I was recently asked by a client to tell them what the developers needed to do in order to get the website ready for me to SEO.
After telling them to finalize the site with the developers, and give me a mostly finished website – copy completed or not – I mentioned errors needing to be cleaned up.
Today I got this question:
Desi, about a week ago, you mentioned the code errors on the site. Can you tell us more about them, and if you’ve run another test since? I didn’t bring it to the dev team’s notice, figuring it’s a standard thing and [they] must know of it. But I mentioned to [the boss] today and he’d like to know which test you ran; was it Google Webmaster Tools, or something else?
Here’s my response:
Make sure all involved know that many websites generate errors like this before being optimized, it’s nothing unusual. Sometimes this is the fault of a poorly built software plugin that was not built by the developers.
I ran the pages against the W3C standards, they are the arbiters of the rules of website coding. The W3C provides a useful form for web developers as a public service, which is here:
http://validator.w3.org/I also took a look in Firebug – a very useful FireFox plugin I’m sure the developers have access to, which focuses on script errors rather than page errors.
This is the ultimate standard for making any web document error-free, and is the tool that I use when checking validity of any website that produces front-end pages containing HTML, XML, or CSS.
Some errors or warnings are not really that important from an SEO stand-point, but some are. Not to get TOO technical, but if an error can slow down the server, such as calling something repeatedly that only needs to be called once, broken links inside of scripts, calls for files that aren’t there, that can affect SEO. Or if an error causes a section of the page that SHOULD appear to be the primary content area to look like a footer note, because of incorrect structural elements, that also affects SEO. Even some small failure to include required attributes and descriptors, all of these can affect SEO a little.
Basically, anything jumbled up that might be able to help the search engines better understand the subject matter in any way, and any excess code in the page that is redundant or problematic or that might generate browser errors – whether visible or not. We want to resolve that during the build.
It makes it much harder for me to do my job, much like trying to cook in an already messy kitchen.
But, rather than confuse the issue about what to fix and what to leave as is, or accidentally end up arguing over what tech is important or not to have operate validly, I usually ask the developers to work to bring these error and warning levels down to a minimum. Obviously, some snippets of code simply won’t ever validate, especially if they are cutting edge tech that has not yet been agreed upon with the W3C. But very little if any tech involved in most websites’ operation will fit that criteria. It ought to be able to be cleaned up.