Did Google drop PR in favour of keyword density?

That's the way it looks after the last update.

[rencke]

Did Google just lower the weight of page rank in favour of keyword density and proximity? So it seems.

Consider this search for a two word phrase: kwd1 kwd2. The search is made about 10.000 times/month and yields 387.000 pages. The following ranking is brand new as of the recent update and represents a huge jump up for #1 and #2.

#1 PR2 Title: "Fluff kwd1 Flash detect" (a warning to gentlemen to keep their trousers zipped up?) Description: fluff kwd1 kwd2. No text on page.

#2 PR3 - a page from the #1 site. Title: "Fluff kwd1 kwd2 fluff fluff kwd1 fluff kwd2 fluff."

Google shows <50 inbound links for the #1 site.

#3 PR6 Title: "kwd2 fluff fluff fluff kwd1 kwd2 fluff"

#4 PR5 - a page from the #3 site. Title: "kwd2 fluff fluff fluff fluff kwd1 kwd2"

Google shows >1000 inbound links for the #3 site.

The way this looks to me is that PR is no longer the most important factor in determining rank. Keyword density and proximity are. If this is right, Google has become a lot more like other search engines and suprisingly easy to spam. Any thoughts?

[Ove]

I totally agree with you and there is one more thing and i will say it again kw in the domain name will give a big boost.

/Ove

[korkus2000]

I have noticed that too. For all of the keywords I use domain keywords really matter.

[WebGuerrilla]

I saw the same thing last update. Had a site (PR5)that consistantly sat at #4 out of 2.2 million pages. Spot 1-3 have always been occupied by 2 PR7's and a PR6.

After the June update we jumped to #2 (above the the 6 and one 7). We stayed there for a large part of July, but then ended up dropping back into the #4 spot prior to the last update. After everything setteled this month, we are still #4.

Clearly, they seem to be toying with on-page criteria, but it doesn't seem to be something that is being done across the board.

Popular/competitive searches don't seem to be showing any signs of on-page factors being tweaked, but many less competitive areas do seem to be producing many more lower PR pages at the top.

[ciml]

It is very difficult to draw solid conclusions when looking at 'real' groups of Web pages. Google, the Web's link structures and page structures can make things too complex to isolate cause and effect.

I'm also seeing this sort of change. Whether it's greater weighting to keyword density, HTML title or link text is beyond me.

[Grumpus]

I don't know that it ditched PR, but there is something about density - or rather, in my experience, KEYWORD ORDER.

In my home page title, I recently changed it to contain keywords. I put:

"Site Title - Keyword1, Keyword2, and Keyword3 Database"

Now, I screwed up because I used Plurals on everything and nowhere in the singular (more popular search terms) appear. That's beside the point, it was the order that I put them that screwed me up...

One of the top 100 search terms on the web is the singular version of "Keyword2 Keyword1" from my title. But, I was noticing that I was getting some hits every day from the plural version if they typed Keyword1 first. So I looked at it.

For the term: "keyword1 keyword2" I was ranked #1 in google (too bad it only gets searched for about 500 times a month according to the Overture suggestion term.

BUT, if I typed in "keyword2 keyword1" a term that gets about 80,000 (if I remember correctly) searches a month. I searched to about 200 sites deep and my site still never showed up.

Now, I understand that one term is more competitive in its exact incarnation than the other because of the way people speak and would think to enter the term, but it stunned me at how big a difference the rankings would be in google just by reversing the order. It'd be like webmasterworld's homepage being ranked number 1 for "web professional" (two consectutive words in the title) but being ranked 400th for "professional web".

I was stunned. (And I, of course, fixed my mistake, let me tell you!)

G.

[angiolo]

Recently, we decided to place keywords in the URL.

For example we changed from

mysite.com\widget.html

to

mysite.com\keyword1_keyword2_widget.html

I do not know if this is related, but after few days all the changed pages got the PR of the main page; before the change all that pages had a PR diminished by 1.

Maybe the main page is going to increase the PR.. (I hope so)

[chiyo]

I noticed the increasing advantage of keyword domains (even non hyphernated) 3 updates ago. But I think that is the symptom rather than cause.

Agree that on page factors now seem to be more important, but Im also seeing that the hyperlink text from external AND internal inward links may also be given more weighting. This explains, as many here have pointed out here, why keyword domanins go well - ie. it is far more likely that people will link to you with the keywords you want!

In just about every SERP we follow where lower PR's beat out higher PR's the former seem to have the keywords in their domain.

[doc]

I have been trying to use keywords in the domain names or subdomain names with fairly good results. On the last update I have achieved a #3 ranking with a PR3 ahead of at least two separate PR7's.
If you do a search for 411 Canada you will see what I mean. I am still not convinced the update has settled yet though as the benefit only intermittantly shows.

[mortalfrog]

others have commented on a much greater emphasis on incoming link text this time around.

perhaps this is part of the reason that PR seems less important.

if it's just keyword density / order, then I have a couple of sites that should be in #1 spots instead of #2, since the site ahead of them has equivalent PR, no instance of the keyword on the page or in the source code, and several solid incoming links containing the keyword.

really, I think it's all about the incoming link text now more than ever. PR is still crucial - but not so much your PR as the *relevant* PR of the incoming link with your key-phrase in it.

[vitaplease]

Tweak the algo to be a bit more sensitive to the proximity of words in the search query

Example: search for: Google searches from your web page (without "..")

sadly for Google, the number one result is Altavista..
and the Google pages presented are not the ones I copied the exact sentence from (it was in the first sentence of [google.com...] )

from this posting: [webmasterworld.com...] msg 121

I forgot to take a screen shot, but if you do the same search: Google searches from your web page (without the "..")

the Altavista page does not turn up first anymore.

However Google's searchcode page still only turns up 16th (of Google's own site results).

So proximity alone is still far to go.

Hence my suggestion in that posting:

And while on the subject of proximity, what about a variable control bar in the toolbar that allows you to adjust the proximity factor?

My observation, that covers too few examples to carry any value, was that two or three updates ago, Google put more emphasis on external incoming textlinks. If these textlinks form part of your multi-word query, proximity plays a smaller role.

[highman]

It appears that google has slightly less importance on links into the site, which is kinda wierd, but then seeing the increase in requests for reciprical link exchanges I get, maybe its an algo tweak to account for the growing awareness of inbounds?.

Certainly done us no harm...always a case for not putting all your eggs in one basket, but then I never believed PR was the 'be all and end all' anyway

[JuDDer]

Just to add to the discussion regarding keyword domains, I have a new site that is PR2 with a domain like keyword1-keyword2.com and it's beating PR6 sites on the first page of the Google SERPS.

The site was built without any real optimisation for the main keywords so I was surprised to see it beating much much higher PR sites like PR5 & 6.

It's certainly nice to see something like this happening as it gives hope that people with smaller sites can compete with the big boys and that PageRank is not the 'be all and end all' of good ranking. I can only put it down to the keyword domain.