It might seem elementary, but there isn’t much study around the value of the various HTML page elements in search engine results pages (SERPs). For SEO professionals there are always discussions about what is and isn’t indexed and used as a value for search engine rankings. In fact, it has been discussed at length as to if Yahoo indexes the Meta keywords element and uses it as a ranking factor, then it was confirmed that Yahoo does index the Meta keywords, but that it has the lowest value for ranking purposes with this statement from Yahoo:
What changed with Yahoo’s ranking algorithms is that while we still index the meta keyword tag, the ranking importance given to meta keyword tags receives the lowest ranking signal in our system.
Words that appear in any other part of documents, including the body, title, description, anchor text etc., will take priority in ranking the document – the re-occurrence of these words in the meta keyword tag will not help in boosting the signal for these words. Therefore, keyword stuffing in the keyword tag will not help a page’s recall or ranking, it will actually have less effect than introducing those same words in the body of the document, or any other section.
However, when no other ranking signal is present, unique words that only appear in the meta keyword tag section of documents can still be used to recall these documents.
So, with this in mind, I started thinking through what elements are important on a page and what of these have the most value or may have none at all. I spoke to other SEOs and took an informal survey of what is thought of being the most valuable elements and which one’s are considered more important as the others. Here are the eleven elements that keywords should be used to rank well, in the order of importance based upon my own assumptions and links to each of these pages:
- Page Title
- Heading 1
- Heading 2
- Heading 3
- Body Copy (tied with all other versions of the body copy)
- Body Copy – Bold
- Body Copy – Italic
- Body Copy – Bold & Italic
- URL
- HTML Sitemap (no ranking value for anchor text)
- Meta Description (no ranking value but still important)
- Meta Keywords (no ranking value)
- XML Sitemap (no ranking value)
So, to find out what is real and what is assumed, I tested it. Here are the details of the test (here are all of the specific details):
- Choose a three word keyword phrase (methodical verbal induction) – none of the words could be contained in any other page on the domain
- Create a unique page with the keyword chosen and insert said keyword in the particular element to be tested (published on 12/09/09)
- Link the pages from only one source (my HTML sitemap)
- Wait for Google, Bing and Yahoo to index each of the pages
- Test the rank of these pages by using the site command with the keyword and variations of the keyword to find the most important element
Notes about the test
Google was the first engine to index each of the pages with Yahoo following right behind. Bing did index four of the pages pretty quickly, then four more of the twelve, but after monitoring it, Bing dropped the eight pages from their index, so now, I cannot test Bing for the ranking factors. I have lost my patience and executed the test across Google and Yahoo only.
So, before you loose total interest in this test, here are the results of the test (sans Bing):
Google
1. Title
2. HTML Sitemap (keywords are only in anchor text to another page)
3. Heading 2
4. Heading 3
5. Heading 1
6. Body Copy
7. Bold Body Copy
8. Italic Body Copy
9. Bold/Italic Body Copy
10. URL*
– Meta Description – Did not rank
– Meta Keywords – Did not rank
– XML Sitemap – Did not rank
notes
* Keywords in the URL has more value when only one word query is used
Meta details are not considered as a ranking factor
Does not index or rank XML Sitemap
Yahoo
1. Page Title
2. Heading 2
3. Heading 3
4. Meta Description
5. Meta Keywords
6. Body Copy
7. Bold Body Copy
8. Bold/Italic Body Copy
9. XML Sitemap
10. HTML Sitemap (keywords are only in anchor text to another page)
11. URL
– Heading 1
– Italic Body Copy
notes
Heading 1 is not indexed for ranking
Italic body copy is not indexed for ranking
Body copy variations are indexed and does not appear to have any priority (except italic only)
Meta details are indexed and valued for rankings
XML content is indexed and ranked in SERPs
Wow, what results. Here are the values of page elements in order of value based upon the average across Google and Yahoo:
Page Title
Heading 2
Heading 3
Meta Description
Meta Keywords
Heading 1
Body Copy
Body Copy – Bold
Body Copy – Italic
Body Copy – Bold & Italic
URL
XML Sitemap and HTML Sitemap are not elements that are controlled in the on page elements, so they shouldn’t be considered for this test, so they are not pulled into the list.
Here are some high level details:
The page title has the most value on the page
Meta elements are still important
The H1 tag is overused and has been lowered in value
Keywords in the URLs is not that valuable
With these results, it shows what everyone assumes is right (even myself) isn’t always the case, neither is what others say including what a Yahoo engineer claims. The most important thing that I found with these results is that you can never assume that everything stays the same and don’t assume get the results first and then make an informed decision.
Feel free to share your questions, cheers!
Related Posts >>
- Measuring the Value of ALT and Title Elements of Image Tags
- Further Study on the Value of Image Optimization
- Top 10 Posts of 2009
- Blogger Offers New Widgets
- Holistic Link Building Defined
{ 37 comments… read them below or add one }
First WOW I have never seen such in depth research covering SEO. I find it quite interesting that the URL came in as low as it did and H1 was pretty much discounted. I have learned quite a bit from this post perhaps if possible you could include the importance of keyword anchored backlinks + on site SEO factors. Anyway excellent post.
@Kris,
Thanks for the comment. I would love to test the value of a link, however, that would be difficult seeing that it would be harder to control, but if I can build a good concept, I might just try getting some results.
Cheers!
Yes. This is one of the most complete posts I came across on this topic and I have to admit that I still consider myself a beginner when it comes to SEO so all I concentrated on is meta keywords and reasonable content.
I didn’t know that search engines use such sophisticated methods to rank pages. And I assume that this is only what you know of.
Strange coincidence: I was looking for a list like this a week a go and didn’t find much.
I think that the H1 element’s lack of significance isn’t surprising at all. It seems that the vast majority of CSS-oriented designers override the H1 element and use it to display the site’s title on every page. The page title ends up using H2.
Sampling a handful of PR9/10 sites shows about half of them using H1 for the site title, including: Slashdot, NASA, and even the all-knowing W3C.
- Roy
Very interesting. I wouldn’t have guessed that heading 1 ranks lower than heading 2, but when you think about it it makes sense.
Not only is H1 overused but, when you think about the traditional layout of a website, H1 tags are used for the website title while H2 tags are used for section titles. Section titles are usually more relevant than page titles.
Great research!
Any sign that google understands new HTML5 elements, as anything other than normal body text?
Great post as usual! Now I have to go back thru my pages and rearrange my heading tags. I recently reworked my site and had a huge drop for one of my keywords – that keyword happens to be in an H1 tag on my home page. Thanks for your scientific approach to SEO!
What an indepth analysis, great resource that I will use and test. Would be interesting to know if google knows the new HTML5 elements, as anything other than just normal body text?
Stephen:
What an awesome compilation of data! A couple of questions, if I may…
When did you do this study?
How long did it take?
Frankly, I’m not surprised as to your finding re: META tags and the H1 tag. The spammers have really screwed a lot of things up as far as how search engine algorithms assess and assign value, relevance, return position, and PR.
As you can see I am in the process of developing my new site and sincerely appreciate the follow attribute.
I’m definitely bookmarking this site and will followup on more of your fantastic info.
BTW: If you would be interested in a reciprocal link exchange, or otherwise would be interested in doing a guest post on my site, please do not hesitate to let me know.
~Scooter
The H1, H2, H3 discovery was quite interesting, it’s one of those tid-bits you keep in the back of your mind when developing sites (thanks).
What would be really cool, would be to automate your test, so that it gets conducted regularly on a per quarter interval in order to compare results.
“Bing dropped the eight pages from their index, so now, I cannot test Yahoo for the ranking factors.”
Still making my way through the post, but don’t you mean “cannot test Bing”?
Wow, so according to this test, the H2 tag carries more weight than the H1? Definitely wasn’t expecting to see that, very interesting…
Great catch, you are absolutely correct.
That’s interesting. I’m surprised to see that the H1 tag carries less weight. But good to know.
Its really interesting to see that h2 has more weight than h1, but why? it doesn’t seem logic.
It actually does … most H1 tags are a repeat of the TITLE tag, so its like giving double relevance/weight …
@Blogging Topics you are right, thanks for the reply.
Nice page/indexing breakdown. One of the easiest to understand that I’ve found. I’ll stop by here more often.
Very helpful information. It seems that title and heading are most important for Google and Yahoo ranking.
It is surprising that XML Sitemap is not ranked by Google but is ranked by Yahoo (although of less importance). AFAIK, XML Sitemap is the byproduct of Google’s Google Sitemaps and, in fact, was one of the inventors of the Sitemaps protocol. I wonder why.
wow such a nice post, and m bit suprised that you say H2 has a more weight compare to H1 tag? can you explain a bit? and shall i use H2 tags in website or both H1 and H2 ?
wow great research, such a nice post to read, m also bit surprised by the H2 tag details
I’ve also been conscious for some time that H2 could and often is deemed more important than the H1 tag
Spammers are always looking for the quick and easy win, so remove the temptations (or devalue them) and good content and developed sites will win through
thanks for the useful post, i enjoyed it!
This was one of the greatest analysis for on-site SEO. I think the use of the H1 tag is something that everyone who makes SEO should reconsider.
You have a very interesting test setup. I think it would be particularly helpful if the test was repeated on other domains and the data aggregated to look for patterns. Right now, if this is just a test on your domain it is not enough data to draw conclusions. Just my two cents.
I liked the way you presented everything that a webmaster wants to know in a single post.
It was really useful
Awesome tests, I figured that Google would behave like that. I get alot of traffic from bing but I don’t think it is the most targeted traffic. Either way these tips will help your search engine rankings. Good post!
This is really an amazing search i have ever find out.I was really surprised to know that Heading 1 is not indexed for ranking.This is really very interesting test.These tips helps us for improving the ranking of your site in search engines.
Great post. I am kind of shocked that H1 is weighted lower than H2 and H3. Great information as well.
I’m glad that most important elements are:
1.Page Title
2. Heading 2
3. Heading 3
B’cose i largely heed on them when develop sites.
I liked what you said in the end, to keep in mind that these things are always changing.
My personal strategy is to *attempt* to follow all the suggestions, as much as possible, because as you said, what works today, might not work tomorrow. I attempt to produce qualitity content, make use of all the meta fields, in text links and header text. A little contrary to what you said however, in my personal experimentation, I’ve noticed that I can place well by making good use of the H1 or H2 tags. Content and url seems to have the most weight. Meta fields do very little. Also, by placing a few external dofollow links to trusted, relevant sites, that seems to have a positive effect on targeting traffic.
@Rohit
You should always use an H1 if you are going to include an H2, so both should be used.
Thanks for sharing. I am new at SEO, and learning more and more everyday with great articles like this. I too, was unaware H1 was less than H2, but i can see the logic behind it. I wonder for how long, these factors will remain in place, by Google, Yahoo and Bing!
Very interesting article. I’m very very surprised at the h1 vs h2 results though.
Would you recommend putting your main page headings in h2 tags then? However, this seems un-semantic to me?
Never thought that bolded or italic body texts would even be on the list. Ima have to change a few things on my site if that’s the case.
thank you this for information
woow nice tips friend