A Better URL Structure for SEO?

Posted by Marios Alexandrou, SEO Resume.

A couple of years ago web site owners scrambled to adopt the latest and greatest URL structure to help their rankings. What people finally settled on included characteristics such as these:

  • Free of session IDs
  • As few parameters as possible
  • As Short as possible
  • Keyword-rich
  • Dashes as word separators


And for the most part I agree with all of these things, but lately I've been wondering if there isn't a better way. The three areas of weakness I see with the above approach include:

  • It's possible to have duplicate URLs if there are just words with no unique identifiers. Some people include the data in the URL, but I don't like doing that, but it seems to guarantee fewer visits for articles that are old even if they're relevant.
  • If you want to rename a URL to use different keywords, you need to also set up a 301 redirect so users and search engines can find the content when using the old URL.
  • Google News which is slowly being incorporated in to Google's main search results (Google Universal) requires a unique 3-digit (or more) identifier in every URL that isn't the date.

My proposed URL structure is this: http://www.allthingssem.com/######/one-or-more-keywords/ where ###### is a unique, numeric identifier with 3 or more digits. Advantages to this approach include:

  • Numbers in the URL are unlikely to dilute the keyword relevance of the URL.
  • You'll still get the keyword benefit from sites that link the page without creating custom anchor text.
  • The number can map to a content record in a database for guaranteed lookup success EVEN IF the keywords change i.e. the CMS can be programed to ignore everything except the number.
  • Should the keywords be changed down the road, automatic 301 redirects (based on the number in the URL) can updated search engines and users. No more manual redirects!
  • Using a unique ID should allow for technology and database independence as long as you can carry the ID to whatever new system your web site migrates to.

The biggest drawback to this new URL structure is that numbers sometimes scare people and there's no question that they make it near impossible to type the URL correctly. Sadly, I don't have a suggestion to deal with either of those two issues.

Anyone want to take me to task for proposing such a crazy idea?

[Update - October 4, 2007]
So a month after I wrote this article with what I thought to be an original idea for a URL structure, I discovered that another site is already using this structure. On the SeekingAlpha web site the URLs are structured like this:

http://www.seekingalpha.com/article/#####-keyword-keyword-keyword

An example of a real article:

http://www.seekingalpha.com/article/48921-etfs-looking-beyond-expense-ratios

And all of these URL versions will redirect to the correct artle:

http://www.seekingalpha.com/article/48921
http://www.seekingalpha.com/article/48921-etfs
http://www.seekingalpha.com/article/48921-i-can-put-whatever-junk-i-want-here

Kudos to the SeekingAlpha site! The only problem is they messed up with their redirects and used 302s (temporary) instead of 301s (permanent).

If you liked this post, please consider sharing it with others.
  • Sphinn
  • StumbleUpon
  • Mixx
  • Reddit
  • Digg
  • del.icio.us
Related SEO/SEM Articles:
4 Responses to “A Better URL Structure for SEO?”
  1. Amit Agarwal says:

    Marios, How about placing the unique ID after the keywords - something like http://www.xyz.com/one-or-more-keywords/######/

    If the URLs are long, search engines display only part of the URLs in the organic resultsets - if the keywords are more on the right side, it might help the visitor get a better idea about the page.

  2. Marios Alexandrou says:

    Amit,

    Having the ID at the end would technically work too, but it introduces a couple of potential problems. The first is that some people might think the numbers are some sort of session ID and remove them. The second issue is that you want to make sure your CMS knows where to look for the ID i.e. always in the same spot after the domain. If you put it at the end of everything and your URLs have a varying number of folders, you could run into problems.

  3. Leonard says:

    Amazon does something like that, e.g.:
    http://www.amazon.com/Habits-Highly-Effective-People/dp/0671708635

    The keyword rich part of the URL does not play any part in the Web page resolution, it's there just for SEO. I.e. you can edit the URL and it still works:
    http://www.amazon.com/Habits-of-losers/dp/0671708635

    I don't think that's a good way to do it :)

  4. Nilesh says:
Leave a Reply