SEO Best Practices for URL Structure

July 3rd, 2008 by Ann Smarty | 34 Comments

I’ve decided to make up a short all-in-one guide to summarize what we know about SEO for URLs. And if you have something to add, please do. So he we go:

Length of the URL: no more than 3-5 words in your URL. According to Google’s Matt Cutts if there are more than 5 words…

[Google] algorithms typically will just weight those words less and just not give you as much credit.”

Here is one more evidence in favor of short URL: recent research shows that short URLs within Google SERPs get clicked twice as often as long ones. So by sticking to short URLs you get both better rankings and better clickthrough.

Short URLs will also help in direct type ins of URLs (if anyone still uses that instead of Google).

Dashes are better than underscores. Although Google has no individual preferences (meaning you won’t be penalizes for either of the versions), dashes are more preferable as Google “sees” each hyphened word as an individual one:

So if you have a url like word1_word2, Google will only return that page if the user searches for word1_word2 (which almost never happens). If you have a url like word1-word2, that page can be returned for the searches word1, word2, and even “word1 word2?.

Unlike a domain name, URL is case sensitive - meaning that if by any reason (your choice or CMS) you stick to a an upper-case version, remember that this can cause a few issues: people are most likely to link to the standard lower case one and you might both lose link juice and suffer from duplicate content issues.

Moving to static URL structure: my (and actually not only my) favorite tactic is to use 301 redirect only for most powerful (in terms of linking and traffic) pages and leave all others to be handled via 404.

If you hesitate if your URLs may be perceived as spammy, check out SEOMOZ URL Spam Detection Tool that will estimate:
- spam words;
- hyphens;
- subdomain depth;
- domain length;
- digits.
Mind your file extensions (i.e. don’t end your URLs with .exe) as they might prevent your pages from crawling.

Comments

34 responses so far ↓

Michael Martinez on Jul 3, 2008 at 10:42 am

Google announced earlier this year that they are now treating underscores in URLs the same way they treat hyphens in URLs. I’m not sure of how this may affect their ability to zero in on technical content (which is why they treated underscores as indexable characters from the beginning).

It would also be a good idea to offline some of the keyword burden to sub-domains, although I’m not a fan of thin sub-domaining, where you create a 1-page sub-domain for the sake of offloading keywords. I would prefer to see robust sub-domain content.
Mark Barrera on Jul 3, 2008 at 10:57 am

Michael is correct with the underscores. Take a look at Digg for example - all of their URLs have underscores and they don’t have any problem ranking. There are MANY other sites that use the underscore and it isn’t an issue.

And as far as file extensions, don’t forget to stay away from ending the url with .0 (think web-2.0)
PR AD on Jul 3, 2008 at 11:05 am

One question:

When choosing a domain or creating URLs - should one focus on keywords to get potentially higher rankings on the SERPs.

Or should on focus on creating a URL or Domain with a name that would intrigue or be easy to remember regardless of the keywords validity.

Perhaps even if the URLs were long, they would be clicked on if the wording was enough to stir the imagination.
Pratheep on Jul 3, 2008 at 2:10 pm

@ PR AD, the answer is simple, to select a domain name add your points 1 and 2 and do not go for long domain name even if they are easy to remember.
Jeffrey Smith on Jul 3, 2008 at 2:25 pm

Thanks for the tips Ann.

In addition, here is another brief and useful post on SEO Naming Conventions.
Shinil Payamal on Jul 3, 2008 at 6:39 pm

Thanks Ann for the great tips. I was about to register couple of domains for online stores.

@ Jeffrey: Thanks for the SEO Naming Conventions link.

Regards,
Shinil.
Sally on Jul 3, 2008 at 11:02 pm

Thanks for the post Ann.

But URL is case sensitive? Is that so? I have to find out more about it.
Terry Van Horne on Jul 4, 2008 at 7:34 am

Ummm, windoze is not case sensitive ie: Thisurl.htm will also load for a request for thisurl.htm

NIX has always been case sensitive and windoze has always not been case sensitive.
Manoj Goure on Jul 4, 2008 at 11:30 am

Because of google’s monopoly, favoring hyphen in the url, it is getting prevalent over underscore.
Joey Ross on Jul 4, 2008 at 12:24 pm

Hey, nice post, but I am a little confused over the lenght of the title. Most sites that I submit to want you to have up to 60 characters in lenght.
I rank very well for my keywords, but I have 6 words in my title, should I try and reduce this to a shorter title?
Trezora - Art, Glass, Jewelry on Jul 4, 2008 at 8:07 pm

I am considering switching to a new shopping cart for my glass jewelry site and most of these carts are absolutely pitiful when it comes to SEO considerations. Their URL structures are beyond bad.

My top choice so far is X-Cart but I need to know if there exists any add-on modules to enable the following URL structure (following Ann’s suggestions):

http://www.trezora.com/jewelry/glass-pendant/sku1234.html

Thanks,

Mendy
vignarajan on Jul 6, 2008 at 11:44 am

Very nice and useful information about the URL structure. I think it is the one of the way to improve the rankings in the SERP.
sweetievale on Jul 6, 2008 at 10:37 pm

Thanks for the tips! ;-)
jay paul on Jul 7, 2008 at 4:39 am

Great Job Ann, superb info. I like Tools from seomoz and also specially spam-detection by Nick Gerner, Danny Dover, Ben Hendrickson after that there is specified like “See the blog post for more information” but that that URL is not opening for more info. so Ann, can you confirm about this URL of seomoz if you know.

or should we ask to randfish?

Thanks
Jay
Ann Smarty on Jul 7, 2008 at 8:45 am

@Jay : the post is here.
Spanish SEO on Jul 7, 2008 at 9:31 am

To continue the list…
1. Don’t use diacritics (i.e. á, é, ó, ú …etc..and..ñ) in URLs when using Spanish and other languages that contain them. People still use old browsers that struggle with these characters.

2. If using non-Roman characters, it will help to go with an IDN too, even though this may not be full non-ASCII TLD (wait until second half of 09). Chinese, Arabic and Cyrillic are good examples.

3. Keep some variety in your keyword rich URL to make it more inclusive and less spammy when comparing it to the title page and headings.
berto on Jul 7, 2008 at 11:22 am

I know that Matt Cutts referenced the 3-5 words in the URL as being “perfectly normal”, but he doesn’t reference any magic number that becomes the maximum number of words or directories that still get indexed.

Has anyone out there found that point where Google simply stops indexing after X words and/or Y sub-directories were reached in the URL?
Jeffrey Smith on Jul 7, 2008 at 11:29 am

Hi Berto:

Personally, I have not seen the magic number where Google omits a site based on naming conventions. I have however seen where the folder names are tuncated with a / / after the domain name, yet the keywords referenced still rank based on allinurl conventions.
Altum on Jul 7, 2008 at 3:05 pm

Personally I have found that sites attain the best SERP if hyphens are used and if the extension is a straightforward .html or .htm. It all goes back to trust and keeping it simple. I concentrate on just 2 or 3 keywords in the URL (except blogs), after all the more words that are used the more diluted the effect and the perceived content is less targeted.
m7mmad on Jul 17, 2008 at 11:08 pm

Thank you for this useful information i liked that tools .

peace
Brock's blog on Jul 23, 2008 at 8:47 am

Very good information here ( also in the replies :-) )
Vipin Kumar on Oct 8, 2008 at 5:47 am

Thanks all commenter!
افلام on Oct 16, 2008 at 9:56 pm

thanks
توبيكات on Oct 24, 2008 at 6:05 pm

thanks work cute
kumar on Nov 3, 2008 at 3:00 am

Can i have % symbols in static url’s
Holz-Export on Nov 15, 2008 at 8:49 pm

Great Job Ann, super info. I like Tools from seomoz and also specially spam-detection by Nick Gerner, Danny Dover, Ben Hendrickson after that there is specified like “See the blog post for more information” but that that URL is not opening for more info. so Ann, can you confirm about this URL of seomoz if you know.

Thanks Holz-Export
Paul on Dec 1, 2008 at 10:03 pm

I have pretty high SERPs but I only use 2 of my 4 words in my URL as keywords? Should I ad to my keyword list?
seo company on Dec 18, 2008 at 5:23 pm

I have seen urls have a effect on rankings but only in very competitive serps. We should do a article on htags… they are more relevant at this time..

seo agency
John G. Walter on Dec 22, 2008 at 12:04 pm

In the following URL: http://pa.photoshelter.com/c/johngwalter/gallery-img-show/G0000QvpSZCun1ls/I0000feH4FkPbbfk/0would it be best to break up my name ‘johngwalter’ into ‘john-g-walter’ to help list this URL to searchers looking for any combination of my name? And is case-sensitivity an issue?
Ann Smarty on Dec 22, 2008 at 3:35 pm

@John, for the better results, yes, ‘john-g-walter’ would be better. As for case sensitivity, it doesn’t matter for rankings - so either version would be fine.
John G. Walter on Dec 22, 2008 at 3:41 pm

Thank you, Ann. That’s what I figured. Probably doesn’t mean a whole lot with a common name(s) like mine, and as a dynamic e-commerce link, but I wanted to have what I could control set as the most effective before I invest in a lot of link coding.
Brian on Jun 2, 2009 at 3:42 pm

Thank you Ann. Do you know of any good resources to determine the best initial URL structure for a blog? For example, which is better - http://www.xyz.com/blog or blog.xyz.com?

Thank you for the useful info above.
Ann Smarty on Jun 3, 2009 at 4:33 am

Hey Brian, would this discussion be helpful? http://www.searchenginejournal.com/subdomains-or-subfolders-which-are-better-for-seo/6849/
Tam on Jun 9, 2009 at 4:47 pm

Hi Ann,
I don’t see google index most of my url
on this page http://www.bailmeoutcoupons.com/coupon-code/discount-stores.html
Is this because most of the url are too long ? Here is one of the sample url
http://www.bailmeoutcoupons.com/coupon-code/stores/Samsonite-Company-Store-Coupon-Codes-Deals.html

shoud I rename it to ../Samsonite.html or ../Samsonite-Company-Store.html ?
If I rename all of these url so shorter url, will I get penalty for pages not found for some of these pages (200+) already index on google ?

Thanks
Tam