Dashes are better than Underscores ...

... between keywords in file names.

[skyhawk133]

In a guest post by Vanessa Fox on Matt Cutt's blog, Vanessa points out dashes are much better than underscores. Matt later bolds this line in the blog entry for emphasis.

I had not thought about this, but apparently blue_widgets.html is considered 1 word, where as blue-widgets.html is considered 2 words. So for multi-word titles used as static HTML names, it looks like using a hyphen is far better than using an underscore. A brief look at the SERPS does confirm this.

[mattg3]

Hey MattG3, I would not want to redirect to the root(provided I understand those redirects)

Only one, the first redirects to root as it makes sense when having outdated pages.

All others do take what is after the domain name and redirect it to a new server/directory leaving whatsafter the domain name.

so what comes in on

blafasel.com/olddir/page.html
will be redirected to
blafasel.com/newdir/page.html

or blafasel.com/somepage.html
will be redirected to
www.blafasel.com/somepage.html

here are hints how to continue
[webmasterworld.com...]
if you have issues with QUERY_STRING

you can also look for RedirectMatch on apache.org etc.

You can of course also do

RedirectMatch Permanent /yourdir/^(.*)_(.*)$ /yourdir/$1-$2

or
RedirectMatch Permanent /yourdir/^([0-9a-zA-Z]*)_([0-9a-zA-Z]+)$ /yourdir/$1-$2

or whatever fits your pages. These are just examples.

[annej]

I agree with the information we have now it makes sense to start using a dash instead of underscores on new pages. But why would anyone change a page that is doing well and has a lot of links to it. I don't see that much difference in how pages with underscores are doing from pages with dashes. It's just plain paranoia to rush out and change your whole site on things like this. MC didn't say everyone with underscores in their urls was doomed. He just suggested the other is better.

[Asia_Expat]

Sniffer... is 'My.product' not also a misude of the dot?

[WW_Watcher]

Hey matt3g, thanks for the redirect examples, I saved them for future reference!

Hey anniej, The new, smaller, non-established site is the one I renamed a bunch of files on. It is only about a month old, and has not had many of the pages indexed yet, and has not had many visitors yet.

The bigger, older, established site, I will not be changing the file names. I need to break up a bunch of the product pages(they have gotten too large over time, as I added more and more products to the pages) and I will use the dotted naming convention for all the new pages I have to create to move the products to. It was not out of paranoia, I have to change the pages, and create new pages anyway, I was only considering the re-naming, while I was changing each page.

Hey Asia_Expat, I chose the dotted naming convention over the dashed, because IMHO it is eaiser to read.

Back to watching, (I have got to stop posting, I have posted more today, than I have in months)
Thanks!
WW_Watcher

[edited by: WW_Watcher at 4:40 am (utc) on April 23, 2006]

[georgeek]

I just do not understand the logic of treating them differently. A dash is a seperator, and an underscore is a seperator, and IMHO, they should be treated the same.

They are by Yahoo!

[sniffer]

Sniffer... is 'My.product' not also a misude of the dot?

'tis, but the difference is that it doesnt make you see (as opposed to read) the two words differently imo. The dash connects them and is quite awkward, especially if there's more than 2 words

[annej]

The new, smaller, non-established site is the one I renamed a bunch of files on.

That's different, no problem renaming those pages. I know a lot of people read who never message though and I wanted to be sure some of them didn't start a mass URL renaming without understanding the whole picture.

[fathom777]

I have many of my files named using underscores. I would like to change to dashes, but I do not want to take a big hit. The underscore pages have PR and are listed in the engine. The dashed ones will be brand new.

Wont I take a hit for switching over? Can I do a 301 without dropping in rank? What would you do?

Thanks!

[g1smd]

Use hyphens or dots for new pages. Leave old pages exactly as they are (unless they don't rank, or are badly indexed).

If you really have to change some pages, then use a single regex redirect from underscore URLs to dot or hyphen URLs to catch all visitors still using old listings, old links, and old bookmarks.

[Robert Charlton]

When I saw all those hyphens in Matt's blog page filenames, it was a signal to me that Google really didn't pay much attention to words in filenames, so it really doesn't matter which way you do it.

We could argue that hyphens help when the url is in anchor text, but how many times does anyone get a url-link to an inner page? And who uses urls as anchor text in their onsite navigation.

As for the domain name, a little off-topic but it's getting mentioned... I've always felt that the domain name boost is largely about your company name, and that's what gives you the biggest boost. Only about 10% or so, if even that many, of home page inbounds I see are domain name links. Many of my clients have multi-word, non-hyphenated domain names and they do just fine with them.

[tedster]

As a side note, this conversation goes back 3-4 years on the WebmasterWorld boards. For example, see confirmation of the same thing from GoogleGuy here in 2003: [webmasterworld.com...] -- but the discussion was even older than that thread, and was conclusively settled, in the Supporters Forum.

With all the noise being made right now about this hyphen/underscore factor, I fear we have a new "Flavor of the Month" for aspiring SEO efforts. It's really NOT worth changing all the file names on a site just to get the minor boost that might, sometimes, come from having a "keyword in the url". The challenges of getting a new url indexed -- or not -- far outweigh the potential payoff for the effort, IMO.

This is not the next magic bullet for better ranking on the SERPs.

[BeeDeeDubbleU]

This is not the next magic bullet for better ranking on the SERPs.

Yes, let's not carried away with what is surely just a minor point.

[Chris_D]

Tedster, I'll see your 2003 GG reference - and raise you a 2002 reference

:)

I would go with hyphens, personally.

[webmasterworld.com...]

[mister charlie]

i'm surprised this is being discussed. i had thought that this was pretty much common knowledge. as ted notes

As a side note, this conversation goes back 3-4 years on the WebmasterWorld boards.

i heard this topic debated ad nauseam years ago.

[pageoneresults]

Tedster, I'll see your 2003 GG reference - and raise you a 2002 reference.

I'll call both of your references and and raise you a 2001 reference. ;)

[webmasterworld.com...]

Bluffing of course. ;)

[wildegray]

Thanks to Tedster again for redirecting me here.

I'd originally tried posting a thread on this myself.

Mister Charlie; the reason this is being discussed more, I think, is because of the huge number of greenthumbs in the game now (like me) and because of Matt Cutts' recent post about the use of hyphens (http://www.mattcutts.com/blog/guest-post-vanessa-fox-on-organic-site-review-session/)
(well, actually Vanessa Fox's post on Matt's blog) where Matt *bolded* the line *hyphens are better than underscores*.

At least, that's why I'm here. I thought they'd be treated the same.

Obviously, regardless of three or four years ago, Matt has left that question in the dust as of two days ago.

My question is, if I wish to change the file extensions on my site from _ to -, how do I go about it without killing my pagerank or facing duplicate penalties?

I've only got about 20 pages on the site, but each is named with a key_phrase. My PR is okay (it exists, at least; and in Yahoo! I'm on the first page for my prime keyword) but in Google I have no ranking in the top 100. Granted, the site needs to be re-written (it's an artifact from my mentally challenged predecessors), but I want to do everything I can to increase it's strength.

CAN I CHANGE one page at a time, wait for PR to filter over, and bleed through the process that way? ANY THOUGHTS?

[Robert Charlton]

...the reason this is being discussed more, I think, is because of the huge number of greenthumbs in the game now...

Some of whom may be so green they haven't read the preceding posts in this thread. ;)

ANY THOUGHTS?

Yes, forget about it. As several earlier posts in this thread indicate, the infinitestimal gain that you might get from hyphens in filenames is not worth the disruption that making such a change will create. Write some good content instead.

[tedster]

To back up Robert, when this issue first became clear to me a few years ago, I realized I had one site with underscores used in the filenames for almost all of the 2000+ pages. At the time, I felt I had better things to do - and I was not aware of any major ranking problems, although we can always use a boost.

Still, I blew it off, mostly out of time pressure plus not feeling that it was all that big an issue. Those pages rank beautifully today for those very keywords, and I'm glad I never touched them.

[wildegray]

Robert and Ted...

(sucking my thumb) Thankths.

[Hissingsid]

Hi,

I thought that all of this changed with the Florida cock up.

We noticed during and since that fateful update that Google can sense individaul words in strings.

So for example:

stylecodes
style-codes
style_codes
style.codes

are all read by Google as:

style codes

FWIW I don't think that it does this perfectly and a hyphen is a pretty strong failsafe way to ensure that Google reads the individual words.

So the hyphen folk-law drives me to hyphenate.

Regards

Sid

[kaled]

It's a fact that underscore is not a letter. Essentially, it is used by programmers, etc., instead of space in identifier names.

Search engines should treat underscores as :-
spaces in documents
binders in searches (e.g. "my world" == my_world)

However, so far as I am aware, search engine designers haven't figured this out yet.

Kaled.

[g1smd]

As was previously explained by several people, underscores are used in compter programming, so those people want to make searches that include an underscore and look only for exact matches: like HTTP_HOST for example; and not including any pages that merely mention the word HTTP and the word HOSTS.

Did I mention that dots work the same way as hyphens, and look nicer too?

[Hissingsid]

Hi,

I'd just like to apologise for talking out of my ar-$e earlier.

If you try a search for each of these you get different results.

stylecodes
style-codes
style_codes
style.codes

Sid

[Halfdeck]

If you try a search for each of these you get different results.

stylecodes
style-codes
style_codes
style.codes

Easy test:

Build a few pages, e.g. redwidget.com/redwidget.html, redwidget.com/red_widget.html, red-widget.com/blue.html. Don't use the words "red" or "widget" in the page copy. Use some random text to link to those urls.

Wait till the three pages are indexed.

Now, run "site:redwidget.com red widget" in Google. It will return 1 match: red-widget.com/blue.html.

[g1smd]

Hmm. I made a post over at Matt Cutts blog about dots in URLs. It was in amongst a discussion about hyphens in URLs and I now see that the comment was deleted.

Hoever, that was not before other people had seen it, I guess, as there are now several (unanswered) questions about dots in URLs that have been made in later postings.

Hopefully they will find this thread...

[texasville]

I am absolutely positive that I saw a Google Guy post somewhere (could have sworn it was here) in the past year that underscores are NOT treated as spacers by the bots. He said green_widgets would be read as greenwidgets.
I remember it so well because that is when I got the stupid idea to make pages as:
green widgets.htm

and then I saw that since I was on a ms server it was automatically changed to 20% or whatever in between the words. Then I found that unix servers would not read that at all. Gad my learning curve was painful.

<edit to add>

I just googled the matt cutts underscores and dashes thing. It was his blog. Not GG.

[g1smd]

Yes. Avoid spaces and underscores in URLs.

Always use hyphens or dots between words.