« Instant SEO Books of Bullshit | Home | Go Back To Playing With Your Crayons »

WordPress

By Sebastian | February 9, 2010

My bullshit detector ticked loud on a tweet today. Now that I’m digging deeper, it rattles deafening. I’m smelling some serious SEO bullshit.

By the way, do you like the clean and short title? I do. Especially when it’s displayed right below this site’s mission statement. Stay tuned.

So what raised a red flag in Dan Sharp’s tweet? He linked to Many SEO Experts Give Wrong Advice Regarding Wordpress Permalinks. I’ve condomized this link for a reason. Read on.

Rob from a hosting service based in Portland, Oregon, explains why a 30 days money back guarantee is a good idea, at least when it comes to hosting content rich high traffic WordPress blogs. Full stop. Actually, he claims that some of the most popular search geeks (for example Scott Hendison, Aaron Wall, Joost de Valk, Michael Gray, Matt Cutts, Jordan Kasteler, Vanessa Fox, Adam Audette, Stephan Spencer, Danny Sullivan …) are handing out bad advice on WordPress permalinks. He forgot to mention yours truly, but he’s cursed anyway, so this tiny sin is forgiven easily.

According to Rob, /%postname%/ permalinks shouldn’t be used, because they don’t perform with WordPress. He recommends /%year%/%monthnum%/%postname% and /%post_id%/%postname%/, or crap like that. Of course WordPress tells us that those temporary performance issues with /%postname%/ permalinks are solved:

Starting Permalinks with %postname% is strongly not recommended for performance reasons. *** Note – this has been changed and is ok to do since ver. 2.0 [Source]

Rob didn’t update his SEO bashing when the bug was “fixed” (in fact, it was just kinda optimized and is not yet really fixed, but at least /%postname%/ permalinks don’t totally jam a blog any more), perhaps because in its current shape his post is nice celeb-bait driving traffic to his hosting service. I don’t care about his reasoning. I’m outing him because he spreads false advice. Even in the first place, before WordPress increased the performance of /%postname%/ permalinks, he took a devious approach.

When a popular CMS develops such a critical bug, the right thing to do is posting a warning like “Don’t update WordPress to version x.xx until the permalink bug that slows down your blog’s performance is fixed”. Accepting such a bug as a feature, and slamming SEO experts for recommendations that were based on a WordPress version that could handle /%postname%/ permalinks without significant performance issues, is malicious. Doing that as a spokesperson of a hosting service is damaging to the company’s reputation. In other words: Bullshit.

Done with this guy. Next.

Apropos bullshit. Reread this rant’s title. Obviously it’s not about a tiny hosting company’s failures. This Weberz hosting dude was about to reveal a real bummer, only he didn’t notice it:

WordPress is fucked up beyond any repair (FUBAR)

In other words: WordPress is utter bullshit. It’s a neat CMS for a personal blog, but it’s architecture is not suitable for professional use. Given its popularity and the number of trafficked blogs utilizing WordPress, that’s seems to be a controversial statement.

Not really. Just because a piece of shit is popular, that doesn’t mean it’s good. Stalin, Hitler, Mao and Pinochet were “popular”, but as dead bodies I like them better. Cobol and RPG/II were popular, but we don’t code in these languages any more, at least not in Web development. WP-Cache and similar plug-ins can prolong the dying of a trafficked blog.

Fortunately, some IT principles survived all paradigms so far, for example normalization, aka good database design. Not that WordPress developers ever considered database design important. Franky, their ERM doesn’t even deserve the predicate “Bullshit” – it’s way worse than any permutation of bullshit I can think of.

If you really want to puke, read discussions about WordPress code, its bug reports, or confusing statements like this one:

For performance reasons, it is not a good idea to start your permalink
structure with the category, tag, author, or postname fields. The
reason is that these are text fields, and using them at the beginning
of your permalink structure it takes more time for WordPress to
distinguish your Post URLs from Page URLs (which always use the text
“page slug” as the URL), and to compensate, WordPress stores a lot of
extra information in its database (so much that sites with lots of
Pages have experienced difficulties). So, it is best to start your
permalink structure with a numeric field, such as the year or post ID.

The latter is interesting, but not understandable without some technical backgrounds. WordPress stores data used to build URIs in different database tables (wp_posts, wp_terms, …). The rules necessary to identify a piece of content (post, page, tag or category page, …) from its URI are stored in a text field in one tuple of the wp_options table. WordPress evals these rules to determine whether an URI points to a post, a page, or whatever, performing one or more database queries per rule.

On a blog with many posts, these “redirect rules” can exceed the limits of a MySQL LARGETEXT attribute, that means they can’t be stored at all. Before this happens, IOW before WordPress crashes with a database error, an HTTP request of a post can result in 2,400 or more database queries just to map its URI to a piece of content. Wow. The next best procedure to slow down page load time is an infinite loop.

Looking at this idiotic software architecture, I’m wondering how much extremely compressed bullshit can be shoved into a developers skull before it explodes out of stupidity.

I don’t know how WordPress increased the performance of /%postname%/ permalinks, and I really can’t be bothered to research it in a conglomerate of unreadable spaghetti code that these CMS clowns produce, mixing up markup with badly abstracted PHP stuff. However, it doesn’t scale, and without significant structural optimization it cannot scale.

What I can tell is that without a database table that stores URIs (URI as a unique index) with references (uri_id as foreign key in all related tables or so) to various entities like posts/pages/categories/tags and so on, preferably iplementing historizing to handle changes of the permalink settings with 301 redirects, WordPress cannot solve the performance problem.

CMS developers should know that “universal” in URI doesn’t mean “anything I can grab from various attributes stored in a couple of databse tables without properly implemented functionality that guarantees that each URI is unique on the Internet”. “URI” is an object that as an attribute of anything that counts as “piece of content” maintains its unique address, across the Web.

If you aim at high traffic, WordPress is not for you. Since SEO is about generating high volumes of traffic, WordPress is not for your clients.

The reaction by Michael Gray Is Wordpress Good or Bad for SEO? might as well answer a few questions from the comments.

Topics: Server-side silliness | Comments

Posted on 09 February 2010 by Sebastian

In your code fixing your bugs.

Sebastian has written 2 articles for SEO Bullshit

Antti Kokkonen

Interestingly, 89 out of top 100 blogs use postname somewhere in the permalink. But only one use just /%postname%/ - one! (or two if you count /blog/postname/ to the list).

Nearly half of the blogs run with the "most blog like permalink" /year/month/date/postname/... Many blogging platforms run that as default, and the top blogs post several times a day, so it makes sense.

All this doesn't make /year/month/date/* best for SEO, or best for anything, but it's what people are used to, so that might make someone think it's the best overall. But saying /%postname%/ is not good is just plain stupid.

Phil

Is this blog on wordpress?

Matt

Hi, I run a large WordPress-powered site. It's called WordPress.com and it got 1.8 billion pageviews last month, about 900 million of those directly from Google. (Is that the SEO thing?) I found that it scales great, and it's easy to scale up in a horizontal fashion cheaply and efficiently. Sometimes we get links from the homepage of Yahoo, or host blogs about Apple keynotes that get over 1,000 pageviews per second, and it works great.

We also share all of the improvements we've made to the core of WordPress so every version gets better, and we release all of our plugins and infrastructure information so other people can use a 100% Open Source stack to do exactly what we've done.

If you're having trouble, and your site is smaller than ours, maybe get in touch and we can help you through whatever is causing problems.

Sebastian

Matt, thanks for stepping in. I appreciate that you reply to a polemic rant with so nicely put irony, and friendliness.

I highly doubt that your blog, or any single blog hosted on wordpress.com, attracts 1.8 billion page views per month. You're comparing apples and oranges. If I'd install WordPress with a few plug-ins on a pretty fast box and publish tens of thousands of categorized and tagged pages and posts attracting a gazillion comments, what do you think how much traffic it can take?

Sascha

nice response. print screen.

paullopez

only one word for that 'priceless'

andreasklinger

comment of the year.

Justin Parks

"If you're having trouble, and your site is smaller than ours, maybe get in touch and we can help you through whatever is causing problems."

Im printing that out and putting it on my wall.

mvandemar

Rob didn’t update his SEO bashing when the bug was "fixed"...

Sebastian, Rob wrote his post 3.5 years after Wordpress 2.0, which is when the bug was supposedly fixed (according to the current codex version, anyways).

Sebastian

I'm not sure the codex is accurate. If you're right, then Rob deserves a BBQ party featuring the grilling of his balls.

AndyBeard

Whilst things could have been worded a lot more delicately, that would have ruined his attack hook. In some ways afaik (I not enough of a coder to fix/test) it is still valid, and I generally regard the codex as accurate as Wikipedia on many topics.
How many of the SEOs called out were actually aware of the potential issue before his post? How many were still giving advice?

Sure WP should fix stuff... there are tons of other things they still have to fix as well, and maybe that is stuff more prominent SEOs should rake WP over the coal for, & Matt Cutts as well.

alanbleiweiss

Though I've dabbled in programming through the years, I'm not a programmer by trade. Same with data architecture, CSS, and design. What I am, most of all, is and has always been, a facilitator of the flow of information - even before frakkin computers were around in the work-place.

Given that as my foundation, I need to say that I'm in total agreement as far as any site that is going to be truly scalable should, yes, be built from scratch. It's the only way to ensure the site's optimally created for the unique needs of that particular site. From data architecture to output, and everything in between, it's the right way to go.

Yet even then, the coders, data architect(s), network administrator(s), designer(s), and UI specialist(s) all better be sharp fuckers. And the project manager better be world fucking class. And the initial spec document better be a mile fucking long.

And every step of the way, the evolution of the site's functionality better be fucking handled with the same level of skill and fore-thought.

But that's the biggest downer to all of this. Because honestly, 90% of the worlds sites are NOT given that level of participation. Because they're sites that are driven by money-seeking suits who don't give a rats ass about ensuring the highest quality, even though they pay lip service to it. Which I ultimately hate about this world and yet fully understand.

As for WordPress, I am going to bet that the initial screwed un-scalable nature of it came from the fact that the guys building version 1.0 didn't put what I've just described, into play. And let's remember that initially it was scaled for internal hosting, not for portability either. And they've chosen, over time, to adapt it as they saw times changing. With more than a few classic bolt-on aspects.

While it's extremely far from perfect, the truth is that most development companies are not set up to provide from-scratch solutions that even come close to the depth of functionality WordPress has achieved to this point. And that, my friends, is another reality that we need to accept in our global turn-on-a-dime Internet time society.

Sebastian

I agree, except for the "functionality is equally important as scalability" part, provided I got that right.

If someone needs lots of functionality for a low traffic site that never will get slashdotted, WordPress is free so why not give it a try. Fuck scalability.

If someone aims at high traffic volumes, then scalability, robustness, speed and such criteria become more important.

So it's not all about WordPress vs. Whatever, or free blogging tools vs. CMSs. It's about the fine art of making the best decision based on all facts and assumptions, not to forget financial limits. Pre-launch that's not easy. Prototyping or even starting a site that has great potential with something like WordPress can be a wise decision, too, if a later migration to a more scalable system is a budgeted option.

a name

ROFL! I was just about to give WP a try, good I didnt. Maybe somebody should give them a hint to store all URIs in a seperate table and give it a 10 or 20 byte index on the URL column. Its not really difficult to figure that out.

Rob @ Lost Dawg

MY BALLS

Sandy Allen

Perhaps this is a dumb question - but why is there one camp that rants on application performance, and another camp that rants on server configuration and performance -- but these groups seem to hang out at two very different bars.

You can dramatically improve a WordPress-driven site's performance the same way you can improve Drupal, Joomla, your CMS de jure, or even hand-coded HTML: by properly setting server cache control headers, utilize output caching, application caching, single-spriting structural elements, etc.

On the flip-side, you can slow your bad-@#* Rails site or custom-coded PHP app to a crawl if you no-cache, no-store the entire thing.

Isn't the real point - get a web architect that understands servers, apps, performance, and databases. Hire one if you can, at least buy one a beer if you must, but never let anyone with "SEO" in their job description design your web site architecture. And she might actually tell you "WordPress is fine for what you are doing and the budget you have!" Of course, letting a DBA design your UI is probably not a good idea either. Don't order sushi in a steak place.

And lets get the server performance and app performance people in the same bar so end customers like me only have to buy one round?

Sebastian

There's no such thing as a dumb question, provided the questioner isn't dumb. ;-)

I don't totally agree, because a few SEOs skillfully utilize every bar out there, and have lots of experience with all sorts of booze.

Seriously, good Web architects are rare, and when you request an expert, as a rule you don't get one, or one with the wrong skill set. In other words, the problem persists in eternity, there's no way out.

Justin Parks

"Stalin, Hitler, Mao and Pinochet were “popular”, but as dead bodies I like them better."

Bloody brilliant line.

In regards the SEO bashing, the guys obviously a bit of a muppet. Things break, things get fixed. If perfection happened first time round you would all be like me (honestly :P).

Next point, OK, so wordpress has its faults. What CMS dosent? Im thinking Joomla, Drupal here. All of them have plus and minus qualities and for different reasons. Horses for courses really, but if you look at it in terms of flexibility and fun (yes I said fun), with ease of use, then Wordpress will inevitably win hands down. Its the Plug and Play of CMS systems.

Once you at your level Sebastian then hell yeah, your gonna see faults and issues but for most, its just to good and effective to be ignored. Who knows when the development of the core functions will reach critical stage and require the complete overhaul of the structure. It may happen sooner than we know.

Sebastian

Exactly Justin. The problem is, that eventually the shit will hit the fun, erm, fan. Knowing about risks involved with a technical decision, preferably long before the risk potential can reach its critical mass, can't be that bad. I totally agree that my zero-BS tolerance leads to rants that many will dismiss as plain paranoid geek stuff. Say, I'm ranting for the few that accept Murphy's Laws. So I'm not expecting that hordes of SEOs will ditch their beloved WordPress now.

Justin Parks

Me and Murphy are mates. He is a bollocks really and knows exactly when to cause crap so I know what you mean. :)

Navin Poeran

Hey Sebastian,
Nice insightful post.

One question though: how much expected traffic could actually be a bad signal to consider WP as the CMS to build a website on?

SEO Nepal

its open source. i believe you expect people to give their voice for the software improvement rather than rant so harshly against it.

Sebastian

Try changing the WP database design. So much for OS.

As for the ranting, I think that's well deserved. Of course WP worries about backwards compatibility, because so many folks have developed something around it. But that's not a valid reason to stick with ancient structures that are designed for troubles on trafficked sites. They should have the balls to discard yesterday's stuff, starting over with a completely redesigned application. Of course that hurts, but avoiding it hurts even more. If I were in charge, I'd have done that years ago.

Seodeluxe

Great article, what would you recommend als an alternative to wordpress. Due to the amount of good plugins its nice to use for seo reasons.

SEO Mofo

Hey Sebastian,

I actually read that post about permalinks back when it was first getting SEOs fired up. It still makes me cringe.

One of the benefits of using a /category/post-name.html permalink structure (a benefit that I've yet to hear anyone mention) is that you can seamlessly integrate WordPress with a static file system. I think Apache is configured by default to prioritize .html files before .php, so I can actually serve a page from WP, view the source code, copy+paste it into an html document of the same name (and directory), and FTP it to my server. No database queries or php scripts necessary; it just serves the static html file.

It sounds like a trivial benefit, but I've found it very useful for testing out new formats, layouts, javascript, etc.--without the need for any rewrites or WP file editing. It can also be used as a "cache" for static WP pages, or as an easy way to customize archive pages, the home page, etc.

Mostly though...it just reminds me of a happier time in my life...a time when I wasn't using WordPress. ;)

Tom Bianco @centripetal

Your right seomofo the /category/post-name.html works great.
But are we not we all missing the point word-press is CMS,a content management system. It is not designed to run scripts or php or run complicated database stuff outside of its own programing.. it is designed for the newbie to the moderate internet user to easily manage some content. If your going to do something serious like ecommerce or saas site you better ground up it. but for what it is its no better or worse then anything else out there and from an seo standpoint it works find by the way no one not even matt cutts knows what the fuck they are talking about when it comes to seo with all these theories the algorithm is to complex. But I know this, original content and links work and the meta stuff does not hurt either

there is said now yall can blow me up

cheese link with little or no link love value , but someone will inevitable click on it
Local Search Marketing

Sebastian

Darren, given a permalink structure like /category/page/ you can just upload /category/page/index.html|php|htm|... (DirectoryIndex index.html index.php index.htm ...) and view it. WordPress doesn't get invoked when Apache finds something that matches the requested URI on disk. When you upload /index.html (given the directory index sort order above) you disable WordPress's main page (example.com/). So you don't need the "category/" thingy for your purposes.

I wouldn't use /category/anything at all, because a post can have more than one category assigned, and then WordPress uses the category with the lowest primary key value or so in URIs. Also, when you can assign a piece of content more than one category, or tag, categorizing must not be expressed in its URI.

Watch out for Peter Young, maybe he'll put an end to your misery.

Ani Lopez

That is one of the reasons, between tons of other more, why I do not use WP for professional sites, not even for blogs and I came to develop my own CMS along years.

Stephen Cronin

So, what software are you moving this site to? :)

Sebastian

With my personal blog I'll stick with WordPress. I've never used WordPress for bigger projects than this. And I won't, obviously.

Joe Hall

So let me get this straight, You are saying that WordPress is BS because some of the moron bloggers that talk about SEO give bad advice?? I don't understand. I mean if some one from Automattic said that WordPress is the best blogging software ever, then that might be open for debate. And if they said something like WordPress is the best for SEO, then yeah that's pretty BS. But you can't call something BS just because some idiot blogger doesn't know what they are talking about, I mean hell everything in the universe is BS then!

Sebastian

Nope Joe. The BS some blogger posted has nothing to do with the crappy WordPress architecture. You can look into the WP site for yourself to research that, or at any WordPress installation

Joe Hall

Ok that's fine. If that's the case why didn't you write that or instead re-title this. "BS bloggers write BS about WordPress"?

Sebastian

Because WordPress doesn't scale, has a totally screwed database design and more performance issues than a snail on a highway. The bullshitting blogger was just the opening for the core message.

Joe Hall

But Sebastian, you can't make a very convincing argument by just making statements and no supporting ideas. For example, tell me WHY the database design is screwed. What are the performance issues? If you were going to rewrite WordPress how would YOU design the database?? I have built many custom CMSs in my day and it seems that everyone builds theirs differently. The only way that we can learn from each other is compare techniques. Or atleast describe our thoughts fully instead of just saying "totally screwed database design", and thats it.

Sebastian

Joe, of course I can do that in a rant ;-) but actually I didn't "make a very convincing argument by just making statements and no supporting ideas".

A database design is screwed when it's not normalized, for example when someone stores instances of a not modeled entity in a text field of a generic options table. There are very few (technical) reasons for denormalization: usually performance issues as a result of other design errors (could apply to WordPress), or restrictions/insufficiencies of particular databases (doesn't apply here), or the necessity to answer multi-dimensional analytical queries (doesn't apply). I've mentioned that.

A database design is screwed when it lacks an abstraction of important objects. I've mentioned that, and I've provided a clue how one could persist URIs as an attribute of various objects.

I'd fire a developer who names a foreign key other than the primary key it references, and I'd question a foreign key's right to exist in a table when it references an already referenced entity with another role. I do not allow arrays as attributes in database tables, even when the DBMS supports them, and in most cases I'd fire a developer who stores serialized arrays in text fields. However, there's no rule without exception, and in extremely rare cases breaking such rules can make sense. When a database design reveals at the very first first sight that its developers don't know the rules, but are way too used to exceptions and lazy workarounds due to lack of experience and education, I call it crap.

What would I do when I'd have to revamp WordPress? I'd throw away its architecture, hire top-notch developers I can work with, and start over at the white board. It would cost Matt Mullenweg a fortune, but the result would be a scalable high-performance CMS that comes with functionality every other blogging tool out there can only dream of. Should I do that in a blog post, for free? Nope. I'm not this kind of idiot.

I'm aware that all this sounds arrogant. It's not really arrogant. It's based on experience. I agree that learning from each other is important. I share a thought every now and then, and daily I learn something from my peers, or from reading brilliant stuff. But just because there's an Internet where a mega ton of good information is available for free, that doesn't mean everyone can become wise by reading brilliant stuff, and it doesn't mean that a developer without years and years of experience can become a good developer.

Without experience knowledge is worthless. That means that sharing design patterns that are based on decades of professional experience can be pointless for savvy developers, because publishing those doesn't help the noobs who can't assimilate them, and sharing for free doesn't pay the author who makes a living from her/his hard earned skills.

I'm not arrogant enough to say that I'm brilliant, but I've worked with brilliant folks. I do know that my work experience is worth something, therefore I write about it every once in a while. I owe many great folks who helped me out, so I try to pay back to the community -- but there's a threshold.

Joe Hall

Excellent! This is exactly what I was looking for! I think the case with WordPress is an interesting one, in that many open source web projects begin by folks that aren't professionals with lots of experience. Because, if they were, they wouldn't be coding for free! And then after a project has come a long way, and has become popular the bugs in their design are so heavily integrated into the community that it would be impossible to reverse. Take for example the options table....fixing that mess would mean basically destroying about 85% of the plugins. You destroy 85% of your community's code and you have effectively guaranteed that your project quits growing! So, they are in a pickle, one that they created themselves.

As for giving away free advice, I don't usually do that either. :) Anyways, thanks for the reply!

Johnny

What would you recommend to use besides Wordpress?

Terry Van Horne

geez ... there were websites before WordPress came along... they didn't take all day long to load... and set the bar higher than anyone "who can press submit". That's partly why the blogosphere is mostly a WoDS...

Johnny

No kidding?! Thanks for that. So for a blog / client site they want to update you recommend building a site from scratch?

Terry Van Horne

Yes. of course I know programming and been building dynamic sites w/000's of pages for years. If you don't know programming and how to build a site other than w/Wordpress then I strongly recommend you hookup w/someone that does and start to learn how to build a real site. And Yes all my clients can update their own sites if they ask... there are lots of components to upload and edit or create files.

Johnny

Do you mean like your SEO site? No offense, but that thing looks like it was designed about 12 years ago and is in huge need of a fresh look. If I built a site like that for a local client they would think I am crazy an run away screaming.

Terry Van Horne

Hehe... you silly fck it was designed close to that... but what is significant... you're the moththpiece here shooting off yer yapper w/no link to his site! Just a silly anonymous fcker talking shite. At least I put my name here and link to a site. Nice graphics are easy to find... real webmasters can take that and build real sites w/out 101 plugins doing the real work like some silly fck named... who doesn't know aesthetics are not any indicator of site performance.... and that is what the discussion is about... or were you too busy trying to be a smartasse fcker... yes... I take offense to unwanted site critiques from silly fcking anonymous users named Johnny who obviously doesn't know shite from shinola...

Johnny

Ooops, looks like I struck a little nerve there. All that because of a little design critic? You must be a bit unstable. First, be a man and swear like everyone else you pussy. It is fuck not fck. I am pretty sure wit a site called SEO Bullshit they don't care if you cuss. Second, your site may have the best structure ever (highly doubtful) but that still doesn't take away from the fact I would never take you seriously landing on that piece of shit. Your tsworldofdesigns.com site is just as much a joke. Dude, you started this with your sarcastic bullshit response to my first comment. Truth is, I was hoping you would blow up like this.

Yeah, I am anonymous. I like it that way. Keeping under the radar is a good thing. I may use Wordpress for my sites but they look a hell of a lot better and load and function just as fast as your turd of a site. Yes, your site isn't pulling off blazing speed, I tried it. Google indexed 145 pages to your site in my Data Center. With your "programming knowledge" I would think a site that size would load faster for me (no its not my connection either). I'm not saying it is slow, but it isn't any faster then an average cached Wordpress install with 145 pages either.

theGypsy

Ok kids... play nice or I am taking the crayons (and comments) away.

Yea Terry, learn how to swear for fucks sake! Uh huh Johnny, throwing stones from glass houses is a bit of a wussy move too. So watch it with the 'be a man' shit.

I understand the points being made by both of you and will leave them if it says there. M'kay?

WordPress

WordPress is fucked up beyond any repair (FUBAR)

Search the Site

Other Shit

Latest Bullshit Busted

Crap on Tap

Authors

Ranting Hall of Fame

Want to post on SEOBS?

On the Radar

Stuff