[wp-hackers] Permalink Structure - Performance vs. SEO

[Mike Schinkel]

On Jun 14, 2011, at 9:47 PM, Lynne Pope wrote:
> If I see an URL like http://example.com/2006/11/06/great-wordpress-tricks/
> I wouldn't bother going to that link. I would assume the article was written
> on 06/11/2006 and that it is no longer relevant.

I agree, I wouldn't either. And there is a pretty good chance that is no longer relevant.

> The problem with this is that some sites will post a series of posts with updated
> information. So, there might be
> http://example.com/2007/11/06/more-great-wordpress-tricks/ and
> http://example.com/2011/11/06/even-more-great-wordpress-tricks/.

Although there are certainly special cases where this is not appropriate in general those URLs are pretty much the correct way to do it from a content strategy, site architecture and user trust basis.  That structure groups posts by year, month and day. It gives the date to indicate to the reader when it was published and the likelihood that it is still relevant (or not.) And doing this engenders trust from the reader towards the site, especially if the site has lots of content that the reader finds frequently via search engines.

As an aside, unless the site gets a large number of posts per day I prefer the shorter URL with just year and month, like these:

http://example.com/2007/06/more-great-wordpress-tricks/
http://example.com/2011/06/even-more-great-wordpress-tricks/

> BUT, on other sites, the 2006 post may be getting continually updated,
> keeping the content fresh and relevant. So,
> http://example.com/2006/11/06/great-wordpress-tricks might have been updated
> yesterday but the URL makes me think its  nearly five years old.  

And if the authors at a site do that they are, simply put making content strategy, site architecture and user trust basis errors.  We shouldn't justify a URL structure to bandaid a poor content strategy and lack of attention to site architecture.

What would be much better would be to create an "evergreen" page with a URL like the following and then continually update it instead

http://example.com/great-wordpress-tricks/

The evergreen page should link to each of the posts and then have each of the post's should link back to the evergreen page.  At the top of each post, or in a sidebar you could have an "ad" that promotes the evergreen page. This cross-linking will help concentrate link juice back to the evergreen page and the "ads" on the evergreen pages will generate awareness of the evergreen page causing more people to be more likely to link to it when they want a (more) authoritative reference.

And if "Pages" aren't a good idea then create a "Feature" custom post type and add these evergreens as features, i.e. 

http://example.com/features/great-wordpress-tricks/

Some people might say that using Features like this is essentially what category pages and tag pages do but the fact they are auto-generated tends to minimize their value and thus their impact on readers and search engines. If, on the other hand the evergreen pages are hand written and manually optimized with links to posts in context within the content then users (and search engines) will recognize they add much great value and give them appropriate credence.

Or at least that is what makes sense to me.

On Jun 14, 2011, at 3:06 PM, Alexander Concha wrote:
> Why does the user need to remember the post_id ? The autocomplete
> feature works if you write only the relevant part (in this case the
> post_name) in both cases. By applying your reasoning, the users would
> need to remember the exact url (including the http protocol, the domain
> name, the port if any, etc) they already visited.

Not true. If a user visits a site frequently then the user will have lots of URLs in the autocomplete history and the browser will only show a handful.  Meetup.com is a perfect example of a site I visit often but that use sequential numbers for their URLs of their meetups rather than a year, month, day and topic. I can't tell you how infuriating it was to constantly have to navigate the site to fine a meetup I was organizing rather than just being able to use the browser history.  Take a look at these:

http://www.meetup.com/atlantawebmarketers/events/11954596/
http://www.meetup.com/atlantawebmarketers/events/9711037/
http://www.meetup.com/atlantawebmarketers/events/11918229/

I got to the point I would remember the first four digits of a meetup URL so I could get this, but I'd often forget. Meetup.com could have made my life as an organizer so much easier if they had just designed their URLs for users and not to make the life of their web developers easier.  (Truth be told, most of Meetup.com's URLs are well designed, just not the meetup-specfic URLs.)

I will admit that it matters more for a site with a large amount of traffic and/or where users spend a lot of time using, but it's still useful as a usability tool by following Jakob's Law of web user experience:

http://j-town.co.il/web_site_design/resources/presentations/2008-10-26/slides/03-jakobs-rule.html

But browser history is but one use-case for URL design; there are many others. I wrote a blog on the topic[1] with over 50 posts several years back, until I realized that people either got it or they didn't and the latter either didn't care or would passionately try to justify why it didn't matter. So I decided to focus my attentions elsewhere:


FWIW.

-Mike
[1] http://blog.welldesignedurls.org/