• the content
• the ratio between content and code (cleaner code is better code)

So, the validation doesn’t help with SEO, despite all the buzz that validation talibans are trying to create recently.

Don’t validate your pages, it’s useless. The HTML language it’s an alive language, yesterday we were using HTML 4.01, nowadays we are using more and more HTML 5 elements, including elements that aren’t yet in the official validation specifications. For example, a few years ago people used images for rounded codes. Now we can use the border-radius property, with browser specific extensions: -moz-border-radius and -webkit-border-radius, that will never validate.

Instead of validation, focus on:

• Cross-browsing code: is your website readable with every big browser (Internet Explorer, Firefox, Chrome/Safari, Opera)? No, do not believe what the validation talibans are saying: a valid code will not display always in the same way on every browser. Just check with the “margin” and “padding” properties and see how they are displayed on Opera and Firefox: it’s different, no matter the validation.
• Clean and semantically clear code. A good code will make the website usable both on computers and mobile devices. Search engines will understand better your content. Clearly define your content areas and don’t mess the <divs>.

In fact, all this validation buzz is a pathetic marketing strategy of a bunch of losers. They have nothing of real value to offer, so they must differentiate somehow from:

• the talented designers, who are offering to their clients good, functional and clean designs;
• the good SEOs, who are offering to their clients good backlinks and valuable advices;
• the functional coders, who are offering to their clients clean templates and a good code/content ratio.

Stop this “validation” buzz. Web languages aren’t the Holy Bible. Don’t be a taliban, it’s pathetic.