HTML Compression and SEO?

Question

At work, we have a dedicated SEO Analyst who's job is to pour over lots of data (KeyNote/Compete etc) and generate up fancy reports for the executives so they can see how we are doing against our competitors in organic search ranking. He also leads initiatives to improve the SEO rankings on our sites by optimizing things as best we can.

We also have a longstanding mission to decrease our page load time, which right now is pretty shoddy on some pages.

The SEO guy mentioned that semantic, valid HTML gets more points by crawlers than jumbled messy HTML. I've been working on a real time HTML compressor that will decrease our page sizes my a pretty good chunk. Will compressing the HTML hurt us in site rankings?

Answer 1

I would suggest using compression at the transport layer, and eliminating whitespace from the HTML, but not sacrificing the semantics of your markup in the interest of speed. In fact, the better you "compress" your markup, the less effective the transport layer compression will be. Or, to put it a better way, let the gzip transfer-coding slim your HTML for you, and pour your energy into writing clean markup that renders quickly once it hits the browser.

Answer 2

Compressing HTML should not hurt you.

When you say HTML compressor I assume you mean a tool that removed whitespace etc from your pages to make them smaller, right? This doesn't impact how a crawler will see your html as it likely strips the same things from the HTML when it grabs the page from your site. The 'semantic' structure of the HTML exists whether compressed or not.

You might also want to look at:

• Compressing pages with an GZIP compression in the web server
• Reducing size of images, CSS, javascript etc
• Considering how the browser's layout engine loads your pages.

By jumbled HTML, this SEO person probably means the use of tables for layout and re-purposing of built in HTML elements (eg. <p class="headerOne">Header 1</p>). This increases the ratio of HTML tags to page content, or keyword density in SEO terms. It has bigger problems though:

• Longer page load times due to increased content to download, why not use the H1 tag?
• It's difficult for screenreaders to understand and affects site accessibility.
• Browsers may take longer to render the content depending on how they parse and layout pages with styles.

Answer 3

I once retooled a messy tables-for-layout to xhtml 1.0 transitional and the size went from 100kb to 40kb. The images loaded went from 200kb to just 50kb.

The reason I got such a large savings was because the site had all the JS embedded in every page. I also retooled all the JS so it was correct for both IE6 and FF2. The images were also compiled down to an image-map. All the techniques were well documented on A List Apart and easy to implement.

Answer 4

Using compression does not hurt your page ranking. Matt Cutts talks about this in his article on Crawl Caching Proxy

Your page load time can also be greatly improved by resizing your images. While you can use the height and width attributes in the img tag, this does not change the size of the images that is downloaded to the browser. Resizing the images before putting them on your pages can reduce the load time by 50% or more, depending on the number and type of images that you're using.

Other things that can improve your page load time are:

• Use web standards/CSS for layout instead of tables
• If you copy/paste content from MS Word, strip out the extra tags that Word generates
• Put CSS and javascript in external files, rather then embedded in the page. Helps when users visit more than one page on your site because browsers typically cache these files

This Web Page Analyzer will give you a speed reports that shows how long different elements of your page take to download.

Answer 5

Use gzip compression to compress the HTML in the transport stage, then just make sure that code validates and that you are using logical tags for everything.

Answer 6

The SEO guy mentioned that semantic, valid HTML gets more points by crawlers than jumbled messy HTML.

If a SEO guy ever tries to provide a fact about SEO then tell him to provide a source, because to the best of my knowledge that is simply untrue. If the content is there it will be crawled. It is a common urban-myth amongst SEO analysts that just isn't true.

However, the use of header tags is recommended. <H1> tags for the page title and <H2> for main headings, then lower down for lower headings.

I've been working on a real time HTML compressor that will decrease our page sizes my a pretty good chunk. Will compressing the HTML hurt us in site rankings?

If it can be read on the client side without problem then it is perfectly fine. If you want to look up any of this I recommend anything referencing Matt Cutt's or from the following post.

FAQ: Search Engine Optimisation

Answer 7

First you check on the code. The code is validate w3c standards like HTML & CSS