beasmeister's opinion about keywords in URL's a few months ago:

http://www.sitepoint.com/forums/showpost.php?p=892175&postcount=19

I believe they are factoring it when used in conjunction with anchor text which is the main factor Google uses to rank sites.


I agree, that is the main factor and that is the main benefit of having a hyphen in a URL. But I don't think that's the only benefit, however slight the URL parsing thing might be.


Originally posted by Best.Flash Regarding your example the term South Beach Diet is effectively that sites company name and Google has always been good at finding official sites, eventually ;)

True, but that's not the only instance I noticed of sites suddenly appearing on the first page that don't contain the word "southbeach" only "south beach" in their backlink anchor text and their content. There were quite a few of them.


Originally posted by Best.Flash If we'll ever know :)

Aint that the truth! :)