SEO Toxin: Directory-like URI Structures

Great post. I especially liked your liberal use of expletives, but that’s beside the point.

Mimicking a directory structure gives an observant user a decent indication of where a page falls within the whole of your site, and if you use breadcrumbs, it gives extra internal links to your meaty ‘category pages’ — this is all.

Pagerank is so 2004. And to give you an idea of how long ago that was, I had friends then. Which says both a lot and nothing at the same time.

I agree in when you say “your information architecture should be based on estimated user behavior”.
“even better: tested!” of course but if you are managing a huge site redirections game would become a nightmare so benefits would be quite difficult to measure.

Directories like structure, while reasonable, create fantastic room for content optimization (again, specially for big sites) so I would not like people believe from this article they are not useful at all and embrace another stupid SEO fashion.

Said that I guess we’re going to enjoy this site.

You’re just crabby because the nodes in my webbed structure have prettier URIs than the nodes in your webbed structure.

Eh? In websites with varied types of content, you need a formalised directory structure so you can route the web application to show the right content on the right page.

example.com/green-widget offers no “pattern” for the web app to determine whether that page should return a product or an article (for example).

Whereas:

example.com/products/green-widget tells the app that the it is a product page about green widgets (the “/products/” part of the url is the pattern)

and

example.com/article/green-widget tells the app that the page is an article page about green widgets. (the “/article/” part of the url is the pattern)

These forms of urls have very little to do with SEO and have everything to do with building a web app that does what it is suppoosed to.

There may be cases where breadcrumbs in URIs make sense, no rule without exceptions applies here, too. It is, however, very unlikely. Usually we’re dealing with historical contamination then.

Breadcrumb navigation with pretty URIs that don’t contain any categorizing is doable, and it’s a sensible thing to do.

As for PageRank, given it’s so 2004, you can still see its bad impacts on gazillions of weak site structures “optimized” for toolbar PageRank.

Toot, there are so many ways to skin a cat.

SebaastianX: Routing is routing. Any developer who is asked to produce a site with urls that have no logical and easy routing structure is just going to laugh and/or charge you a lot more for what is essentially a cosmetic difference (I’ve been the one being laughed at…)

Also considering Google utilises breadcrumbs in their own serps (one assumes for user benefit), ignoring them for the newest SEO craze would, in my opinion, be rather foolish.

damn right, it just doesn´t matter for rankings!
the only positive effect is (once you have good rankings), that clickrates might be higher because of “speaking” URls…

Toot, did you get the impression that breadcrumb navigation at all is crap? I don’t agree. Prolly you’ve just skimmed the post.

Sebastian: No, I read the whole post. I just think your apprasial of breadcrumb naviagation is perhaps more combative than perhaps you think it looks (you are pretty harsh)

I’m defending breadcrumb/hierarchical navigation from the point of view of the development of medium size/large sites that need definable routing patterns to display content in a logical manner. You make no mention of this in your post and how costly not developing like this can be.

I’m not necessarily disagreeing with you. :)

Well, breadcrumb navigation is a great thing to do, if and where it makes sense, and that can be in particular site areas or site-wide. That doesn’t mean that navigation path components should become URI path components. Especially not when depending on different views on the contents many paths can lead to a piece of content, regardless whether that’s obvious at the time of the first IA draft, or a possible future insight. If the implementation of flexible nodes (that don’t rely on URI structures to know what to do) seems too expensive, probably a second look at the IA, the development team, and chosen instruments/tools makes sense.

But.. but.. the status quo! What have you done? :(

Breadcrumbs leave a lot of room for canonical errors, and a lot more work. Siloing however does make sense as long as it's not more than 3 clicks deep for user friendly reasons. It's been suggested that the deeper the link content the less relevant Google considers the content. I however can not confirm this but Matt Cutts implied it in one of his videos. I believe what he was suggesting is the user friendly links side of things especially in reference to getting to content in 3 clicks or less. Taxonomy in relevant structure helps search engines/users to easily determine what directories are relevant to user queries. This is especially true for geolocation with /ca/ /us/ /de/. Using webmaster tools we're able to use these heirarchies to specify regional information so in certain circumstances the URL structure is important for functional reasons.

I believe you're right about search engines changing their algorithms for overly optimized sites as they're trying more, and more to indentify more so by content rather than SEO tweaks. I think we can expect the Google slap a lot more readily in the future.

While URLS are not especially important when considering stuffing in keywords(especially if over optimized) they are a variable of significance when it come to being user friendly. Google highlights the keyword content in URLS which indicates that it does in fact find it relevant. Google is also now highlighting synonyms more so than in the past which indicates that LSI is also a factor when considering variable expert verbiage pertinent to the content at hand. Keyword research to determine the most powerful of these synonyms with Google Wonderwheel, SERP pages, SEOmoz, and Adwords tool should give strong determining factor of what keywords to be using.

It does not make sense to make duplicate content pages just for different keywords such as /XL/ /XXL/, as that's just setting yourself up to get wiped out by the next site that uses a single page. It does however make sense to structure your site in a user-friendly manner that helps both your users, and search engines communicate more effectively.

Sebastian – haven't read something of yours in a long time and like in the past, I wasn't dissapointed.

I've wrote this comment five times over already becaue I find myself agreeing and then disagreeing with some of your points in this post. I'm not sure I'll ever be able to successfully put my words on paper.

I'll cover two things and hope I make sense:

#1. Back in 2005 the company I work for LOVED web directories with a straight forward URI structure such as example.com/ca/Riverside-County/Corona/. We felt the Corona page itself along with the county page before it would help boost the relevancy of our customers link on the page.

#2. Now days when I go back to visit those exact same directories with pretty URI, I find that 70% – 90% of their internal pages aren't even indexed anymore. The biggest fault IMO not being exactly the URI structure but the lack of internal linking. So as you put it.. I fully agree with your statement, “Interlink the hell out of your page portfolio”.

Well written boss!

The directory structure just helps with site upkeep and organization from a webmaster point of view. It's just good practice. And in turn that can make SEO tasks easier.

But from an overall myth debunking point of view, you are right – a URI could theoretically be located anywhere, get indexed, and get ranked. Makes no difference.

I am not a fan of the MVC style URIs. I understand that it's necessary but I also think it's lazy. I'd like to see a Web application that handled URIs with the same intelligence a Web server handled URIs, especially if nearly all of the URIs are going to be routed through one central location.

Thanks for pointing this out, I agree totally~John online logo design

URI is totally new concept for us.Thank for sharing

lol new concept?

Thanks Man! Very clear language and academic overview – usually SEOs just post a bunch of unuseful crap but your stuff is great, direct and matches the facts! Cheers, Anton