Google Prefers Invalid HTML?

113 comments

Site Reference performed a study recently to try and confirm whether valid HTML can actually help your rankings. The result was a bit surprising.

Not only does it appear that Google does not give preference to valid HTML, it seems as if they actually prefer invalid HTML.

Comments

Jill, in the overall scheme

Jill, in the overall scheme of all this, it is a major factor in SEO

I don't claim to be the best SEO of all time (far from it), but I know enough to state unequivocally that that is absolutely not true. And I'm guessing a lot of the smarter SEO's in here agree with me, like Jill -- (sorry to bring you up again, hehehe)

Proof is in the Pudding

Jill, Andy, let's put this to the test. If each of you will give me access to one of your top performing sites, I will show you how not having valid HTML will effect your SEO. Come on, don't back out now. And no, I'm not going to anything that the rest of the web isn't already doing. But, I do know which errors are likely to cause problems in crawling. So, I will implement those errors on your top performing sites and then we will revisit this topic in about 30-60 days from now.

Sound like a reasonable plan?

You have to be kidding if

You have to be kidding if you think I'd show my best performers to anyone :-)

But further, I think SEO 'tests' are simply misguided.

You can never really test a search algo (as an outsider) because you never EVER have REAL controls. And you need REAL controls to do a REAL experiment.

If you are saying some errors...

But, I do know which errors are likely to cause problems in crawling.

Of course some errors will cause problems. Specifically the ones that Mike raised will cause HUGE errors. That is a subset of all the types of errors you can introduce to your code without it impacting your ranking negatively.

Not Testing the Search Algo

Quote:
But further, I think SEO 'tests' are simply misguided.

I've clearly stated that in a previous post. This isn't really an SEO test but more of a test of how invalid code can affect the SEO. Jill states adamantly that validation has nothing at all to do with SEO when in fact it has everything to do with it. There are plenty of documents to support this claim too with a couple of them posted already in this topic. What more do you need to be convinced?

I know, let me at your top performing site and let's give you a real taste of how invalid code can have a negative impact on your SEO.

let me at your top

let me at your top performing site and let's give you a real taste of how invalid code can have a negative impact on your SEO.

let me at YOUR top performing site and let's give you a real taste of how I can efficiently exploit the niche you've found ;-)

Take my site...please

I don't know what you want to do with it, but feel free to use highrankings.com as an example, P1R. According to the validator:

Failed validation, 34 errors

This site has been pretty much in the same state since it was designed way back in hmmm...2000? Not designed by me, and I believe the original designer probably used FrontPage to create most of it. I hack at it with my 1994 copy of WebEdit whenever I need to add or change anything.

Until Big Daddy, it had been pretty much #2 - #3 for the past couple of years or so for the phrase search engine optimization. For the past few weeks, it's hanging at around #5. If you think you can validate me and get me to #1 I'll be your best friend for life. :)

Although, I have to say that I'm fully expecting it (okay praying for it) to be back at 2 or 3 once big daddy stops shuffling. So it might be better to wait until things are more stable if you really want to do something.

To me, P1R, it's common sense. It just wouldn't make any sense for the search engines to use valid code as a ranking criteria. That and the fact that pretty much any site I've ever worked on most likely didn't validate (I don't know as I don't check them), but do fine in the rankings, is good enough proof for me. I don't care what any of your tests, or Mark's tests, or anyone's tests show. I just know what I see with my own eyes, and can tell from my own common sense.

ah if we're talking

ah if we're talking 'official homepages' take andyhagans.com ... might make a good juxtaposition to jill's site as it's compliant up the wazoo (even 508!)... and yes it ranks #1 for its keyword... but that's because of HQ inbound LINKS not the validitation trust me :-)

unvalidate!

I know...why don't you unvalidate it, Andy, and see if you lose rankings!

Validation, Cell phone et al

I'm scared shitless of validation as every time I get too close things go off the deep end like AdSense targetting, SERPS, so I just leave the same old crappy tables that Google seems to love and just converted all the text to be pure CSS and to hell with the rest.

When Google won't crawl pages that don't validate and the browsers refuse to display those pages, then it's worthwhile to complete the upgrade.

Until then, I'm happily raking in money hand over fist with a site that spews errors like Linda Blair in the Excorcist and could care less.

Ever made sites for cell phones? One line of screwed up XML and they choke; usually refusing to load the page. Know why?

Quirks mode takes FARRR more CPU to process (and far larger codesets) than compliant XML. Plain tabled HTML takes FAR MORE processing power than XHTML+CSS when semantically done.

I block cell phones, f'em. Not like I'm spending my time for the 10-20 users that bounce off the firewall daily vs. the tens of thousands that get thru every day. FWIW, the excuse of CPU power for quirks is silly as a 3rd party proxy could pre-process and clean up the code for the cell phone and send it along.

Trust you?

lol, last time I heard that it cost me a lot of money. ;)

Quote:
But that's because of HQ inbound LINKS not the validitation trust me :-)

Okay, I'm well aware of the link value, no argument there. But, I would be willing to dispute your statement that the validation is not helping. Both you and Jill have clearly stated that validation has nothing to do with SEO. I claim that you are both wrong. The only way to really put this to the test is take that valid site and let's make it invalid, just for the sake of argument. If validation doesn't have anything at all to do with SEO, then your site should remain right where it is, correct? I'll take a peek at your code, provide the invalid code for you to replace it with and we'll wait and see.

Guess what? Neither you or Jill will take me up on the offer because you are not absolutely sure of what your saying. If you were, you would succumb to the test just to prove me wrong and have your day. Then pageone will scurry away with his tail between his legs and feel like a total fool. But, I know better from experience that what I say will happen, will happen. And, until someone can prove otherwise, I remain with the fact that validation is an extremely important part in the overall SEO process whether you want to admit it or not. Those who continue to fight this, are those who do not understand what validation is and cannot figure out how to correct their errors. Have you checked the validator lately? They've made some major improvements on how the errors are presented and they almost go as far as giving you the corrected code. How much easier can it get than that.

:-)

If I did your test, no matter what the result, it wouldn't prove ANYTHING. Because there still wouldn't be a CONTROL.

Although I do like your taunting.

How bout this: nyah nah na nana !!

Challenge Me?

Okay Jill, I will take an hour or two out of my schedule and help you bring highrankings.com up to at least HTML 4.01 Transitional. Heck, I might even be able to take you to HTML 4.01 Strict. It would require cooperation from your side and access to at least the home page of your site. Or, I can walk you through it over the phone error by error. Some of the errors are cascading so, once we clean up one or two at the top, a few below will go away. You've got some nesting errors there that would concern me from a technical standpoint.

http://validator.w3.org/check?uri=http%3A%2F%2Fwww.highrankings.com

That first error you see can be easily corrected by opening your home page and eliminating the duplicate data being generated in the head section. Shame on you Jill, that is totally uncalled for. Get rid of these two lines, right before the closing head element. Both of these are malformed anyway! Why do these RSS calls contain TYPE="text/css"?

LINK REL="alternate" TYPE="text/css" type="application/rss+xml" title="RSS" href="http://www.highrankings.com/high-rankings-advisor.xml
LINK REL="alternate" TYPE="text/css" type="application/rss+xml" title="RSS" href="http://www.highrankings.com/forum/rss.php

And then, what the heck is this garbage?

BODY BGCOLOR="#FFFFFF" TEXT="#000000" LINK="#8596B6" VLINK="#A8A8A8" LEFTMARGIN="0" TOPMARGIN="0" MARGINWIDTH="0" MARGINHEIGHT="0"

Don't you dare blame it on FrontPage either. For those of you who don't know me, I use FrontPage exclusively for site development and management. Don't even think about going there because I'll spank ya!

The above can be replace with this in your external style sheet...

body{color:#000;background:#fff;margin:0;padding:0;}
a:link{color:#004080;background:#fff;}
a:visited{color:#004080;background:#fff;}
a:hover{color:#004080;background:#fff;text-decoration:none;}
a:active{color:#c00;background:#fff;text-decoration:none;}

Not only did we just eliminate deprecated and invalid markup, we've also trimmed a lot of fat from the code, we all know how you ladies like to trim the fat. ;)

Once you've implemented the changes, we will be well on our merry way to becoming valid. And you know what, if you implement these changes sitewide, I'd be willing to pay you $500.00 in 90 days if you've not seen an overall improvement in the performance, usability and overall indexability of the site. We'll take a snapshot now, and another one in 90 days.

New Topic

Aaron, can we start a new topic on this and get away from the connotation of the first thread topic? Also, there is a bug here that when a topic moves to multiple pages, the refresh takes you back to the first page, arrgghh!

Can we just close this

Can we just close this thread? Enough already :-)

Before we do...

Jill, you may want to read this...

http://ln.hixie.ch/?start=1037910467&count=1

You may have some of the above issues with your site.

Andy, that's it? I can get you to bow down that easily? ;)

Both you and Jill have

Both you and Jill have clearly stated that validation has nothing to do with SEO. I claim that you are both wrong. The only way to really put this to the test is take that valid site and let's make it invalid, just for the sake of argument. If validation doesn't have anything at all to do with SEO, then your site should remain right where it is, correct?

Also worth noting that even if you made Andy's site invalid it may not prove a point, as he already used the valid code angle to write an article to get high power links (that they would probably pull if those site owners read Andy's comments in this thread hehehe).

Of course the irony of it is that the links he got from playing up the importance of accessibility are what make him #1, while some of his other sites say "almost valid" on the link to the validator hehehe.

Like a few others, I never expected Andy and Jill to argue FOR invalid code in the same thread. Funny really.

I think this thread is about done its course though?

If not, we can start another, but if we do I would so prefer to see P1R write an uber compelling post proving your point all at once...like something that became the ultimate document on why validation is a fundamental part of SEO.

You could post it here, or post it elsewhere and reference it here if you like.

Funny you should bring this up...

Quote:
If not, we can start another, but if we do I would so prefer to see P1R write an uber compelling post proving your point all at once...like something that became the ultimate document on why validation is a fundamental part of SEO.

We're working on it right now.

I'm happy to have you fix my

I'm happy to have you fix my code, but I'd prefer that you don't publicly embarrass me here with just how bad it is, by posting all the errors! Please feel free to email or PM me the info. I'll put it on (or let you do it) but I have no interest in learning it at this point in my life!

Like I said, I don't claim to know anything about writing valid code. I learned how to create webpages by hacking the the code from other websites. If I put some code in and it makes the page look like how I want it, then it's cool with me. (And obviously it's been cool with the engines since my site was #2 for one of the most competitive phrases there is for many, many years). How that doesn't prove my point to you, I'll never understand!

Regardless

of your site's position for that gem...

If I put some code in and it makes the page look like how I want it, then it's cool with me.

I hope that you're not doing any client code tinkering! ;-)

Thank you

I for one like false premises because they simply give me an edge.

I'm a vehement standards

I'm a vehement standards advocate and also believe in the SEO value of adhering to them but whilst there are strong external offsite factors involved in ranking a page/site, it is simply not possible (in absolute terms) to accredit one's ranking to valid code alone.

I think it was GoogleGuy who said they work from a baseline of HTML 3 (no doubt with error correction built in too) and we all know how badly marked-up pages were in those days!

--

Quote:
I hope that you're not doing any client code tinkering! ;-)

Nope, don't worry!

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.