Anchor Text Question

[cameron]

Suppose that the term "fake blood" is my main target and "funny fake blood" is something I want to target as well, but it's not searched for nearly as much as fake blood. If I have sites link to me using the anchor text "funny fake blood" will it help me as much for ranking with "fake blood" as it would if I had the anchor text just set as "fake blood"

Would linking with "funny fake blood" be a good choice, or should I just go after my more important term of "fake blood"?

[r2d2]

"fake blood" would be better for targeting "fake blood" than using "funny fake blood".

I would go for "fake blood" I think.

[chromate]

Depends how competitive the term is. For something that's not so competitive, I would go with "funny fake blood". However, if there's a lot of competition, just go for "fake blood" to start with and then when you're in a position you're happy with, get any subsequent links to use the anchor text "funny fake blood".

[nohaber]

"funny fake blood" targets both keyword phrases. You get the benefit of targeting both of them. My only concern would be, if the visitors who search for "fake blood" will actually *click* on a title such as "funny fake blood".

[tomek]

I think nohaber is right...

[Chris]

Yes. Funny Fake Blood gets you a 100% score on Funny Fake Blood and a 66% score on Fake Blood. Its the one I would use.

[cameron]

Thanks.

Nohaber, it was just an example. I don't have any sites targetting fake blood or funny fake blood. I don't think there is really such thing as funny fake blood.

[nohaber]

Chris,
are you saying that keyword density in the anchor text matters?
Cameron,
I also didn't mean exactly these keywords. I meant that the title should appeal to the searchers for max CTR, and for some keyphrase that might not be the case.

[Chris]

Yes, keyword density in everything matters.

Density is the best way to combat the original form of SE spam, repetition.

[nohaber]

IMO, density has never mattered in anything, at least for Google

[chromate]

As I said, in the example "Funny Fake Blood", if "Fake Blood" gets you 1000 searches a day and is a competitive phrase, but "Funny Fake Blood" only gets 50 searches a day then I would go for "Fake Blood" Because the benefit of the 100% score would be felt more than the 50 extra searches a day. You have to balance up what everything's worth.

[r2d2]

As I said, in the example "Funny Fake Blood", if "Fake Blood" gets you 1000 searches a day and is a competitive phrase, but "Funny Fake Blood" only gets 50 searches a day then I would go for "Fake Blood" Because the benefit of the 100% score would be felt more than the 50 extra searches a day. You have to balance up what everything's worth.

This is pretty much why I reckoned going with 'fake blood', should have put some reasons with mine.

It all depends on the ratio of the two phrase's search frequencies I guess, if it was 1:1 I would go 'funny fake blood' if 20:1 prolly 'fake blood'.

[nohaber]

imo, there's no such thing as 66% and 100%. "Funny fake blood" has exactly the same relevancy as "fake blood" for "fake blood". It is the number of instances that matters, no the density.

[nohaber]

And to support my view on the keyword density, here's an extract from the Google's prototype paper. Where do you see density? I have bolded the text that shows how spam is coped with:

===
4.5.1 The Ranking System
Google maintains much more information about web documents than typical search engines. Every hitlist includes position, font, and capitalization information. Additionally, we factor in hits from anchor text and the PageRank of the document. Combining all of this information into a rank is difficult. We designed our ranking function so that no particular factor can have too much influence. First, consider the simplest case -- a single word query. In order to rank a document with a single word query, Google looks at that document's hit list for that word. Google considers each hit to be one of several different types (title, anchor, URL, plain text large font, plain text small font, ...), each of which has its own type-weight. The type-weights make up a vector indexed by type. Google counts the number of hits of each type in the hit list. Then every count is converted into a count-weight. Count-weights increase linearly with counts at first but quickly taper off so that more than a certain count will not help. We take the dot product of the vector of count-weights with the vector of type-weights to compute an IR score for the document. Finally, the IR score is combined with PageRank to give a final rank to the document.
For a multi-word search, the situation is more complicated. Now multiple hit lists must be scanned through at once so that hits occurring close together in a document are weighted higher than hits occurring far apart. The hits from the multiple hit lists are matched up so that nearby hits are matched together. For every matched set of hits, a proximity is computed. The proximity is based on how far apart the hits are in the document (or anchor) but is classified into 10 different value "bins" ranging from a phrase match to "not even close". Counts are computed not only for every type of hit but for every type and proximity. Every type and proximity pair has a type-prox-weight. The counts are converted into count-weights and we take the dot product of the count-weights and the type-prox-weights to compute an IR score. All of these numbers and matrices can all be displayed with the search results using a special debug mode. These displays have been very helpful in developing the ranking system.

[chrispian]

nohaber, I seriously doubt your theory. Density is clearly important. I have a page where I target a specific keyword phrase and my site is #1 and there is less than a paragraph on the page. The competing sites are all very large pages with many many more words on their page. Even though they use my term more often, I use less words making the keyphrase density much higher so I rank ahead. I would say keyword density is one of the #1 factors in ranking well.

The paper you're quoting is old and outdated by at least 4 years. They've had to adapt many times over to the changes in the search market.

[Chris]

All search engines use keyword density. It is seriously the best spam fighting tool ever invented.

For instance it completely kills the notion that you could put an entire paragraph in h1 tags to get a boost.

[nohaber]

Why would a search engine reward a page for keyword density? It would mean that a wordier author will be less relevant, than someone who writes shorter texts.
The h tags are used to determine just the font size. Google keeps font size, capitalization and position info. It keeps no H tags info. If you put a whole paragraph in H tags, the parser will think it is too long for a big headline, and just write it in the index as normal font text, not big bold one.
The way the paper is written, spam is fought with no problem. You may repeat a keyphrase 1000 times, but only (example) 10 of them will count. So, you'll have to have PageRank to compete. The two major scores are PageRank and the IR score. And the major contributor to the IR score is anchor text and title. So the onpage factors, don't have too much room for influence. When you cut off the relevancy after a preset number of matches, you limit keyword spamming to the minimum.
Let's say you have two pages with the keywords 5 times repeated. If page 1 has 200 words, and page 2 has 300 words, then according to your logic, page 1 is more relevant, although, there's more information in page 2. Keyword density pushes user-unfriendly pages upfront. If that is quality, then Google might be out of business.

chrispian,
why are you sure that your page outranks the rest by keyword density, and not by some of the many more other factors? I can give you counterexamples.

[chromate]

nohaber makes some interesting points. I know I've seen examples of pages that have ranked well with very low keyword densities in anchor text, title and on page content.

Why can we not just set up an experiment to prove or disprove this? It would be so easy to do. Maybe you could do this nohaber and report back.

[Chris]

Why would a search engine reward a page for keyword density? It would mean that a wordier author will be less relevant, than someone who writes shorter texts.


Actually under what you propose that would be true. You are favoring repetition or length over density.

Also density is just one aspect and it is used in a variety of ways.

Are you really of the opinion that anchor text using the words

"Pearl Jam Sex Cars Britney Spears"

Is as helpful as anchor text using just the word "Cars" assuming you are targetting Cars?

Were you online in the late 90s, say 97 or so, before search engines were using density? Do you remember how searching on alta vista was a joke? Every site would litter their meta tags and invisible onpage text with irrelevant pop culture references to get ranked well on all searches, even those they had nothing to do with.

[nohaber]

"Pearl Jam Sex Cars Britney Spears" has the same relevancy as a "Cars" title for "Cars".
How do you know that keyword density solved the SERP problems of the dumb search engines? Even google's prototype returned great results without density.

How would density solve spamming? In fact, if density is used, it would prevent pages from ranking high for multiple keywords at the same time, which is dumb. A high-quality page should be able to rank high on a variety of keywords.

Google introduced word stemming which is just the opposite of keyword density. That's a step in the opposite direction.

During the late 90's I was a university student. I know how bad the SERPs were. It was difficult to find good resources for papers. It was even difficult to find a decent free porn site.


chro,
if anyone has any doubts about density, let him do some testing. I am 100% sure that density has never been used, and won't be used by any decent search engine. I won't waste my time on sth I am 100% sure. Chris, thinks just the opposite. Let the uncertain ones do some testing. :D

[Chris]

How is stemming, extrapolating word matches based on different forms of a word, the opposite of density?

[nohaber]

Density favors the repeating of a single keyword phrase. That means scoring high on a given phrase, and scoring badly on other phrases.
Stemming is just the opposite. If you optimize for "diet supplements", Google also takes into account "dietary supplements". A high PR page optimized for "dietary supplements" can rank very well for "diet supplements". Google wants high quality pages to score high on a variety of related keyphrases. Density favors one or a few phrases. Although "dietary supplements" will score less than "diet supplements" for "diet supplements", it can still get to the top with sufficient PR.
Of course, my reasoning depends on the definition of "density".
Density favors -> one or just a few phrases (pages not written for the users, text looks strange)
Stemming favors -> scoring high on a variety of related phrases (pages written for the visitors, text looks fluent, you don't have to cut off words to bump up the density)

I just don't get it how density improves relevancy. It simply favors over-repeating keywords. Density favors spammers (they could rank well without so much PR). Counting keyword instances and cutting them off after a certain count, simply makes spamming unfeasible (it could work only with high PR).

[Chris]

Density can mean many things. For overall keyword density on a page you're not looking at 50%, 20%, or something crazy like that. You're looking at probably 3-8%. With stemming you have all related words helping the density as well. The bonus is also likely weighted so that 3% gets you "10 points" and 8% gets you "20 points" and 50% gets you the same "20 points" that 8% does. (just to be sure, this is all just a hypothetical example, I do not actually know what kind of "points" are given for what exact percentages.)

Density for anchor text and title tags is different. Here you need as high of a percent as you get. They're simply smaller.

In both cases density discourages spam. Although there are different types of spam.

The worst kind of spam, I'm sure you'll agree, is the kind where a site gets ranked on an unrelated keyphrase. Simple repetition allows spam like this to run rampant. Using my example of anchor text above for instance.

If things worked as you say then I could simply put "Britney Spears" 10 times at the bottom of my page and append "Britney Spears" to the anchor text of my incoming links I would recieve absolutely no penalty to my main keywords for doing so. I could do the same with "Pearl Jam" "Sex" "Mp3" "Porn" and everything else.

The principle of density is very much the same as the principle of PageRank flow. There is an absolute amount of PageRank you can send and so it must be divided among all outgoing links. Likewise there is only a total of 100% density and so trying to target multiple keywords in turn makes each one perform less.

[nohaber]

Chris,
to spam with titles/anchors such as "Pearl Jam Sex Cars Britney Spears" you STILL need PageRank. No one is going to give you such a link. So, you'll have to use your PageRank power. It is obvious, that no one with a bunch of sites and great PR power will ever do this. So, this is quite an unrealistic situation. If there is some dumbster with a lot of PageRank who stuffs anchor text with these irrelevant keywords, all his competition will spam report him. He'll get all of his sites banned, and will be cited in all webmaster books as the dumbest whoever lived.
In the 21st century, everyone is trying to be relevant for a bunch of keywords actually relevant to his/her site. This leads to overrepeating of a small set keywords. That is what a lot of webmasters do. Because this is a much greater threat to search engines, they should concentrate on preventing overrepeating. And density is counterproductive.
PageRank and similar link popularity score, simply make situations such as "Pearl Jam Sex Cars..." practically non-existent. If you own a search engine, are you going to chase a non-existent problem (solved by PageRank) or are you going to limit the over-repeatedness of keywords. Plus, if all webmasters knew it, the overall quality of the websites would be higher.
And Google's prototype paper is not outdated. It is one of the most insightful texts one can read. The problem is that not everyone can understand its beauty. That's because most SEOs are ex-journalists. :D

[Chris]

No one will ever do that? With the rampant selling of links that goes on nowadays?

[nohaber]

Chris,
Who's going to buy a link "Pearl Jam Sex Cars Britney Spears"?
Who's going to sell a link ""Pearl Jam Sex Cars Britney Spears"?
Even if it happens, spam reporting will work it out.
Plus, buying (esp. high PR) links can be devalued very easily. Here are 2 ideas:
1) make high PR links count after a couple of months (let's say 6 months). buyers will need to pay additional 6 months before getting any value
2) high PR unrelated links can be identified very easily by a simple graph algorithm that measures the distance between sites on a hostgraph (Bharat has one related paper).
etc.
Anchor/Titles like "Pearl Jam Sex Cars Britney Spears" + high PR are so rare, that no one should give a *** about them.

[Chris]

So you implement an imperfect system in hopes that people will report enough spam to make it perfect?

It could be that a shady SEO company puts hidden links on a client's site without their knowledge. Quite a few do this actually. Their "SEO" consists of hidden links on all client sites pointing to other client sites.

Of course "Pearl Jam Sex Cars Britney Spears" is the extreme too. What about something less extreme. "Web Hosting & Domain Names" or "Web hosting, Development, & Domain Names" or "Web Hosting, Development, Domain Names, Web Templates" Do you really think that you will get the same help on the phrase "Web hosting" for each different anchor text?

If you do, then thats where you disagree with me and it seems most others.

[nohaber]

yes,
a "Web Hosting, Development, Domain Names, Web Templates" title is just as good as any short title with just one phrase. Of course, you disagree with me :)

Consider this: you have a site about "web templates".
title#1: "Web Templates, Web Templates, Web Templates, Web Templates"
title#2: "Web Templates"
title#3: "Web Templates - a large database of bla ..".

Which title is the most user-friendly? Of course, it is #3. Density will favor #1 and #2. According to your logic, a page shouldn't rank high for a couple of keywords because density prevents this. But, I think that a page should be perfectly able to rank high on a variety of keywords. Density favors keyword titles, and prevent webmasters from writing descriptive titles. That does not improve quality of the SERPs. Bad SERPs = no one uses your search engine.

The quality of the results should be improved by methods other than density. Density is useless.

[chrispian]

Well, personally I hope more people think like nohaber. It means I'll continue to out rank them.

[nohaber]

Your statement must be qualified.
It means I'll continue to out rank them (if I am outranking them because of density, and if nohaber is wrong)