Google Ignores <h1> tag

[janeth]

A member (google junky) here talked about the fact that Google ignored his <h> tag in his in description in the Google results.

The thread is here http://www.webproworld.com/viewtopic.php?t=27981

I asked if anyone had done any test but both mine and his post went pretty much ignored, with some members even going as far as saying that Google penalized you for not using them.

I went to other forums and asked the same question, with pretty much everyone saying no resent test had been done.

This morning I found this on one of the forums I had asked the question on.

The post is here http://www.webworkshop.net/seoforum/...pic.php?t=3015

One of the members dmilford did some test using two made up phrases. One in an <h1> tag and the other in regular text.

The one in the regular text shows up in the searches but the one in the <h> tag does not.

Thanks google junky for taking the time to post here. It may change the way I build and set up websites.

[google junky]

Originally Posted by Janeth

Thanks google junky for taking the time to post here. It may change the way I build and set up websites.

Your'e welcome. Glad to see someone else tested it also. Good luck in all you do.

[greeneagle]

I have just started using the header tags in the last 3 months. So far I have not had an occurance that required that I go beyond H3.

I am not sure whether GOOGLE grants precedence to the “H” tags any more or not. I have not tested this at all. We are in the process of launching our new site and are using 3 levels of “H” tags, so I am also interested in the state of the art here.

If anyone can shed light – let it shine!

Thanks,
Ken

[janeth]

I agree Ken,

What there saying so far is that Google ignores the text in those tags. If that is correct then by putting those tags on your site you could be hurting yourself.

[crankydave]

I have text in H1 tags that is unique to a particular page, is found in only one spot,and it is indexed. Whether or not the tag is given more or less importance is another matter but it is not ignored.

Patron Saint Medals

The text at the top of the page under the row of medals has h1 designation and is indexed and cached.

Dave

[janeth]

Hi Dave,

The problem is all those words are also found in the nav. on the left hand side of the page.

[crankydave]

Perhaps I'm looking at it the wrong way. The precise phrase "Who is your Patron Saint?" is only found in exactly that spot and nowhere else. A search using that phrase in quotes shows that it is indexed.

Dave

[activeco]

Indeed.
It seems H1 tags with defined lower than default size are not the problem.

[janeth]

You are correct Dave

[crankydave]

You are correct Dave

About looking at it the wrong way or about the unique phrase "Who is your patron saint?" within the h1 tag being indexed?

[janeth]

Originally Posted by crankydave

About looking at it the wrong way or about the unique phrase "Who is your patron saint?" within the h1 tag being indexed?

about the unique phrase "Who is your patron saint?" within the h1 tag being indexed? (-;

Thanks, makes me feel better. I will see how the rest of the test come out over at WWS.

[PhilC]

It's not such a well defined example Dave. The words "Patron Saint" appear in the page's Title tag, "who" and "is" are discarded words for the search, the whole phrase "Who is your patron saint?" appears twice - once as link text and again as the title attribute for the link text - both within the H1 tag. The fact that the phrase is used as link text could in itself cause the page to be listed in the serps for the phrase, regardless of the fact that the link is inside an H1 tag.

So, as an example of Google's use of the H1 content, it is not so well defined.

One test that has been done in the last few days is that a unique phrase, that turned up no results, was placed in an H1 tag, and none of the words appeared anywhere else on the page. The page is indexed and comes up as the sole result for other unique phrases on the page, but no results are returned for the unique phrase in the H1 tag.

More tests are underway. One of them is to place a unique phrase (no current results) in an H1 tag, and also include a further instance of each word in the phrase in the body text, but not adjacent to each other. There are other tests too. It should be interesting how they turn out.

[janeth]

I'm very sorry PhilC, I got lazy and did not check any further then on page text.

[Peter (IMC)]

Thanks for telling me about this thread Janeth,.. :)

I don't believe in the logic that a heading tag will be ignored. However, we have to realize something first to understand why I believe that the heading tag is not ignored.

A heading tag is not a stand-alone tag. It may have been in the past, but search engines do get more sophisticated over time and get better at judging the structure of the copy in a page, rather than just looking at the individual parts.

Heading tags are just part of the HTML codes available to structure text.

A search engine will give you higher rankings if you structure text properly. The tags are important, but not all powerfull.

Which brings me to the test that was done. A heading tag is meant as a heading and at least some of the words (but more likely all words) in the heading should show up in the following paragraphs. When it doesn't show up, you can come to 2 conclusions:

First, the heading is unrelated to the rest of the text, and therefore has no meaning.
Second, Google probably has a filter set up for headings that are unrelated to the rest of the page.

I would like to see that test done again but now with some real text in the following paragraphs that uses the heading tag in the copy at least once.

It´s all about good writing and proper structures. Search engines, and especially Google, love the proper use of HTML tags.

[janeth]

Hi Peter,

Glad you made it. (-:

If you look at the thread in my first post, they are still testing it and have a couple more test there doing.

Did you get a chance to look at the other test ?

[Peter (IMC)]

Not sure which test you mean, but perhaps you are refering to:

Indeed.
It seems H1 tags with defined lower than default size are not the problem.

Inside a heading tag there should be no Font tags. It would not surprise me that Google simply ignores these heading tags and treats the content as normal text. This would explain why this text is not filtered out.

And another thing,.. that much text in a heading is abuse of the heading tag. But as strange as it may sound, it is not in Google's interest to penalize the wrong use of HTML code. If they would penalize everybody for wrong HTML their index would fit on the oldest Hard disk that still exists. Not even their own home page would be indexed.. :D

They generally don't give penalties for the wrong use of HTML code, but they simply ignore the wrong use and consider this type of text to be normal text.

[clasione]

[quote="Peter (IMC)"]Thanks for telling me about this thread Janeth,.. :)

I don't believe in the logic that a heading tag will be ignored. However, we have to realize something first to understand why I believe that the heading tag is not ignored.

A heading tag is not a stand-alone tag. It may have been in the past, but search engines do get more sophisticated over time and get better at judging the structure of the copy in a page, rather than just looking at the individual parts.

Heading tags are just part of the HTML codes available to structure text.

quote]

I second Peter's ideas.... It wouldnt make any sence at all for Google to skip the Htags.......

That would be absolutley rediculous, and I really hope the test doesn't prove that is happaning....

That would be a horrible way for Google to attempt to mearly shift things....

[simonm]

I agree with Peter. Google, in placing a page will look at (using automatic processes) the sense of the page based on content and any emphasis on alt tags and headers.

My guess is that google takes into account 1st the paragraph text then treats alt and H tags in a similar manner. This has the benefit for non SEO experts that their pages have a reasonable chance of being placed, ie they are not using alt and H tags. It avoids an over reliance on SEO but also rewards a balanced page.

The lesson being, write your text - that's what google looks at, then add headers and alt tags and all the other stuff to complement that text!

Simon

[PhilC]

I can't agree with the "lesson", Simon, since nothing has been learned yet - as you said, it's your guess ;)

Originally Posted by Peter (IMC)

I would like to see that test done again but now with some real text in the following paragraphs that uses the heading tag in the copy at least once.

That test is already underway. The H1 tag contains a unique phrase that appears nowhere else on the page, and that currently produces no results. Each word in the phrase is used once in the body text, but the words are not adjacent to each other, and they are the wrong way round - the first word in the H1 tag is the last one used in the body text, etc.

[roam_dx]

Hi,

Very interesting subject. Im not much at figuring out SEO, but wanted to test some things I was learning(CSS, and some ideas from this forum). I have put up content from a site I am working on for "experimental purposes" on www.orangeboom.com to see how Google will treat it before I take it live.

The pages have been up for 20 days now. Google has only indexed the homepage and refuses to go further and index the website. I was actually experimenting with the effectiveness of <h> tags and so have used <h2> for the main titles
"Get a Lake Tahoe rental house - fill your life with romance!"

and <h1> for the rest of the titles.
"Book a Lake Tahoe rental house for an unforgettable vacation!"

Both seem to be indexed. But maybe this is being considered spam since there are too many <h1> and few <h2> tags.Should I try switching?? Or are there too many of both. Is there anything like too many <h> tags.

As for the discussion, when I use the site search the content of the <h1> tags does show up.

Wonder why Google is not showing the rest of the pages. any ideas?

Hope this site helps I will be checking in to see what the conclusion here is. Need to know what the problem is before I take this live. Or is it normal for Google to take ages to index a new website?

[janeth]

Hi Peter,

The guys doing the test are here http://www.webworkshop.net/seoforum/...pic.php?t=3015

As you can see there are a lot more test being done.

I know in the past that <h> tags played a role but then meta tags also once played a role.

I also know that adding <h> tags and removing them do not change the ranking of any sites I have played with.

SO far from what we have seen it does appear that if a word only appears in the <h> tag and no where else that Google does ignore it. But at the same time if it appears in the text Google does not ignore it.

[/quote]

[crankydave]

SO far from what we have seen it does appear that if a word only appears in the <h> tag and no where else that Google does ignore it. But at the same time if it appears in the text Google does not ignore it.

Which would make sense. If a <h> tag were to appear on a blank page or be completely unrelated to anything on the page it should be ignored. Shouldn't the h1 be setting up or describing the page? Why would anyone want to use the tag on something completely unrelated to the page? Google may not be able to filter out unrelated text on a page but might be able to filter out unrelated <h> tags.

Dave

[crankydave]

And another thing,.. that much text in a heading is abuse of the heading tag.

Perhaps it might be Peter. I think it appropriately describes, to my visitors, as the heading, the page and what to expect as they navigate through the sections.

If I were to place <h1>Hurricane Ivan<h1> on the page, this would be abuse and should be ignored.

Dave

[Peter (IMC)]

The pages have been up for 20 days now. Google has only indexed the homepage and refuses to go further and index the website. I was actually experimenting with the effectiveness of <h> tags and so have used <h2> for the main titles
"Get a Lake Tahoe rental house - fill your life with romance!"

and <h1> for the rest of the titles.
"Book a Lake Tahoe rental house for an unforgettable vacation!"

Both seem to be indexed. But maybe this is being considered spam since there are too many <h1> and few <h2> tags.Should I try switching?? Or are there too many of both. Is there anything like too many <h> tags.

You should start with an <H1> and then use a <H2>, etc. Don't skip a level. That´s what w3c says.

I also know that adding <h> tags and removing them do not change the ranking of any sites I have played with.

SO far from what we have seen it does appear that if a word only appears in the <h> tag and no where else that Google does ignore it. But at the same time if it appears in the text Google does not ignore it.

I think what is important is to realize that it is about the combination of all the parts, rather than just about the <h> tag.

The conclusion you make here is direct evidence of the importance of text structure. Here is an interesting page about text structure: The 5 paragraph Essay. If you know how a 5 paragraph essay is structured, you will understand better how Google looks at text.

Taking out the <h> tags but leaving the text in the page, doesn't make that much of a difference. The structure is the same, though it is a little less clear to a search engine. But, the difference isn't so big that it would radically change the positions.

[Peter (IMC)]

Perhaps it might be Peter. I think it appropriately describes, to my visitors, as the heading, the page and what to expect as they navigate through the sections.

I am sure your intentions were good, but you´re addapting the logic to fit to your benefit,.. :)

The heading tag is meant as an heading and not as a paragraph. So if you have enough text to create a paragraph you should use the

tags.

You use way too many <h> tags in that page. The best way to go about it is to put "14K Gold Patron Saint Medals" in <h1> tags. The paragraph in between

tags and that´s it. No need to put the copyright information in between <h4> tags.

By the way,. there are a lot of empty <h3></h3> tags in in your code. you may want to clean that up.

[crankydave]

Thank you Peter.

My problem is the template design from my hosting company. It, for whatever reason, automatically place the h tags, except for the h1 which I placed. I cannot, for example, place the page title "14K Gold Patron Saint Medals" within the h1 tag. Often, trying to make corrections the "page" disallows them and goes back to whence it came.

Changing host, at this time of year, is not an option. Making due whith what I have to work with is.

I do appreciate your suggestions very much.

Dave

[PhilC]

There is often a tendancy to think of Google as a person. It's often seen when people recommend not having much Javascript in the head section of a page, because it pushes the body text, with the target searchterms, further down the page, and it is thought that the higher up the page the searchterms are, the better it is for rankings. But Google isn't a person - it a set of programs.

Programs have no problem seperating the body text from everything else on the page, and the Javascript idea is almost certainly a mistake. People may find the 5 paragraph essay idea to be good, but a program doesn't see it in the same way. Programs are not that sophisticated.

Creating 'good' pages for people does not help rankings. Programs are very crude, they have no intelligence, and they can only do certain things. All of Google's main patents are very crude. PageRank (not their own patent) is a cold calculation that has nothing to do with intelligence. Their expert system and Topic Sensitive PageRank are the same - take this, extract that, compare to something else, etc. and print what you're left with - no intelligence. So the idea that creating pages that are good for people will help the rankings doesn't hold up. To improve rankings, it is necessary to understand the cold, unintellgent, elements. And, if they change, we need to keep abreast of the changes.

[Peter (IMC)]

Originally Posted by PhilC

There is often a tendancy to think of Google as a person. It's often seen when people recommend not having much Javascript in the head section of a page, because it pushes the body text, with the target searchterms, further down the page, and it is thought that the higher up the page the searchterms are, the better it is for rankings. But Google isn't a person - it a set of programs.

Programs have no problem seperating the body text from everything else on the page, and the Javascript idea is almost certainly a mistake. People may find the 5 paragraph essay idea to be good, but a program doesn't see it in the same way. Programs are not that sophisticated.

Creating 'good' pages for people does not help rankings. Programs are very crude, they have no intelligence, and they can only do certain things. All of Google's main patents are very crude. PageRank (not their own patent) is a cold calculation that has nothing to do with intelligence. Their expert system and Topic Sensitive PageRank are the same - take this, extract that, compare to something else, etc. and print what you're left with - no intelligence. So the idea that creating pages that are good for people will help the rankings doesn't hold up. To improve rankings, it is necessary to understand the cold, unintellgent, elements. And, if they change, we need to keep abreast of the changes.

I´m sorry, but even if that was true it is still the best way to write and structure copy because very simply, once the visitor does get to your website, you want them to buy something as well.

I don't agree at all with what you are saying. The 5 paragraph essay is just an example and obviously there are many other ways to structure copy. But lets use the 5 paragraph essay as an example:

It tells you to make the point in the first paragraph, and to describe all subtopics you want to use in the next 3 paragraphs. After that in the 5th paragraph it tells you to repeat the 3 subtopics and make the point again.

So what does this structure result in?

It results in the keywords being used in the first paragraph,.. in the middle paragraphs, and the last paragraphs. So which program is not able to find the common keywords in this structure? Esepcially when a heading is used it gets even easier for a program to determine what this text is about.

Then we have a text that wasn't written very well structured, but just starts somewhere and ends somewhere.... In this situation a programm is going to have a much harder time figuring out what the main keywords are in this page.

The best way to optimize for high rankings is to optimize for your visitors. That´s really true, but you do need to know what your visitors want and that is probably more difficult than figuring out what a search engine wants.

However, I have noticed often that websites improve a lot for the visitors when they are being optimized for the search engines.

[janeth]

From the test so far Google did find the <h> tags but ranked it under the page that only used plain text.

[crankydave]

Janeth,

I caught that this morning. Seemed to be a lag time in finding the h1 h2 h3 vs. the plain text. Was this the page with the titles only and no content? I think it would be interesting to see how...

1. A page of content with related <h1>title</h1>

Ranks against....

2. Identical page of content with related text only title.

Dave

[Peter (IMC)]

From the test so far Google did find the <h> tags but ranked it under the page that only used plain text.

I don't want to be the one that is being a pain in the a$$, but there is another thing that needs to be taken into account:

<head>
<meta http-equiv="Content-Language" content="en-gb">
<meta http-equiv="Content-Type" content="text/html; charset=windows-1252">
<title>test</title>
</head>

The title of this page is "test". Please don't forget that the title is the most important tag in a page. And in this test, the title "test" may be interfering with what you want to investigate.

All goes to show that you can't really say anything about an individual tag. It´s about how everything works together. You will notice that while doing these tests, every time you need to add something to make sure that it didn't interfere with the previous test(s). You will end up with the conclusion that <h> tags will work best if it matches the title, it´s in the meta keyword tag, it´s in the meta description tag, it shows up in the copy, etc. etc. etc.

The word I always use is "Harmony". All factors need to be in Harmoney. That´s what works best! In my opinion, the best SEO consultants are the ones that know how to bring Harmony to a page/website. And I noticed that many people have difficulties with bringing Harmony to their pages.
It may not always get you position 1 results, but it will get you many high rankings and even more important: success for the website.

[janeth]

Hi Peter,

There still testing to see what works and what does not work.

You have to take it one step at a time so we can see how it all works together.

When we add the titles they need to be added to both test pages with and with out <h> tags.

It maybe that the page without the <h> tags will alsways rank above the page with it but until we get all the test done it is hard to say.

the best SEO consultants are the ones that know how to bring Harmony to a page/website.

The best SEO's are the one's that test to see what works and what does not work and change there sites accordingly. (-;

[Peter (IMC)]

The best SEO's are the one's that test to see what works and what does not work and change there sites accordingly. (-;

hmm,... (-; I agree that it is important to know what works and what doesn't. But it´s not an exact science. Lately I have seen some threads about people having high rankings, but no success, even if they do get lots of visitors.

The tests that are being done now are interesting because it shows that the <h> tag is not a stand alone tag.

When we add the titles they need to be added to both test pages with and with out <h> tags.

What I meant to say was that the results of the tests so far may be interpreted wrongly simply because the title of the page is "test". Do we know if an <h> tag gets ignored if the title says something totally different? It´s just a question, but could be fundamental in this research.

Perhaps it would be better to have no title at all.

[janeth]

Originally Posted by Peter (IMC)

hmm,... (-; I agree that it is important to know what works and what doesn't. But it´s not an exact science. Lately I have seen some threads about people having high rankings, but no success, even if they do get lots of visitors.

We both agree here. It takes a lot more then a number one ranking to make a site work and the things we're talking about are very small. But it is just one part of the big picture.

Originally Posted by Peter (IMC)

The tests that are being done now are interesting because it shows that the <h> tag is not a stand alone tag.

What I meant to say was that the results of the tests so far may be interpreted wrongly simply because the title of the page is "test". Do we know if an <h> tag gets ignored if the title says something totally different? It´s just a question, but could be fundamental in this research.

Perhaps it would be better to have no title at all.

There are a lot of things that need to be tested before we decide what works best. (-;

[PhilC]

Peter, we are talking solely about the current effect that the H1 tag has on Goole rankings. We are not talking about keeping visitors happy or about making sales. They are important, of course, but that's not the subject of this discussion. Unless a page is ranked highly for its target searchterm, there won't be anyone to keep happy or to sell things to.

I'm sorry Peter, but you are mistaken. Your 5 paragraph structure does not naturally result in what you described, unless you take specific steps to make it do so.

Originally Posted by Peter

So which program is not able to find the common keywords in this structure? Especially when a heading is used it gets even easier for a program to determine what this text is about.
....
....
In this situation a programm is going to have a much harder time figuring out what the main keywords are in this page.

You are still making the mistake of thinking that programs are like people. Firstly, search engines programs do not determine "common" or "main" keywords in the page. That's not the way they work. Secondly, they don't determine what the "text is about". That's not the way they work either. People do that; search engine programs don't. So your conclusions are based on false assumptions.

Your "harmony" idea is sound *if* you mean harmonizing all the elements on the page to match a search engine's algorithm (it's called optimization). But before we can optimize/harmonize for search engines, we need to know which elements are being used in the algos. We pretty much know that answers to that, but we still need to keep up to date with them and their effects - hence this thread about H1 tags and Google.

Incidentally, there's another recent thread here that talks about Microsoft's spam doorway pages - and they really are spam doorway pages. If you've seen one you'll know that they don't even begin to resemble a 5 paragraph structure, or any kind of structure that is useful to visitors, and yet they occupying a stack of #1 rankings in Google. They are just a drop in the ocean of top ranked pages that are like them or very similar.

Getting top rankings in any search engine means matching its algorithm as closely as possible. It's a cold, clinical thing, and has nothing to do with satisfying visitors, or making nice, user-friendly, well layed out pages. It's necessary to combine the two, of course, unless you want to spam, but getting top rankings requires cold, clinical algo-matching. Nice, user-friendly, well layed out pages, including 5 paragraph style pages, will not get top rankings, except perhaps for uncompetitive phrases - by accident; optimization will.

[Peter (IMC)]

You are still making the mistake of thinking that programs are like people. Firstly, search engines programs do not determine "common" or "main" keywords in the page. That's not the way they work. Secondly, they don't determine what the "text is about". That's not the way they work either. People do that; search engine programs don't. So your conclusions are based on false assumptions.

I don't think that programs are like people. I really don't!!! Though I am not a programmer, I did learn some languages like Pascal, C++, and machine language. I also forgot most of what I learned quickly as programming really isn't my thing.
But I do know what programs can and can't do.

I can ask a program to find which word or group of words (phrase) is used most often. Which (group of) word(s) is marked bold or strong, which words in the heading are repeated in the text most, which of these words are used in the last paragraphs,... etc. etc. etc. A program can perform all these functions without any problems. Then you can place a set of rules on how to use all the information that the program gathers. Give this program a text and in most cases it will tell you the main keyword phrase for this text.

Perhaps I am crazy, but in my opinion, this program will perform better with texts that are structured and written well.

Indeed it is just a "cold, clinical thing".

What about taking the title "test" out of the test so that we don't have to worry about interference of the title?

[PhilC]

Give this program a text and in most cases it will tell you the main keyword phrase for this text.

Yes, programs can do what you said, but search engine programs don't.

Search engines don't "find which word or group of words (phrase) is used most often". Pages are not classified as being about certain things. It's a mistake to think that they are. Search engines are simply not up to that sort of level yet.

Perhaps I am crazy, but in my opinion, this program will perform better with texts that are structured and written well.

I disagree with your opinion, Peter, but it doesn't matter because search engine don't work in the way that you seem to think they do.

The test you are talking about isn't mine, so I can't modify the Title. But it wouldn't make any difference anyway. Each word in the phrase in the H1 tag is unique in the page. They exist on the page and that's all that matters. The page should have come up in in the serps for the H1 phrase, but it didn't. It does now though.

If you start removing page elements just because those elements are used in the algo, and might just interfere, where do you stop? It is known from tests that a page will show in the serps for the phrase in the H1 tag, when none of the words in the phrase are in the Title. So the Title doesn't interfere, and there's no need to take it into consideration.

I think that somebody in this thread said that seo isn't an exact science - but it is. It's a *very* exact science. If one person knows the algo, s/he'd be able to create #1 ranking pages every time, regardless of competitiveness. We're not going to know the algo, but we try to learn what we can.

[Peter (IMC)]

Search engines don't "find which word or group of words (phrase) is used most often". Pages are not classified as being about certain things. It's a mistake to think that they are. Search engines are simply not up to that sort of level yet.

No? then what about adwords and adsense? I add a new page to one of my websites, and within minutes, sometimes seconds, the page is crawled and relevant ads show up next to the content of the page. Somehow they figure out what the page is about and here too,.. the better the text is written, the more relevant the ads are.

And then we have Topical Sensitive PageRank,.. though this may not be fully implemented yet,.. it´s another indication of "knowing what a page is about".

With all do respect, but I believe you are underestimating what a search engine like Google is capable of.

Pages are not classified as being about certain things.

This I agree with.

There is a way of talking about how something behaves and a way of talking about how something does what it does. It´s not always necessary to know exactly what the inside of a box does, as long as you know how it interacts with the world arround it.

But lets not get into a discussion about how search engines work. I prefer to talk about how they behave. That is all you can really investigate, and that is what the test is trying to determine.

If you start removing page elements just because those elements are used in the algo, and might just interfere, where do you stop? It is known from tests that a page will show in the serps for the phrase in the H1 tag, when none of the words in the phrase are in the Title. So the Title doesn't interfere, and there's no need to take it into consideration.

Well you would have to stop at the point that you´re sure the results of the test aren't influenced by other factors than what you wanted to test.

But you made a good point. The title does not interfere when trying to determine if words in an <h> tag will show up in a SERP.

I think that somebody in this thread said that seo isn't an exact science - but it is. It's a *very* exact science. If one person knows the algo, s/he'd be able to create #1 ranking pages every time, regardless of competitiveness. We're not going to know the algo, but we try to learn what we can.

Yep,. :) That was me. And I still say that SEO is not an exact science because for a big part search engine rankings depend on the other websites in the results. You have to be better than them (to the search engine). Even if you would know exactly what the algorithm wants, you would have to make compromises. You can be certain that an algorithm gives the best rankings to those sites that have the best combination of all factors that it looks at. The best combination,,... and you would have to combine that with something called "text". and especially in Google,.. if you know exactly how the algorithm works,.. you would still need to get all those backlinks,.. and that isn't that easy either.

Cheers,

[mikmik]

Search engines don't "find which word or group of words (phrase) is used most often". Pages are not classified as being about certain things. It's a mistake to think that they are. Search engines are simply not up to that sort of level yet

So therefore, using keywords is a myth???

That statement doesn't make sense to me in the slightest.

How can over use of keywords, or under use make any difference then? Then there could be no way to tell spamming of keywords!

Counting words is about the easiest thing any program can do, even my simple text editors count and rank words.

[Mel]

No keywords are not a myth, but they are not much about search engines either. As far as search engines are concerned every word on the page (except possibly stop words) is a keyword.

The important thing about keywords and phrases is to find out which are being used the most by real live searchers, and then optimize your page to rank highly for them.

While we may never know fully the details of how search engines rank pages, Sergy and Page have given us a very nice inside peek into how the Google parses pages and how the parsed results are used:

First the pages are parsed and as a result of the parsing the words found are stored in word barrels with information on the position on the page, formtting, if it is anchor text or not,etc and if I remember correctly they assign a decreasing weight to each successive repetition of each word so that eventually more of the same words do very little good.

When a search is done only the data from the word barrels corresponding to the words in the search term is used, and at that only the first 40,000 results or so are actually used in the ranking process (and only the top 1000 ranked results displayed).

So the algo is not using the full text of the page to rank the page, and therefor there are limitations as to what the ranking algo can do.

Peter

Adwords and Adsense are not search engines, they are simply advertising programs that determine the "theme" or topic of a page in order to display relevant ads on that page.

[mikmik]

Thanks, MeL! Everything makes more sense in light of that understanding, a lot.

This whole thread is very interesting and informative for me you guys and gals :O)

[PhilC]

AdWords and AdSense are nothing to do with the search engine. They have CIRCA technology for those, but it isn't used in searches.

And then we have Topical Sensitive PageRank,.. though this may not be fully implemented yet

What makes you think that it's even partially implemented - or that it will ever be implemented? I have seen no evidence for it. It's a patent, and without evdidence that it's actually being used, it is only a patent.

There's a *huge* difference between discerning a page's topic in order to place a few ads on the page, and discerning it to compile the organic rankings. They are not even slightly similar.

In describing why you say that seo is not an exact science, you described why it *is* one, Peter. The ranking algo is an unintelligent program that takes many factors into account when scoring pages - including off-site factors. The serps are ranked according to cold calculations and not by any sort of 'judgement'. That makes seo an exact science. Those backlinks that you mentioned are part of knowing how the algo works. It's all very exact. Programs cannot work any other way - not even CIRCA.

[oceanmajk]

I don't understand how google would benefit improving a "relative" search by excluding pages structured according to industry standards. Penalizing someone for using regular html tags such as H1 seems rediculous to me.

Everything I've read up to this point about optimizing your pages emphasizes how important putting your keywords in header tags. I've even went as far as to include header tags in my pages in random places, but changing the look of them in my style sheet so the text doesnt look out of place. And I use H1 through H3 all over the place.

But it does raise a good question, because only a few of my pages for flaloans.net are indexed in google, but pretty much all of my pages are indexed in yahoo.

[Peter (IMC)]

How do SERPs behave?

Is it possible to predict what a SERP will look like tomorrow, next week, next month, next year?

This answer to this question is that the further in the future you want to predict what a SERP looks like, the less accurate the prediction will be. This is literally the same as with predicting the weather.

Even though the weather is governed by the laws of nature, which are known, there is no way of exactly predicting the weather. The further in the future you want to predict, the less accurate the prediction will be.

So how come that even though we know all the laws of nature, it isn't possible to predict the weather accurately? The answer is that the weather is a dynamic system which basically means it is governed by chaos theory.

This same principle applies to the SERPs. A search engine is a dynamic system and even though they are governed by defined and exact rules, the future of any SERP can not be predicted. Due to the inmense amount of variables and inputs it simply is an chaotic system. The existence or non existence of one single link for example can have a huge effect on a SERP in the future.

SEO is not an exact science. If it was, it would be possible to tell what the position of a webpage is tomorrow, next week, next month, next year, etc.

Even if you know exactly how a search engine works, that still doesn't guarantee a position 1 result.

But that doesn't mean we don't want to do tests of course,. :)

I hope that the tests that we discuss in this thread are targeting to determine the best way to use the <h> tags. Though common sense pretty much tells what the best way to use it is, it is good to know where the bounderies are as this allows for a bit more aggressive SEO.

[PhilC]

A search engine is a closed system, Peter, and the laws of chaos do not apply to closed systems. I'll tell you now that, if you bring string theory or branes into it, I'm outa here ;)

Search engines are not dynamic. They are programs that do not change in any way, except by human intervention. Programs do not obey the laws of chaos - the memory in which they reside may do so, but not the programs. In the timescale that we are talking about, they are completely fixed - and exact.

We don't do seo for what might be in the future; we do it for what *is* right now. It's now when we want high rankings, and it's what the programs are doing right now that we are interested in. We can look to the future with things like getting IBLs from on-topic pages, just in case it makes a difference in the future, but, even then, we are still dealing with now because those IBLs do just as good a job as any other IBLs, and we need IBLs. But if something like the H tag changes in value, because of human intervention in the program ;), we want to know - because it's fixed now, and it won't change until a human changes it.

At any given time, seo is only about cold, hard calculations, and nothing else. It's exact, and not dynamic.

Even if you know exactly how a search engine works, that still doesn't guarantee a position 1 result.

It does if you're the only one with the information. You can't fail.

Btw, the tests are being done elsewhere and probably won't be reported on here.

[Peter (IMC)]

A closed system is a system without input. A search engine has lots of input. Spiders crawl the web for as much imput as they can find.

You are too much focused on the inside of the box and forget how the outside world affects the output of the box. The output of the box is not just determined by the inside of the box, but also by the input. This is fundamental.

Quote:
Even if you know exactly how a search engine works, that still doesn't guarantee a position 1 result.

It does if you're the only one with the information. You can't fail.

Come on,. you know very well that you need external factors to get high rankings. (External factors are inputs to the box.)

If you know exactly how a search engine works you still need to get all those backlinks. Knowing how the search engine works is no guarantee for getting all those backlinks and thus this knowledge is no guarantee for high rankings.

But I understand what you want to say: "One Eye is king in the land of the blind." :)

[Peter (IMC)]

oh,. I forgot,.. a search engine it self (the box) is not a dynamic system, but the search engine is part of a dynamic system. The search engine is nothing more than "the laws of search engine nature". We can't see these laws, but we can see websites and SERPs (input and output).

[beanstalk]

I ran some checks accross a few sites with some unique H1 tags and found that most were in fact read by Google.

If I had to hazard a guess as to how Google treats them it would be as a starter point of sorts. An indicator that the "real" content starts here.

I doubt as though any significant weight is being given to them, what with ability to use and abuse the tag with CSS.

[Peter (IMC)]

I doubt as though any significant weight is being given to them, what with ability to use and abuse the tag with CSS.

Though the looks of an <h> tag can be changed with CSS, there are many factors that can show abuse. For example:

number of words used
placed in between other tags
location compared to other tags

These are all factors that are important. For example, hiding a heading by making them inline and look like the normal text still wouldn't work because <h> tags are not suppose to be placed in between

tags. And if you wouldn't use

tags so that the heading wouldn't be in between

tags, that wouldn't help much either.

[PhilC]

I meant that the program is a closed system - and it is. But we're not talking about chaos; we're talking about search engines.

Peter, if I was the only one who knew the algo, I'd get #1 ranked pages every time, because rankings are decided by an exact calculation - there's nothing dynamic about it. IBLs are part of the algo, and I'd know exactly what's needed. I said that the algo includes those external factors earlier.

To be honest, the "box" and "the laws of search engine nature" (there aren't any) don't interest me, because they are not important to seo. What is important to seo is knowing the algo as well as possible. For instance, if I knew it, I'd know exactly how many words to put in a page, exactly how many instances of the target searchterm, where to place them, what attributes to give them, how many IBLs are needed, what link text to use, etc. etc. etc. If something is considered more important because of something else, I'd know it and match it 100%. It's an exact calculation - nothing more. As long as programs are deciding the rankings, it cannot be anything other than a cold calculation. CIRCA uses a cold calculation to determine page topics. There's nothing dynamic in search engines, or in any other computer program.

How H tags are treated is exact. People, including me, are currently trying to determine whether or not there has been a change in their value.