Thread: The Triviality of On-Page HTML Tag Optimization

I don't agree with this. But it's an interesting take nonetheless:

I have long speculated that on-page optimization was trivial. It meshed with my understanding of how a suspicious Google engineer may treat the content of a page in relationship to its rankings. Why trust anything a webmaster says about his or her content (keywords stuffed into H1, meta, or bold tags)? Why trust anything on a page that a user won’t get to preview before visiting (anything outside the Title and Meta-Description, by-and-large)? However, despite my speculations, I lacked the data to truly start making conclusions about the usage of keywords in specific tags. Until now.....

The Triviality of On-Page HTML Tag Optimization

Thoughts?

Most people (who don't get SEO advice from Digital Point) already knew that. Everything they are testing is out of date by almost a year at least if not longer.

However, while there is little correlation between heading tags, if you're going to use them, always start with h1 and work your way down. This is more of a accessibility and readability as well as page structure rule. You don't start in the middle of your articles, so why start in the middle of the heading tags!

Title tags has always been and always will be a great indicator of what the page is about.

Keyword density, meta description/content/keywords are 80% obsolete/not needed

using alt/title attributes are worthless (and always have been). They are simply used to say what an image is about now for screen readers

placing your words in bold, italic, different colors, etc etc has never achieved anything (unless you used it to hide text/links, in which case it simply hurt you)

Duplicate content isn't even an issue unless your using it in some way to trick your visitors or SE bots.

Originally Posted by Sundance

Keyword density, meta description/content/keywords are 80% obsolete/not needed

I believe there's a lot of value to be found in the little things, such as the way Google lists your site in the results.

Regardless, I believe using nonsensical examples for his tests was the wrong way to go about testing his theory.

Originally Posted by m42

Regardless, I believe using nonsensical examples for his tests was the wrong way to go about testing his theory.

I can agree 100% there!

I ignore on page optimization - such as using H1 and other formatting. And so do the people that write for yahoo news, CNN, BBC. Those authors write for people, and a little bit for search engine. I write most of my articles just like I would have wrote them for my college essay class. And that did not including special formatting of text.

I love real testing data. I can't agree with everything in this authors testing methodology, but it's very good to have more data points.

It's interesting that he missed what has been one of the greatest tools for on-page optimization for me, the URL.

A page with a URL of http://www.example.com/red-widget.html is, in my experience, a lot easier to rank for "red widget" than a page with the URL of http://www.example.com/q54321.html.

Originally Posted by Will.Spencer

A page with a URL of http :/ /www.example.com/red-widget .html is, in my experience, a lot easier to rank for "red widget" than a page with the URL of http :/ /www.example .com/q54321.html

I agree. It also appears that new pages obtain higher rankings faster when keywords are in the URL.

Matt Cutts confirmed the value of keywords in a recent interview:

http://www.youtube.com/watch?v=gRzMhlFZz9I

The position of keywords does not matter, but the presence of keywords does help a little bit. From my perspective, every little bit that helps should not be ignored.

Originally Posted by TopDogger

I agree. It also appears that new pages obtain higher rankings faster when keywords are in the URL.

Matt Cutts confirmed the value of keywords in a recent interview:

http://www.youtube.com/watch?v=gRzMhlFZz9I

The position of keywords does not matter, but the presence of keywords does help a little bit. From my perspective, every little bit that helps should not be ignored.

It looks like Matt may not have read the question carefully. The question wasn't just about the position of keywords in the URL, it was actually about the position of keywords in different parts of the URL.

The examples in the question are:

• www.example.com/keyword/London
• www.example.com/London/keyword

That is very different than a question about:

• www.example.com/keyword-London
• www.example.com/London-keyword

The question was about keywords in the page filename vs. the page subdirectory. The answer was about keyword positions within only the filename.

Matt was also careful not to say that position didn't matter, but only that it didn't matter to him on his blog. Well duh, Matt's not a professional SEO and his traffic isn't largely search driven. He gets all of his traffic because of his position at Google.

I ran some tests a long time ago and the results were pretty clear:

• www.example.com/keyword/London <-- Bad for "keyword"; Good for London
• www.example.com/London/keyword <-- Bad for London; Good for "keyword"

A significant amount of time has elapsed, so those tests should probably be run again.

We all have to be very careful when listening to Matt Cutts. He's pretty skilled at speaking around the truth.

I ran some tests a long time ago and the results were pretty clear:

• www.example.com/keyword/London <-- Bad for "keyword"; Good for London
• www.example.com/London/keyword <-- Bad for London; Good for "keyword"

A significant amount of time has elapsed, so those tests should probably be run again.

While I have never tested this, your results are exactly what I would have expected... Finding the keyword (or keyword phrase) in the name of the page (not in the name of the parent folder, grandparent folder, etc.) I would expect to carry the most weight...

But I would also expect that having keywords (or phrases) in the parent folder name, grandparent folder name, etc. also helps, only to a lesser degree than having them all in the page name itself...

For example, I would guess that if the search phrase were "keyword1 keyword2" then:

http://www.example.com/keyword1-keyword2

would rank better than:

http://www.example.com/keyword1/keyword2

But again, I've never actually tested this to see.

Keyword rich URLs are only a small ranking factor. They don't carry near the weight of things like <title>. So testing such a situation would require LOTS of test cases with repeatable results before any conclusions could be drawn.

Originally Posted by Canonical

Keyword rich URLs are only a small ranking factor.

I really don't think they're that small, in fact I think keyword rich URL's are a very important ranking factor, specially if you have "exact matches".
I tell this based on a few recent experiences where my exact matches jumped to the first page of Google within 2 weeks for HUGELY competitive keywords.