Recently I received a burst of traffic from a reference citation on the popular tech site, Experts Exchange.  When I went to the source to check on what it was all about, the referenced link back to my site was not very pretty; in fact it was downright ugly!  The link supplied by the poster at Experts Exchange was in response to a previous question asked by a user.  It was simply -

also here:
http://www.os.com/?p=9

For a search engine result, this link isn’t an issue, as the results page will include the post title and a brief content snippet.  But for a manually entered link, the reader’s experience isn’t quite the same.  The default permalink doesn’t provide the potential reader with any incentive to click through to the article.  The URL is totally generic and non-descriptive and for all they know, it could be a spam link.  This was something I hadn’t considered.  In the past I was only looking at how search engines would view the URLs, not people.  Unless all one is interested in is ad clicks, a blog should be written for people, not computers.  This alone was enough to convince me to change the permalinks.

The other reason has more to do with overall design considerations than anything else.  I’ve been thinking of expanding the site with additional content and was contemplating moving the blog one level lower in the site hierarchy.  While I could easily re-install WordPress in a sub-directory, doing so would require an additional content management system for the non-blog portion of the site.  I like WordPress and it works fairly well as a CMS.  I have no desire to add yet another management layer to the site.  With WordPress, simply changing the permalink structure is an easy and effective way to push the blog content down a level without the need for a major re-design.  So now that I’ve decided the ugly permalinks needed to go, all that is left is to decide on the optimal permalink format.

There are literally hundreds of posts detailing the ongoing debate over the best permalink structure.  Most of these debates revolve around search engine optimization.  As I have stated before, I don’t believe SEO considerations are germane to a URL formatting decision. Your permalink structure decision should be based on how people interact with your site, not search engines.  Although there are several popular formats that are sound from a design standpoint, in practice they could be technically troublesome.

One popular format is the /%category%/%postname%/ structure.  While this method looks good at first, it suffers from two potential problems.  First, if you ever decide to change your category structure, or move your posts between categories, it will cause havoc with both links on search engines and referring sites.  Second, if your categories contain sub-categories, or your posts are listed in multiple categories, you may experience unpredictable URL paths.  The category WordPress will use is not necessarily the category you want.  WordPress will select the category with the lowest category ID.  Remember, pretty permalinks are the result of the rewrite engine.  WordPress still uses the ugly permalinks internally.

The other method, and I believe the most popular one, is simply using /%postname%/.  From a people standpoint this is the simplest and most descriptive URL.  It does, however, lock your blog at the root of the site.  In my previous post I stated this was good thing, but that was before I decided I wanted to add additional non-blog content to the site.  My preference now is for flexibility as well as readability.  So long as your posts aren’t buried in what looks like a long directory path, having your blog at the second level of your site shouldn’t be a problem for search engines or people.

There is also one other format that I have recently seen that is based on /%post_id%/%postname%/.  I’m not sure what to make of this design.  Creating a URL that contains a unique post ID as a path element doesn’t seem to serve any practical purpose.  It can also hinder user interaction with the site.  How often have you been on a site and manually edited the URL in the browser bar to navigate back to a previous directory?  Using /%post_id%/ in the path makes this type of navigation impossible.

After debating the pros and cons of the various link formats, it came down to either /articles/%postname%/ or /blog/%postname%/.  I decided on /blog/%postname%/ for a couple of reasons.  Having the word “articles” in the URL path is descriptive, but not necessarily accurate.  An article is really something different than a blog post.  Articles are generally longer and more formal.  I felt the posts here, although some are long, are not technically articles.  Blog is a much better description.

Readers identify blogs with dynamic Web 2.0 content.  Articles are static in nature, more like reference material.  Selecting /blog/%postname%/ seemed like it fits both my structure and design requirements.  Now the only thing left to do is to make sure I haven’t created lots of duplicate content.  One of the major advantages of the default structure is that it helps reduce duplicate content which will definitely affect SEO.  Now that I’ve changed link structure, I need to be more aware of the potential for duplicates.

As the content is now available through multiple URLs, I needed a way to control how search bots indexed the site.  I decided on using the Robots Meta plugin from Joost de Valk as it provides granular control over what and how the site is indexed.  This seems a little less ham-handed than messing with the traditional robots.txt.

So now, after eating crow, I’m ready to implement the new structure.  I hope I will not feel the need to switch it back because that will make a real mess with search engines, RSS feeds and manually entered links.  Changing your permalink structure from the default is something you only want to do once.

I have yet to see anything other than anecdotal evidence that changing your “ugly” permalinks to “pretty” permalinks has any significant positive effect on page ranking.  In fact, the reverse may be true, especially for established blogs with many indexed posts.

I have searched the web high and low and have yet to find any credible empirical study on the SEO effects of  permalink structure.   Most of the information I have garnered is purely opinion without much evidence to back it up.  It’s basically an “everyone’s doing it, so it must be right” attitude.  So, always rooting for the underdog, I’d like to put up a short defense of ugly permalinks.

1) Ugly permalinks are short and based at the root of the domain.

Custom permalinks are available in many formats, however, some defaults are offered.  Of the defaults, one of the more popular ones is the “Day and Name” permalink format: http://www.yourdomain.com/2008/07/04/sample-post/.  This URL is built using the /%year%/%monthnum%/%day%/%postname%/ tags.  Unlike the ugly permalink, this URL is three levels down from the root.  It is a bit long and it contains numbers that have no real meaning in relation to the post’s content.

Long URLs can appear as random keywords to the Google search bot unless the post title is very carefully crafted.  Google will penalize you for excessively long and random appearing URLs.  Don’t let your URLs look “spammy.”

While this popular custom URL format isn’t inherently spammy, consider the post: http://www.yourdomain.com/2008/07/04/if-i-have-to-listen-to-my-mother-in-law-complain-one-more-time/.  When reading it, it makes sense, but to a search bot, it really doesn’t contain any useful keywords.  I think with that post, the ugly permalink of: http://www.yourdomain.com/?p=123 is more than sufficient.

2) Ugly permalinks offer an absolutely unique post id.

One of the major advantages of ugly permalinks is that they ensure the post title doesn’t have to be unique.  For most people, this is not an issue, but consider the potential problem if you have multiple authors on your blog.  It is not beyond the realm of possibilities that your authors will try to create posts with duplicate titles.  The use of ugly permalinks make this a non-issue.

On the question of duplicates, one thing I haven’t been able to ascertain is whether using ugly permalinks can help eliminate duplicate content issues.  With the exception of data added after a trailing “/” as in ?p=223/xxx, no other invalid URL will return a page with ugly permalinks.  All invalid requests return a “Sorry, no posts matched your criteria” message.  With the custom URLs I have tested, the behavior is somewhat different.  It seems that any characters added after, or removed before the post name will redirect back to the post or the home page.  I’m undecided if this is the behavior I want, especially when trying to eliminate duplicates whenever possible.

I’m not sure how search bots interpret this.  Is this seen as duplicate content if someone links to you with an invalid URL?  I need to do some additional research to determine exactly how this situation is handled.

3) Ugly permalinks are immune from changes in category or post titles.

Many people use /%category%/%postname%/ format.  This can be a real nightmare if you ever decide to change your category names or structure.  All your indexed content at you favorite search engine is invalidated.  You can certainly use a redirection plugin, but that’s a real pain if you have many posts.

Post title edits, category changes, etc., do not affect ugly permalinks in the least.

4) Nobody cares about the URL anyway.

Who cares what the URL contains?  Most people find a site from a search engine or web link.  While it is probably somewhat more difficult to remember  http://www.yourdomain.com/?p=234 than http://www.yourdomain.com/my-receipe-for-bundt-cake/, who really want’s to type that in anyways?  It is much easier to bookmark the original link, or just go back to Google and google it again.

I have opted, at least for the time being, to stick with my ugly permalinks.  Unless someone can prove to me pretty permalinks have a significant SEO effect, the ugly ones seem just fine.  Your ranking is not about what’s in your URL, it’s about what’s in your post and what other sites think about your posts.  Stop worrying about permalinks and create better content that people want to link to.  That’s how page rank is built.

If I’m full of baloney please educate me as I don’t want to be the last person left using the uglies.

- A permalink heretic.