Keyword Density

[therapists]

Keyword Density, I know is a very sore point here on WebProWorld.

I have written some marketing pages with a keyword density of around 4-5. However, once I add the footer onto the page, I suddenly have keyword densities of a 9-10%.

I use the keyword cloud checker on webconfs (Keyword Density Checker - Keyword Cloud). They take out 2 letter words and removes stop words.

I expect contrasting opinions here, but what keyword density should I be looking for - should it matter that the density flies up when I add the footer (the site is built in asp), and what keyword density tools does anybody else use??

Thanks!!

[wige]

The first concern I have, going by the numbers you posted, is how much content is on the page vs how much content is in the footer. If the footer is adding that much to the keyword density it seems to me that the page either contains only a few paragraphs, or the footer has a lot of links that are being added. If this is the case, that the footer is so heavy it overpowers your content that may introduce problems of its own.

[suzstephens]

I've been watching a site that ranks #1 for real estate in a major U.S. market for a year or more. It contains about 80 repetitions of variations on the primary keyphrase for that area. Here's how it measures up for the primary keywords:

atlanta 60 13.86%
homes 35 8.08%
real 26 6%
estate 14 3.23%

[stuart888]

Once you get to the point of spending time worrying about keyword density, you are at the point Google warns you about. "Would you do it if search engines did not exist?" You are much better off using your valuable time focusing on your customers and making them happier, so they will self promote your web site's value.

Quality content is worth spending time on, keyword density is not.

[craigmn3]

The common thought that keyword density should be 3%-5% and your word count should be considerable. Keyword density should only really be considered as part of your "anti-spam" portion of SEO....you don't want to find yourself suddenly in the spam closet because you didn't check your density.
Another reason not to get to concerned with keyword density is simply there is not one tool that measures it the same. I use 4 -5 different tools (From Web CEO, Submit Express, SEO Tool Box and others) and never get the same answer. Just keep the keyword percentage out of the attic and concetrate on writting good copy.

The reason the search engines consider this, is that in good natural writting your topic keywords fall in about the three- five percentile range
-------------------------------

There are apx 127 words in the above post and the word (keyword) appears 5 times.
that's about a 4% keyword density

[williemanillie]

I don't know if this has too much weight to it, but most the articles I have read state around 4% to play it safe. Although I am not a SEO pro by trade, my site does rank on the first page.

[GoogleRankings]

Umm where did Dan Theis and my posts go? Don't tell me they were moderated

[DanThies]

That'll teach me for spending time here... and although they don't want me to link to a detailed explanation, there's still no such thing as keyword density.

[bj]

I know I'm gonna get beat up for this, but worrying about keyword density is the wrong thing to focus on.

I check my stats periodically and see what natural keyphrases are being searched repeatedly and just blog using those as topics. Great way to get blogposts and to nail your keyphrases in a more natural way. Another thing I do using that keyphrase list in my stats is to keep it in mind while I blog, and after writing any article I'll simply look through and compare what I wrote to my keyphrase list. Then I'll change a few things based on that-- like "site design" to "website design". I never worry much about frequency. I just write naturally, and then tweak just a bit. It's worked well for me. Google says to write for your visitors. That's what I do (with just a BIT of SEO tweak to help things along!)

[DanThies]

BJ, that's exactly the right approach. Your copy has to work as copy - any SEO "problems" that creates (unlikely) can be fixed without stuffing keywords.

[GoogleRankings]

Originally Posted by DanThies

That'll teach me for spending time here... and although they don't want me to link to a detailed explanation, there's still no such thing as keyword density.

Majority rules, right or wrong means nothing in that kind of situation. I have a feeling many of these members sit on the forums so long they don't notice Google has changed quite a bit since the began...


I don't know about you guys but on the forums I moderate when we remove posts we either alert the member, PM or warn them. I received none of the above.
No worries I won't spend anymore time here - That saves you a heck of a lot of time deleting my posts

[sanver]

It is better to focus on delivering right communication to your target customers than getting into things like keyword density.

I am not a "SEO Professionsal", hence I may sound simplistic.

Sanjay Verma
<snip links CD>
Mumbai, India

[DanThies]

That's a very professional attitude, Sanjay. You've solved one of the great mysteries of SEO without even trying.

[incrediblehelp]

Ignore "keyword density" and write for you visitors and your topic.

[Webnauts]

I think Jill wrote something nice about the whole issue. Worth to check it out: Revisited: The Art of SEO » HRA - Jill Whalen

[therapists]

The content I have written is around 4%. However, the site is for overseas property and within the side navs and footer are links to property categories and different properties, all with Property in Bulgaria, Property in Germany etc..

After I get the content into the page, the % density of the word property shoots from around 4% to up to 12%!

The content reads well, and is not stuffed - but will all the navigation on the page make it look like spam?

[Traffic Master]

You all say keyword density is not important and should be ignored. What if your content is so filled with your keyword that the Search Engines suspect you of spamming. I believe that ideally it should be between 3 - 8%.

[live_Tech1]

I have to agree with Stuart888. After proofing the site pages for my organization, I found that my web developer focused more on keyword repetition, than on content. The wording on the site was incoherent and I had to basically tweak the english to make it readable.

In the great quest for optimization, we often lose sight of the purpose, to present our respective organizations as a viable option for potential visitors.

It's seeing the forest through the trees.

Live_Tech 1
Live Online Tech Support

[jtracking]

Originally Posted by therapists

Keyword Density, I know is a very sore point here on WebProWorld.

I have written some marketing pages with a keyword density of around 4-5. However, once I add the footer onto the page, I suddenly have keyword densities of a 9-10%.

I use the keyword cloud checker on webconfs (Keyword Density Checker - Keyword Cloud). They take out 2 letter words and removes stop words.

I expect contrasting opinions here, but what keyword density should I be looking for - should it matter that the density flies up when I add the footer (the site is built in asp), and what keyword density tools does anybody else use??

Thanks!!

Just make regular sounding sentences and regaular sounding/looking paragraphs with your keywords in them. maybe add a few more keywords here and there in those sentence/paragraphs but that normalicy should do you well...

you're not going to get much higher in the serps if you stray too far from this concept. It's just normal speak for the visitors and combine that with good links to your site and again, you should do well...

my take...

[jtracking]

Originally Posted by Traffic Master

You all say keyword density is not important and should be ignored. What if your content is so filled with your keyword that the Search Engines suspect you of spamming. I believe that ideally it should be between 3 - 8%.

I hear ya there but that assuming that you want to use tools to help you spam "properly" or "safely" but if you make normal sentences and normal paragraphs with normal copy with your keywords in them then you shouldn't have to use those fricken tools.

those tools help you fine line your copy but who wants to fine line when you can just offer honest text? not me!

anyways,

do a search for mississauga maid service

there is a doman (not the first one) but about #8 down the list...its totalhomecarecleaning .com - select their whole page (CTRL A) and notice the spammy keywords on the right...they've got a HUGE list of "same color as background text" and not only did they copy my site but they also are still getting ranked and displayed on the first page with yahoo and google (NOT MSN) even though they've been reported to both engines...

does it matter? spam all you want spammers...

[wige]

After reading through the comments here, I went back and took a look at some of the pages on my primary site, checking for the first time the keyword densities of some of my more common entry pages, and saw that I had keyword densities of 45-60%. These pages appear in the first page of the Google SERPS, and consist of about five paragraphs of text, with the keyword in question distributed throughout (probably 4-7% in those parts of the page) combined with a giant table that repeats the keyword 500 times, along with other data on each line (please keep flames about how unwise that decision was to a minimum, my users like the tables). Every keyword density tool I have used has shown differing densities: some include the table, some ignore it, some look at the meta tags, others don't, etc. All I know is that I am not arguing with top 5 placement for these pages. I haven't a clue what the limit is, or what would trip the spam filters, all I can report are my own personal experiences.

[GoogleRankings]

Originally Posted by wige

After reading through the comments here, I went back and took a look at some of the pages on my primary site, checking for the first time the keyword densities of some of my more common entry pages, and saw that I had keyword densities of 45-60%. These pages appear in the first page of the Google SERPS, and consist of about five paragraphs of text, with the keyword in question distributed throughout (probably 4-7% in those parts of the page) combined with a giant table that repeats the keyword 500 times, along with other data on each line (please keep flames about how unwise that decision was to a minimum, my users like the tables). Every keyword density tool I have used has shown differing densities: some include the table, some ignore it, some look at the meta tags, others don't, etc. All I know is that I am not arguing with top 5 placement for these pages. I haven't a clue what the limit is, or what would trip the spam filters, all I can report are my own personal experiences.

That is a silly reply... you are assuming that because you rank in the top ten that your keyword density somehow had an effect on that... This is the response I have gotten from clients "why should I fix my url structure? I rank well for many pages, it must not be a problem" - that is a very flawed idea of things.

There are many sites that rank well but are still held back by not being completely SEF, because they are penalized etc.... now I don't believe keyword density spamming will get you penalized by its self but I do know that if your focusing on keyword density you are very likely to lose sight on your goals...

[wige]

Originally Posted by GoogleRankings

you are assuming that because you rank in the top ten that your keyword density somehow had an effect on that...

Sorry, what I meant to imply was that because I am in the top ten for my competitive keywords, and have a much higher than average keyword density, I have not been penalized for it. I know pretty well what pushed me onto the first page, which was a coordinated off page optimization effort several months after the last time I edited the text on the particular page, resulting in a jump from page three to page one.

I am definitely not trying to imply that loading my page with keywords helped my ranking. I do not believe that to be the case at all. I am simply saying that in my case it did not have an adverse effect. When designing the page, I considered the risk of repeating the keywords in the table, but I saw that many competitors did similar things on their own page with no ill effects, so I gave it a try as it made the tables more clear to my users.

[mjtaylor]

A question about keyword density doesn't imply a focus on keyword density; it's just a question about keyword density. (Teehee, this reminds me of our debates about PR.)

And the question wasn't whether Keyword Density is worth paying attention to or not ... the question seems to be: if I am going to pay attention to keyword density what should I aim for?

Craig is right 3-5% *is* the rule of thumb ... however, I have noticed that on some competitive terms that a higher density seems to found on the top 10 pages ... it's as if the top sites keep optimizing a little more for the term in an effort to outdo one another ...

I would say look at the sites on the first page of SERPs related to your terms and you will see what the SEs tolerate for KW density on a given term.

And Craig is also right, I believe, that it is important to focus on keyword density at least long enough to make sure you aren't *over* doing it ...

My two cents on a hot topic!

MJ (edited for typos)

[crankydave]

I didn't read wige's reply the same way GoogleRankings.

Actually, I think it supports the notion the keyord density unto itself, really doesn't tell you much.

All these tools really do is provide you with a percentage of how a particular word or phrase relates to the text as a whole. Many times, they don't even tell you what parts of the page and or code it is taking into account.

Most importantly, they don't tell you what kind of effect, if any at all, the phrases they take into account have on anything, let alone rankings.

As an example, let's say you add a paragraph of text to your page that changes what percentage one of the density "tools" gives you and your rankings drop.

Is that drop because the density changed? Because the keywords moved position on the page? Because the relationship between the keyword and other text changed? Because the positioning of text around the keyword moved? Because the competition outperformed you? Etc. Etc. Etc.

Several folks have already said it. Simply make your text read and flow naturally. Don't stuff keywords into text simply to raise that percentage and don't think that you should rank #1 for a phrase because you used it once in 700 words of copy.

Dave

[DanThies]

But MJ... keyword density isn't even measured by search engines. It's a complete myth. All this talk about "magic numbers" or rules of thumb... nobody's even saying which method they're using to come up with those numbers.

[DanThies]

Here's how keyword density gets silly:

Let's say you want to rank for [purple widgets]. You write a page that has 100 words on it.

The word purple appears 4 times.
The word widgets appears 3 times.
The phrase "purple widgets" appears once.

What's your "keyword density?" - how do you measure it?
a) 1%, because the phrase only appears once in 100 words.
b) 2%, because it's a two word phrase once in 100 words.
c) 7%, because 7 of the 100 words are keywords

If you answered a, how do you deal with a 10 word phrase's "density?" If it appears once, you have 1% density, if the page is just that phrase repeated 10 times, you're only at 10%.

If you answered b, then simply adding one word to your phrase increases your "density" to 3% even though you only used the phrase once.

If you answered c, your numbers will get higher very quickly as your target phrase gets longer.

So what's the optimal keyword density?
d) there's no such thing - search engines don't measure it, for good reason

[mjtaylor]

Well, you may be right ... perhaps SEs don't measure it; perhaps it is a myth, but I do notice patterns in keyword density when I use IBP to analyze sites ... and it can't hurt to look at the top sites and not exceed their density.

Now, it's been years since this happened ... 4 or 5 ... it was somewhere along the time of the Florida Update ... and I got a little optimization happy on a particular site ... I won't say I stuffed keywords, but I just put them everywhere ... okay it was kinds stuffing: h1s, h2s, h3s, bold, italic, alts, file names, titles, ad nauseum, and the site slid down to #7 ... I reduced the overdone phrase and voila, back to #3 where it remains today .... perhaps that was not a keyword density filter, per se, but it sure quacked like one.

MJ

[DanThies]

MJ, you can absolutely overuse keywords in your copy... but density is not real. As an SEO copywriter, you might find it helpful to read up a bit on how search engines really work. If anything, it will reinforce the value of writing good copy.

[mjtaylor]

Well, thanks, Dan. I am a little more than an SEO copywriter, and I have read a good deal, including Google's patent applications, etc. but I am always grateful good SEO advice.

And no, I don't recall anything about keyword density in what I read from the horse's mouth ...

Still, I am not sure how you can say that it is possible to overuse keywords and that 'keyword density' is not real. Sorry, but that just seems flat out contradictory to me. What's the difference between keyword overuse and excessive keyword density? Semantics.

I agree that keyword density is not a magic bullet to get your site to the top and that it should not be a major focus. I will allow that SEs probably don't measure it in the simplified manner that IBP or some other kw density tool does .. but I am not sure of that, and while you and other knowledgable SEOs are, that doesn' t make it definitive. The algorithms are not known absolutely.

Perhaps 'keyword density' is a myth precisely because it is an oversimplified approach to what is clearly not a myth: the judicious use of keywords in various page elements is critical in ranking well for a given phrase. See SEOmoz | Google Search Engine Ranking Factors for what one panel of experts have to say about keyword use factors ...

Some of you may want to read this Matt Cutt's post which touches on the topic. SEO Advice: Writing useful articles that readers will love ...

MJ

[GoogleRankings]

Originally Posted by mjtaylor

Well, you may be right ... perhaps SEs don't measure it; perhaps it is a myth, but I do notice patterns in keyword density when I use IBP to analyze sites ... and it can't hurt to look at the top sites and not exceed their density.

Now, it's been years since this happened ... 4 or 5 ... it was somewhere along the time of the Florida Update ... and I got a little optimization happy on a particular site ... I won't say I stuffed keywords, but I just put them everywhere ... okay it was kinds stuffing: h1s, h2s, h3s, bold, italic, alts, file names, titles, ad nauseum, and the site slid down to #7 ... I reduced the overdone phrase and voila, back to #3 where it remains today .... perhaps that was not a keyword density filter, per se, but it sure quacked like one.

MJ

First problem... Using IBP to determine patterns... you can also notice patterns with Web position gold yet they are meaningless...

Second problem... 3 years? You are basing your response, your evidence on 3 years ago? That is a little inaccurate to say the least...
Secondly your basing your evidence on one case, you need to try more than one case to prove anything... I have made changes and "voila" I am back to top ten and then after careful testing I realize the ranking flux had nothing to do with those changes.

I am speaking from trial and error... I used to have many misconcepted(new word, add to your dictionary list) ideas and probably still do but once I began researching theories on various sites I began to actually be able to prove things...

One of my bigger researches was the Google Analytics bounce Factor and its effect on rankings, this took a ton of research, 500+ participants and over 50 sites I manually picked over... the results were worth it.

Also work as a team, I have found this is beneficial, your research becomes much more fool-proof.

If someone can provide me test they have made that prove keyword density is still a factor I WANT to see them, it will mean my research went bad

[crankydave]

It's all fine and dandy to discount "keyword density" but until it is defined from an "opinion" perspective and a SE's perspective any conclusions are meaningless.

Dan... If you know precisely how and why SE's "discount" density based upon exactly what they take into consideration and why, please point to it.

GoogleRankings... If you have something that definitely proves that keyword density is in no way a factor as the SE's may/maynot percieve it, please feel free to point to that definitive proof.

Dave

[mjtaylor]

Originally Posted by GoogleRankings

First problem... Using IBP to determine patterns... you can also notice patterns with Web position gold yet they are meaningless...

There's not a problem.

Originally Posted by GoogleRankings

Second problem... 3 years? You are basing your response, your evidence on 3 years ago? That is a little inaccurate to say the least...

Originally Posted by GoogleRankings

Secondly your basing your evidence on one case, you need to try more than one case to prove anything... I have made changes and "voila" I am back to top ten and then after careful testing I realize the ranking flux had nothing to do with those changes.

Yes, of course, 3 years, 5 years ... everything changes.

But I am not basing my conclusions on one case ... it was simply illustrative of what I have seen. And I still see what is clearly a result of a change in on site keyword optimization - which would change the keyword density on the page.

I am not a huge SEO company that does testing - I don't have the time; I busy producing top ten results for many, many web sites over nearly a decade. I trust my observations and I read others' tests and opinions.

That's all these are. Opinions. No one knows anything definitive here. And if there were a problem, that would be it.

Originally Posted by GoogleRankings

If someone can provide me test they have made that prove keyword density is still a factor I WANT to see them, it will mean my research went bad

Can someone provide a definitive test that it was ever a factor?

I think we need to define 'keyword density'. What do you mean by that term, GoogR? Dan seems to agree with me that you can overuse keywords. Do you disagree with that?

When I optimize a page for a phrase, I change the density ... it is a factor ... it's just the SEs look at a host of keyword factors ... and don't count the density as one pure total ... but it is still a number that can be analyzed on a series of pages and it will give you a guide as to what keyword density pattern is among top ten sites.

And I will return to my original point: you don't really want to exceed what others are doing.

It's all relative, whether we are talking about keyword use in a title tag, headings tags or content.

Keyword density doesn't exist? Show me a page that makes it to the top ten for any even slightly competitive phrase with a keyword density of less than 1%.

To say keyword density doesn't exist is to say content doesn't matter. And we *all* know better than that.

It's not quite accurate to say this is another PR Flag to rally round, because one does not need it to succeed in SEs, but it feels like it's a big button.

MJ

[DanThies]

The difference between your mythical keyword density and what really happens is not merely semantic, MJ.

Someone comes in here asking for help figuring out the magical density. All kinds of numbers get tossed out by other members as well as the "expert" moderators.

CrankyDave, I linked to a post on my blog which explains some of this, and I'll offer another link here:
The Keyword Density of Non-Sense

If you want to educate yourself, it's easy enough. If you want to argue about it, good luck with it, guys. I'm just trying to help. Giving people who are trying to learn SEO magic percentages to use for their "keyword density" is not helpful.

[mjtaylor]

No one is offering magical percentages. No one is suggesting anything wonderful will happen if kw density reaches a magic number.

And no one is an expert. We are all members offering opinions.

I will ask my question again; what is the difference between keyword overuse (your way of putting it) and the mythical term of 'keyword density'?

MJ

[DanThies]

MJ,

Overusing keywords to the point of getting filtered out of search results is actually pretty difficult. The mechanism for this is not the same thing as the mechanism for retrieving and ranking pages in search results.

I'm going to assume you followed that link and read it. Summarizing some of what it says here.

Terms which occur less frequently in the index are weighted higher when they occur in your document. Terms which occur more frequently in the index are weighted less when they occur in your document.

There is a theoretically optimal keyword placement, but it is different for every document for every search query, and without access to the algorithm and a copy of the index, you can't determine what it is.

These two facts alone make any magic number meaningless. It also makes simple numerical comparisons between your document and my document meaningless.

To measure keyword density, search engines would literally have to inspect every possible matching document for every search query, or pre-calculate the keyword density for every possible matching query for every single document.

To measure keyword density yourself, you have to decide what it means, don't you? There are three ways to measure it, and all of them are flawed. That's why you can debate magic numbers endlessly - the number you get depends on the search term and how you measure.

Anyway... we really should get out of here and go barbeque something, no?

Because the real answer to "what should I write" has almost nothing to do with SEO. What you should write is whatever gets the job done when a visitor arrives. If that means you don't use keywords at all, you can fix that with SEO.

[crankydave]

Originally Posted by DanThies

The difference between your mythical keyword density and what really happens is not merely semantic, MJ.

Someone comes in here asking for help figuring out the magical density. All kinds of numbers get tossed out by other members as well as the "expert" moderators.

CrankyDave, I linked to a post on my blog which explains some of this, and I'll offer another link here:
The Keyword Density of Non-Sense

If you want to educate yourself, it's easy enough. If you want to argue about it, good luck with it, guys. I'm just trying to help. Giving people who are trying to learn SEO magic percentages to use for their "keyword density" is not helpful.

No Dan...

You show up after not having made a peep for more than 2 years...

You don't even bother to have the courtesy or decorum to provide the OP who has a legitimate question a prefunctory answer to their question...

Instead, you drop a link to YOUR blog...

That, would be self promotional spam.

Shame on you.

Dave

[DanThies]

Self promotional spam, Dave? That's a pretty serious accusation for someone who doesn't know me at all. I don't need to spam, Dave. If anyone had any questions about my motivation, they could as easily have PM'd or emailed me - I am not hard to reach.

Anyway, I'm sorry that I linked to some information there, Dave. The post that I linked to provided a much more complete answer to the question than anything I had seen posted here.

I was actually getting ready to send a shorter version over to WebProNews for publication, so the people I supposedly spammed would have seen it anyway. I didn't write it all into the post, because the truth is actually a long story. I guess that's why people like the convenient myth of keyword density.

But I seriously doubt that my post was deleted because anyone thought it was spam. I'm guessing it was deleted because it didn't follow the party line on keyword density. I notice that WPW sent a pingback (and got a link right back from my blog). Was that spam too?

I came back over here because I got an email about this thread. I've tried to help out in a couple other threads as well. But don't worry, it'll probably be a while before I'm back. Things don't seem to have changed much. Try to help, get insulted. Keep up the good work, Dave.

[mike]

I appreciate Dan stopping by and I don't think he was doing anything spammy at all...

Just my 2 cents.

[Devesh100]

Don't pay too much attention to keywords density as this is not the only consideration by google.
I would advise you to write only keeping in mind the real RELATED keywords to your site.
Be natural and pay ONLY DUE attention to keywords not too much...!

[mjtaylor]

Dan, you do seem to have a legitimate point of view on keyword density. *And* you seem unwilling to address what I believe to reasonable questions.

I repeat (for the sake of healthy debate ):

• What is the difference between overuse of a keyword and excessive keyword density?
• What is the definition of keyword density?
• Can you (or anyone) can show me a page peforming in page 1 SERPs that has less than 1% density?

I do agree with you that there is not a magic number for overall keyword density on a page, but there does seem to be a threshhold (and I believe it varies from term to term, relative to other elements on competing pages) past which too many keywords on a page harm position in the SERPs.

I think the problem is here is that 'keyword density' is an oversimplified way of looking at the overall use of keywords on a page and it is much more valuable to look at individual elements: heading tags, titles, text, etc. and assess whether a keyword is overused or missing, as the case may be.

Could that be the real problem?

And what sort of a problem is this? Outrage that novice webasters are being misled would suggest that there is great harm in analyzing websites in this way. And yet, we know, that if keyword density runs somewhere in the range 3-5% (or maybe higher relative to competing sites) that a page will not suffer for that, perhaps specious, calculation.

I fail to see great danger in this myth - if myth it be. I do not advocate focus on it, but I do not see a reason for much concern if someone does consider it. Much ado about nothing. Yes, there are far more important things to focus on ... *and* I think you asked a legitimate question, therapists: Should you be concerned about the change in density by the addition of your footer?

I don't think 9-10% is high enough to warrant concern if the content on the page is viable ... *and* I think only a test will tell ... However, it might be a good move to consider changing some of the target keywords to synonyms. That could reduce the density of the keyword while potentially improving relevance.

Cheers, MJ

[hawash]

you can check keywords densite using seo quake
which also has helps u to suggest other words

[DanThies]

Wow, Dave. You post that outrageous accusation, then delete my response. Shame on you. You're abusing your power as a moderator. I expect a retraction and apology.

I've had a long relationship with iEntry. To be insulted in this manner by one of their moderators is totally uncalled for. For you to compound that insult by deleting my response tells me that they need to take some steps to clean this place up.

According to my analytics, more traffic flowed to WPW than from, via the Pingback that WPW sent to my blog. Which isn't surprising, given that my post has been viewed by thousands of people, whereas this thread has been read by several dozen at most.

So why in the world would I bother to spam anyone, Dave? I was trying to help the original poster and everyone else by pointing out that they don't need to worry about keyword density.

The truth is, my original post wasn't deleted because it had a link in it... or this post would have been deleted as well:
All Your Jobs Are At Stake
That person got the courtesy of a reply from a moderator.
Besides, you didn't just delete my post, did you? You deleted a post by another member who apparently sinned by agreeing with me.

[DanThies]

Thanks, Mike.

Originally Posted by mjtaylor

Dan, you do seem to have a legitimate point of view on keyword density. *And* you seem unwilling to address what I believe to reasonable questions. I repeat (for the sake of healthy debate ):

• What is the difference between overuse of a keyword and excessive keyword density?
• What is the definition of keyword density?
• Can you (or anyone) can show me a page peforming in page 1 SERPs that has less than 1% density?

MJ, I will try to address these questions directly, but you must understand that the full explanation of what really happens is long and technical. Dr. Garcia's lengthy paper is actually a lot shorter than a full response to the question would require, but on the whole you'd gain a better understanding from him than you'll get from me.

For a particular document and search query, you have a query vector and a document vector. Lots of complicated math determines how closely the two match. If you start with a theoretically ideal document, fewer occurences of a term, or more occurences, will move the document further away from matching the query.

This is only "on page" so you don't have to shoot for perfection - anchor text can easily overwhelm any on page factors for a given search term. Different search engines will also have different approaches to filtering documents out of results, which is likely another layer that happens outside of the document retrieval process.

Definitions... how you define keyword density is part of the problem. What people are talking about is percentages, but as I've pointed out the number depends on the method, and all possible methods lead you to silly places depending on the search term.

I'd define the myth as the belief that search engines measure keywords in a query as a percentage of text in a document (this is literally impossible), and that there is some magical percentage to shoot for.

There are plenty of examples of all-flash pages (no text) getting ranked, MJ. It's a lot easier to get the page retrieved and ranked if the words appear on the page, but Pizza Hut's text-free home page was #1 for [pizza] for years.

And what sort of a problem is this? Outrage that novice webmasters are being misled would suggest that there is great harm in analyzing websites in this way. And yet, we know, that if keyword density runs somewhere in the range 3-5% (or maybe higher relative to competing sites) that a page will not suffer for that, perhaps specious, calculation.

Modifying the text to satisfy *any* percentage targets is the wrong approach, and it's not just inaccurate, it's harmful. Not only does this kind of teaching cause people to waste their time on "on page text" optimization, it causes people to write bad copy.

It's not just novices, because the keyword density myth is so pervasive that even people who have been working at this for years still think it's real.

I'm not trying to pick on Dave here because I think that he's probably a decent guy (albeit overly cranky), but his religious jewelry site's home page is a good example - the first "sentence" of copy isn't even a sentence, it's pretty much just a list of keywords, and he doesn't need to do that to get ranked. That kind of copy can't be helping him convert visitors and make sales. Believing in keyword density is probably costing Dave a good deal of money.

The words on a page should be "optimized" to persuade, inform, communicate, etc. Yes, we should use keywords in the text, because unless you use the words on the page it's a lot of work (anchor text) to get the page retrieved in response to a query. I would suggest (and this is just my opinion) that it's hard to write a relevant page without using keywords anyway.

[mjtaylor]

Originally Posted by DanThies

MJ, I will try to address these questions directly, but you must understand that the full explanation of what really happens is long and technical. Dr. Garcia's lengthy paper is actually a lot shorter than a full response to the question would require, but on the whole you'd gain a better understanding from him than you'll get from me.

For a particular document and search query, you have a query vector and a document vector. Lots of complicated math determines how closely the two match. If you start with a theoretically ideal document, fewer occurences of a term, or more occurences, will move the document further away from matching the query.

This is only "on page" so you don't have to shoot for perfection - anchor text can easily overwhelm any on page factors for a given search term. Different search engines will also have different approaches to filtering documents out of results, which is likely another layer that happens outside of the document retrieval process.

I appreciate the reference to Dr. Garcia and I will see if I can wade through that ... however, I am not real clear on how this addresses my question directly ... I am not sure which question you were addressing at all, but here's the one that has gone unanswered again:

many posts above you wrote:

Originally Posted by DanThies

MJ, you can absolutely overuse keywords

and I am asking you to tell me why overuse of keywords is different than excxessive keyword density.

I am *not* suggesting that SEs measure KW density directly; I am suggesting that such a calculation, however it is done, may still be useful to help someone see when they 'overusing' keywords ...

Originally Posted by DanThies

Definitions... how you define keyword density is part of the problem. What people are talking about is percentages, but as I've pointed out the number depends on the method, and all possible methods lead you to silly places depending on the search term.

Silly places? This doesn't help me understand your point of view.

Originally Posted by DanThies

I'd define the myth as the belief that search engines measure keywords in a query as a percentage of text in a document (this is literally impossible), and that there is some magical percentage to shoot for.

I have said several times that I agree that there is not a magical percentage ... and that I agree SEs may not make any such measurement, but why is this mathematically impossible?

Originally Posted by DanThies

There are plenty of examples of all-flash pages (no text) getting ranked, MJ. It's a lot easier to get the page retrieved and ranked if the words appear on the page, but Pizza Hut's text-free home page was #1 for [pizza] for years.

Good point. Lots of links can overcome 0% kw density.

Originally Posted by DanThies

Modifying the text to satisfy *any* percentage targets is the wrong approach, and it's not just inaccurate, it's harmful. Not only does this kind of teaching cause people to waste their time on "on page text" optimization, it causes people to write bad copy.

Well, maybe .. it might also cause them to stretch a little and learn how to write better copy in order to put keyords into the content seamlessly.

Originally Posted by DanThies

It's not just novices, because the keyword density myth is so pervasive that even people who have been working at this for years still think it's real.

Like me.

[DanThies]

MJ,

I believe I've already answered what the difference is between "excessive keyword density" (which you still haven't told me how to measure) and what really happens inside a search engine.

So let's just talk about keyword density again. Depending on how you define/measure it, your percentage will be very different depending on the search term. Read my post above on the 3 methods and how they don't work. That's what I mean by silly. You can get really high or really low numbers for the same term & document, depending on how you measure it.

Why is it impossible for a search engine to calculate keyword density for every document that might match for every query? Aside from the fact that there's no logical way to define the measurement, just think about it. Let's say you search for [large purple widgets]. Maybe 100,000 pages have all of those words on them, right?

Unless they somehow pre-calculated and stored a keyword density score for every possible combination of words on every page (requiring more storage space than you can imagine), they'd have to go calculate keyword density for a whole bunch of documents really really fast every time you searched... which gets even harder if you want to use one of the methods that counts exact phrase matches. They'd have to reconstruct all those phrases on all those documents from their index, which is a big database of words and where they appear.

Bottom line here... keyword density isn't part of what search engines do. That's not my opinion, it's a fact. So why are we talking about it?

[mjtaylor]

Originally Posted by DanThies

MJ,

I believe I've already answered what the difference is between "excessive keyword density" (which you still haven't told me how to measure) and what really happens inside a search engine.?

Well, that was your term, and I was trying to get how *you* measure it since you seem to feel it's different than 'keyword density.'

As for keyword density measurement, I would vote for word by word, as IBP does it ...

Originally Posted by DanThies

Why is it impossible for a search engine to calculate keyword density for every document that might match for every query? Aside from the fact that there's no logical way to define the measurement, just think about it. Let's say you search for [large purple widgets]. Maybe 100,000 pages have all of those words on them, right?

Unless they somehow pre-calculated and stored a keyword density score for every possible combination of words on every page (requiring more storage space than you can imagine), they'd have to go calculate keyword density for a whole bunch of documents really really fast every time you searched...

Bottom line here... keyword density isn't part of what search engines do. That's not my opinion, it's a fact. So why are we talking about it?

I'm not buying that Google *can't* do it ... Google stores whole pages, does it not and does other calculations. I am not saying it or any other SE does ... I have said all along that I think the concept of keyword density is oversimplified and I have not maintained that there is a pure calculation of it ... but I don't think we can aver that as a fact ... I don't see any clear statements from SEs one way or the other, so I am not sure why you think it is other than an educated opinion. I do see lots of respected SEOs who agree with you and I also agree that it is not a single factor.

But I think many SEOs also agree that the placement of keywords in various elements *is* a factor, and that there are there are limits and that overuse of keywords by whatever name can stink of the SERP garden. Even *you have said you can overuse keywords ...

You seem as interested as I in talking about it ... indeed, don't you have a sort of campaign against it on your blog?

Still must go and look into Garcia ... and must get a life ...

MJ

[DanThies]

Originally Posted by mjtaylor

Well, that was your term, and I was trying to get how *you* measure it since you seem to feel it's different than 'keyword density.'

Measure what? Term weights? You can't, that's the mistake Rand made a couple years ago.

I'm not buying that Google *can't* do it ... Google stores whole pages, does it not

They cache the HTML code of pages, yes, but that's not used in search, it's stored in a compressed format so they can show it to you when you ask to see the cached copy.

For search they have an index. Page & Brin's original paper is clear enough once you understand enough IR theory. The innovation of Backrub/Google was in PageRank though, the rest of what they did was pretty standard. Which is probably why they didn't feel compelled to explain it in great detail.

Anyway, the science of IR continues to advance, and it's likely that many innovations of the past 10+ years have made their way into practice at search engines. Keyword density, however, is not useful for retrieving and ranking documents.

If you make it to SES some time, ask an engineer about keyword density. That's how I got my wake-up call. Then Dr. Garcia decided to show up at SEW and help us all out.

[crankydave]

Originally Posted by DanThies

Wow, Dave. You post that outrageous accusation, then delete my response. Shame on you. You're abusing your power as a moderator. I expect a retraction and apology.

I've had a long relationship with iEntry. To be insulted in this manner by one of their moderators is totally uncalled for. For you to compound that insult by deleting my response tells me that they need to take some steps to clean this place up.

I've no idea what you're talking about Dan. I've deleted nothing you posted.

Dave

[mjtaylor]

Dan, Dan, Dan,

I am not saying and haven't ever said that Google calculates KWD. I am not asleep here.

I am saying they could if they wanted to and that it is a reasonable way to assess whether one overuses a keyword phrase on one's site ...

I do appreciate you trying to educate me; I did read Garcia's article on the Non Sense of KW Density and it was fairly clear to me considering the algebraic feel of it ... thank you for that reference. You might want to read it again, too, as I think it might help you to express some of your arguments more logically or convincingly.

I also wish you would read my posts and see that I have not disputed that KWD is a myth, so to speak ...

And as we do agree that SEs don't calculate keyword density or ratios or excessive uses of it (whatever you want to call it) then what's wrong with keyword stuffing?

MJ