Sat 12 May 2007
Google SEO Test – Google Prefers Valid HTML & CSS?
Blurb by Shaun Anderson (Hobo)
Well we’ve started to get some results in from our qualitative Google SEO Mythbuster Test #1.
Does Google Like Valid Code? and the results we have….is yes!
So here’s a re-cap of the first SEO Mythbuster Test:
- 4 Pages on the hobo site
- New but Duplicate Content of Course
- All html pages
- All page names garbled letters to ensure no preference in Google. Will Google read the pages by alphabetical order, date created etc? Who knows – maybe we’ll find out this too.
- 1 Valid HTML + CSS
- 1 Valid HTML + Invalid CSS
- 1 Invalid HTML + Valid CSS
- 1 Invalid HTML + Invalid CSS
- Anchor Text same text so as not to influence Google (of course this isn’t good for accessibility having links to different pages using the same text phrase)
- As far as we know we’ve tried to duplicate the pages including page titles and meta-descriptions etc. What this test should also test is will Google, today at any least, collect at least one page from this duplicate content or will it ignore them all.
- We assume every thing is equal apart from validation – that includes duplicate keyword descriptions, keywords and title elements etc.
- In this first test we’ll put it in a folder. This may make everything go supplemental but we can wait and seo
So lets see what page Google likes best from this simple test. Google should follow these links below and read the pages, apply duplicate content filters to the pages, and pick one of the pages for it’s index? Right? Which one will it choose?
- link / link / link / link (This test is no longer valid)
Well – the result is clear. From these 4 pages Google managed to pick the page with valid css and valid html as the preffered page to include in it’s index! Ok, it might be a bit early to see if the four pages in the test eventually appear in Google but on first glance it appears Google spidered the pages, examined them, applied duplicate content filters as expected, and selected one to include in search engine results.
It just happens that Google seems to prefer the page with valid code as laid down by the W3C (World Wide Web Consortium).
The W3C was started in 1994 to lead the Web to its full potential by developing common protocols that promote its evolution and ensure its interoperability.
What is the W3C?
- W3C Stands for the World Wide Web Consortium
- W3C was created in October 1994
- W3C was created by Tim Berners-Lee
- W3C was created by the Inventor of the Web
- W3C is organized as a Member Organization
- W3C is working to Standardize the Web
- W3C creates and maintains WWW Standards
- W3C Standards are called W3C Recommendations
How The W3C Started
The World Wide Web (WWW) began as a project at the European Organization for Nuclear Research (CERN), where Tim Berners-Lee developed a vision of the World Wide Web.
Tim Berners-Lee – the inventor of the World Wide Web – is now the Director of the World Wide Web Consortium (W3C).
W3C was created in 1994 as a collaboration between the Massachusetts Institute of Technology (MIT) and the European Organization for Nuclear Research (CERN), with support from the U.S. Defense Advanced Research Project Agency (DARPA) and the European Commission.
W3C Standardising the Web
W3C is working to make the Web accessible to all users (despite differences in culture, education, ability, resources, and physical limitations).
W3C also coordinates its work with many other standards organizations such as the Internet Engineering Task Force, the Wireless Application Protocols (WAP) Forum and the Unicode Consortium.
Summary
It’s clear at least Google seems to prefer valid HTML & CSS. We will keep an eye on the results and see if Google decides to include the other pages, or anything changes.
If these results are correct then their may be a clear case to make sure pages you create are valid html and css – this might be the difference in whether or not you rank higher in Google serps than your competitors.
Note – Take this with a pinch of salt of course. Test everything yourself
Did you know when you link to a Hobo SEO post we have search engine friendly links back to your site if approved? Our comments are also search engine friendly you know (once you've commented on a few posts)! Do you need any more encouragement to get involved in the conversation ;)
Cerbera at Accessify pointed out –
“These tests inevitably mean the pages are not identical. As such, whether it’s the conformance of the markup or the effect of content being offset by a few bytes creating microscopic differences in ranking is impossible to say.
They also send slightly different Last-Modified and Expires headers:
1. Last-Modified: 08 May 2007 00:48:54. Expires: 13 May 2007 11:18:17.
2. Last-Modified: 08 May 2007 00:48:56. Expires: 13 May 2007 11:18:25.
3. Last-Modified: 08 May 2007 00:48:57. Expires: 13 May 2007 11:18:03.
4. Last-Modified: 08 May 2007 00:48:59. Expires: 13 May 2007 11:18:31.
Even though the modification dates are sequential, the Expires header for page “c” is earlier than all the others. It may be that Google picked what seemed like the most current one.”
This is a very good point, Ben and I factor this sort of thing into ALL my tests these days!
Thank you for doing this test! I am in the minority that defends the importance of Validation. In our opinion Validation is the cherry on the top of the SEO cake. Our website has validated pages and they DO rank very well for our keywords/phrases. Likewise our client’s that made the extra investment for our time to validate their sites have also seen very good ranking results (along with our SEO).
In our opinion the timimg issue mentioned above is irrelavent because all pages were indexed with the same process, coding and content drive relevance and the outcome speaks for itself.
Validation takes time and skill. For a web designer this is where character shows . . . are you willing to do the right thing even when no one may notice? But then again . . . maybe it is good for us that our competition doesn’t get it.
Nice test.
Why don’t do the same test with Yahoo?
I think that yahoo take more care for valid xhtml than google.
The answer for Yahoo really doesn’t matter that much; they’ve lost too much market share for it to be important.
I think this test is a great idea, but with duplicate content, it might just come down to which page Google discovered first.
Content is nothing without seo. SEO is nothing without content.
Hittail, hittail, hittail!
Its possible all this means is that google has lots of little tests which give certain pages indexing priority; maybe validated code is one of those tests?
And on some level it makes sense for code that passes the strictness of an sgml/html validator would/should/could be crawled first, since those pages probably take a less memory and cpu time to crawl, assuming it’s not the duplicate content myth at work here.
Extremely interesting test in any case.
The only scary thing about validation after the fact is that it’s horrible to do by hand; hopefully someone will write a good html to xhtml convertor for all that messed up frontpage and dreamweaver code. Hopefully frontpage 2007/expression designer is much better as it’s purported to be.
NVU and html/xhtml tidy is the bloggers best friend.
A webmaster commented that their is no way I could outrank my competitors because my website is full of html errors.
I have my site checked in W3C, It really has more than a hundred html errors. I only know basic html,so what I did is trial and error technique. It took me 8 hour before my site passed the validation.
It feels great to have that W3C logo in my site.
“A webmaster commented that their is no way I could outrank my competitors because my website is full of html errors.” – Robert this is incorrect advice.
You can rank well with invalid html – as long as it is not so screwed up Google cannot read it.
Valid HTML is just, well, cleaner – and good practice. Some of the pages on this site do not validate because we’ve been tinkering about in the background but they still do well in Google.
This seo test was just to see if Google would like the validated page “all else being equal” – in reality, nothing is ever equal ie number and quality of links, keyword density….
I’ve been trying to find out if valid code helps in rankings – thank you.
Hm…certainly an interesting result.
On the one hand, I think the test would need redoing, of course, to attempt to eliminate some of the variables suggested above.
Even still, it’s nice to see an actual live test to explore the issue of W3C standards in SEO.
Cheers Brian (haven’t we met on the UKBiz Forums? Thanks for dropping by).
Yep, that we have.
This is extremely interesting. I’m looking at revamping a popular site with XHTML/CSS instead of tables. I’d love some actual concrete evidence from Google on this.
Does anyone know of any similar articles/tests?
Nice test.
As far as I know, Javascript influences the aspects of the accessibility and SEO of a website, how about PHP?
Whenever I’ve performed SEO work with clients I’ve always told them that validation isn’t as important as content. Still, with the quality of Web Design today validation should not be a problem for anyone and your Google rank would obviously benefit from a clean page structure. I think that’s where validation helps the most. A valid web page just makes sense.
The Web sites must be always validated against W3C standards regardless of the SEO effect, because such sort of programming style indicates about professionalism of the Web design company and capabilities of the team in particular. The output generated will look better in all browsers anyway and this will benefit the end users as well.
This was a very interesting test. Although as others have pointed out, it doesn’t conclusively prove that valid HTML was the causal factor in Google indexing that particular page.
Google itself uses invalid HTML, which might be some indication of how important they consider valid code to be. Though I’m hoping this will change.
Does Google care about valid code?
We are currently doing a similar thing but on a bigger scale taking into consideration other variables.
Nothing new to me – I invest great effort in valid HTML / CSS across all self created pages
those who belief that errors don’t matter just based on the fact that they sometimes have a top serps ..
the type of error matters, different errors have different impact on the parsing software
sometimes your competition has even more errors or WORST errors than you,
never count on competitors errors – always compare to the very best possible
I agree that it is better to build a site that is W3C compliant. But in terms of SEO benefits from doing this – you will get nil. I’ve been doing SEO for 9 years and I have not seen any evidence of this.
Run a validation on Google.com and mattcutts.com/blog
See how many errors they have?
Interesting huh?
Thanks for this experiment. I am new in the field of SEO, so I don’t know much about importance of Valid Html & Css. I prefer to have vaild Html and Css. I have seen lots of our clients sites appear in the first page of google search which don’t have validated by W3C. This thing is confusing me a lot. But I think I should believe in you and your experiments.