April 2008 Google SERP Changes - part 3

Me, too. I was #12 a day or two ago, but #5 today, oops, *the last time I looked* (one minute ago).

Welcome to the Google Top 10 Shuffle.^TM

p/g

Maybe the bounce is part of the new strategy for Google.com what Tedster referred to as 'serendipity'.

Perhaps fluctuation will be part of the new Google.com strategy since there are many more websites nowadays that can offer quality products and services due to increased competition.

Two messages about a PR Update for April 2008 were moved here:

[webmasterworld.com...]

[edited by: tedster at 3:28 am (utc) on May 2, 2008]

Well, one small part of my mystery has been solved.

One phrase I ranked in the top ten for was "Acme widgets." I'm now on page two for that term. A couple of the sites that came out of nowhere to the first page are owned by a widget dealer who sold widgets to the fruitcakes who shot students recently at two college campuses.

The media has covered the owner of the site and his websites extensively, so it looks like the resulting increased traffic has propelled his two sites to the first page.

It will be interesting to see if he holds those rankings for long.

Pretty gruesome way to get rankings, though.

Internetheaven Thought I'd share the results from studies on my own sites showing that it is internal linking that has been devalued...

It's a good move in my opinion, but sadly they seem to have removed other filters like:
- interlinking sites all owned by one guy on one IP with no other inbounds
- article and blog spam links

Sounds like these things are linked.

Total linking (anchor text) = internal linking + inbound links.

Maybe interlinking and blog spam were devalued previously but not completely filtered out. Reduction in value of internal links, in effect, increases value of all external links including those that were previously devalued by a filter. Another example of Google engineers forgetting what they previously did and disregarding side effects of new actions.

Random thoughts: I wonder if the change in internal link value is related to the explosion in the use of xml sitemaps. If I were a Google engineer (wouldn't do it, couldn't afford the pay cut!) I would be asking the question "how do we stop sitemaps passing PR around a site?". Perhaps some sites that had pages that were naturally #1 dropped out as a result of installing a sitemap and they had to react.

Could I be right about sitemaps or are they reducing value of all internal links or only navigation, header and footer links on pages?

This is extremely interesting stuff and some of it could explain what I am seeing in our niche.

Thanks

Sid

It would be trivial not to grant PR functionality to xml pages. And from what I see, html pages are often grey-bar these days,

I wonder if the change in internal link value is related to the explosion in the use of xml sitemaps.

I wouldn't have thought so. Seems to be general across all pages and, like tedster said, the impact from devaluing just that would not be having the dramatic effects we're seeing.

Pretty gruesome way to get rankings, though.

Modern day business plan: start a blog, add adsense ads, shoot someone, blame youtube, press tells everyone about your blog, you make millions, you hire OJ's lawyer, get off and still have a couple of mil's left ...

OK, but are we talking about all internal links or just header, navigation and footer links?

Cheers

Sid

OK, but are we talking about all internal links or just header, navigation and footer links?

Pay for half the staff hours put in and I'll share what our research showed ... ;) I would imagine, though, that anyone with some experience with Google's paid link attack algorithms will see the same patterns emerging for their internal link devaluations.

I would imagine, though, that anyone with some experience with Google's paid link attack algorithms will see the same patterns emerging for their internal link devaluations.

I thought you said you wanted paying for half the man hours for this ;)

I'm pretty sure that this has combined with the UK geo filter to produce some interesting results.

For example one site was #1, on both .co.uk and .com, for a while. Two things noted in analysis were:

1. They had a javascript driven navigation that dropped down links in a clam shell menu style. They were therefore able to control the internal link structure and anchor text in a way which I guess was benefiting them.

2. Lots of links exchanged from their UK site with somewhat topic relevant US sites.

They are now still #1 on .com but have dropped to #4 on .co.uk. It looks like their geo weight has been tipped towards the US. Perhaps reduction in internal link value coupled with a few more US links has tipped the balance.

Cheers

Sid

Here is something you don't need to pay for half the staff hours put in :-)

I have noticed that during and post Update Dewey, the whining-level on this threadv didn't get out of hand. Therefore, compared to previous updates, Dewey might be looked at as a "White Revolution".

Btw, a site, in a sector related to advertising, has dropped exactly from #10 to #20 on the "Mother of all Google Data Centers" 72.14.207.** . Maybe there would be something called "#20 Penalty" :-)

Thought I'd share the results from studies on my own sites showing that it is internal linking that has been devalued. Less power is being shifted within a site for the majority of results sets we've seen (all commercial searches).

Strange enough I got 2 domains which is exactly the opposit where 2 internal pages came from > 18 to Top10, joining the main-page.

so is anyone winning in this update?

I find that Q interesting: Everybody here as well as in other forums is complaining about the update. Where are the people that got up? Perhaps we have made nothing wrong but they did something right? Or is this a kind of "punish all seos and reward all semi-prof webmasters"-update?

In my sector, I am noticing that a number of sites that have managed to retain (and some have regained) top positions have changed the title tags to very short keyword 1 keyword 2 ¦ Company name. Anyone else noticing this?

Hi,

Perhaps related to the internal linking discussion above. allinanchor:blue widget on google.com used to produce the same as blue widget on google.co.uk in our niche, now it does not.

Cheers

Sid

@Optimus That's exactly what I've seen since with this update, one site which company name is a two word competitive keyword with the title 'competitive keyword ¦ competitivekeyword.co.uk' go number one with sitelinks.

It is there 'money term' and there only term but it was usualy quite close between us but G thinks their great now >:(

Where are the people that got up?

My #1 core directory site has so far remained untouched however I have seen a lot of movement below it including my core site #2 which has bounced about considerably in the top 20 for several keywords.

I suspect the real answer to your question is that they are blissfully unaware or, if they do know they have reached the top of the SERPs, have no idea how they did it and won't be particularly concerned when they drop again...they'll have had their 15 minutes of fame:-)

title tags to very short keyword 1 keyword 2 ¦ Company name

Certainly, I have a .mobi site firmly in the top 11~15 for several weeks now. It's a good site and it's applicable however in the regular SERPs?

I just do not expect to see a .mobi site there. Is this wrong of me?

Thought I'd share the results from studies on my own sites showing that it is internal linking that has been devalued. Less power is being shifted within a site for the majority of results sets we've seen (all commercial searches).

Strange enough I got 2 domains which is exactly the opposit where 2 internal pages came from > 18 to Top10, joining the main-page.

I think you have may have mis-read my post. I said internal linking has been devalued, not internal pages. If a page has merit based on other factors than internal links then those pages should move up in ranking (as you have noted). Pages with "self good merit" will start to pass internal pages of websites with simply "site good merit".

I am noticing that a number of sites that have managed to retain (and some have regained) top positions have changed the title tags to very short keyword 1 keyword 2 ¦ Company name. Anyone else noticing this?

Yes. But it is not always the result of actual intentional SEO work.

That's exactly what I've seen since with this update, one site which company name is a two word competitive keyword with the title 'competitive keyword ¦ competitivekeyword.co.uk' go number one with sitelinks.

Makes me feel a little better about all those keyword domains I bought up when I was a young SEO! Just launched a brand new site that had keyword1keyword2.co.uk and it shot straight in at No.2 ... I couldn't stop smiling! That was, until I saw someone else had a keyword1keyword2keyword3.co.uk domain that just pushed me out of the 3rd spot on a much better keyphrase ... :(

Just noticed that my page rank has changed from 4 to 6. I wasn't expecting especially after the big drops I saw last week thanks to Dewey.

I have a two word phrase that I keep track of that is the top desired ranking phrase in my market but not very likely I will ever achieve it and I'm not really trying either. I just watch it to try and figure out if it's something I'm doing that causes it to go up and down and if it's rank affects the rank of other phrases with those two words included.

This month it gravitated from #272 to #99 and back again and it wasn't anything I did to my site cause my daughter was in the hospital most of this month and needing a lung transplant. I've been too busy to do anything with my site for at least 6 weeks so it's not anything I'm doing that is causing that fluctuation.

sorry for sounding stupid. i'm new to all this. what is a DC?

Welcome to the forums, hitchhen1.

The term "dc" is short for "data center". Loosely, we tend to say data center when we really mean "IP address". Google does re-assign their collection of IP addresses (around 40 class-c blocks) to different physical data centers from time to time, but we can't directly observe that process,

What we can see is that when we go to various Google IP address directly (instead of using a google domain, such as goog;e.com or google.co.uk) the results for the same search can be different. Because of load-balancing technology, a search done at google.com at different times can still be routed to a different data center, and so the results are not always the same. We sometimes can see changes in search results, even within minutes or seconds, but the IP address (data center) may have changed when that happens.

Here are a couple of recent threads that can catch you up a bit more:

Current State of Google Data Center Technology [webmasterworld.com]
How Do I Know Which Google Center I am Looking At [webmasterworld.com]

In the Hot Topics [webmasterworld.com], which is always pinned to the top of this forum's index page, you can find a section about Data Centers - and a lot more.

Tedster,

So will different data centers hold different information or it generally the same information? How does google decide what data center to call upon?

Thanks

It varies throughout time - nothing dependable here. Sometimes the data centers all pretty much the same and sometimes there's wide variation, especially when google is working on a major new initiative. Sometimes a data center can be taken out of rotation for experiments, and you only can access it directly.

The load balancing logic Google uses to determine which data center to tap for any particular query is quite complex - and I don't have any further detail about that.

Not much happening on the DCs. Backlinks update and Toolbar PR update running very smooth. So it seems we have arrived to a calm period until July where things start to happen again ;-)

A content site I have on the side that has been getting ~3K/day google searches since January lost 99% of traffic at 2 PM. No changes done recently to trigger this.

Ironically, the site's PR went up 1 point yesterday and gained nice PR on the inside pages. This shows the meaningless of toolbar PR. Also, the site is still very well indexed.

Before January the site would have a few good weeks here and there and then drop of the map so maybe it will spike again down the road.

There are two pages that keep moving back and forth to from above my main page to below it. I am very suspicious they do some link buying and wonder if Google is playing around with some algos they think might catch that.

I'm still seeing movement two or three times a day not just in data centers but on Google.com.

mostly bookmark sites capturing the top SERPS results.

<continued here: [webmasterworld.com...]>

[edited by: tedster at 3:18 pm (utc) on May 1, 2008]