| 9:39 pm on Apr 28, 2006 (gmt 0)|
>And I thought that we are going to have a calm Friday. No way. DCs are moving again.<
I'm just not seeing that much churn except on a few DC's. Some are always in motion, testing etc.
| 10:10 pm on Apr 28, 2006 (gmt 0)|
"I'm just not seeing that much churn except on a few DC's. Some are always in motion, testing etc."
In fact I see today for the first time a site which has been for ages #1 on the serps for my test search keywords, just disappear on all DCs. Can't find it on the first 100 listings. Of course it might return back the next few days.
We might call that a #1 site being hit during testing etc.. :-)
| 6:48 am on Apr 29, 2006 (gmt 0)|
We Live.We Learn..We Share..We Make Friends..
The Passion of Google Datacenters Watching
Good morning Folks
While the DCs are displaying this morning different serps than the ones I saw before going to bed last night, I wish to write few words today about Chess.
I'm a member and webmaster of our local Chess club. For 10 years ago we were 64. Today we are only 40. We don't see many new youngsters joining our club anymore. Today the youngsters prefer online life; listening to MP3, chat, dating, gaming etc..
When there is an international Chess Tournament going on, we spend many whole evenings at the club in analyzing ongoing games and positions and making some predictions on next moves. Tons of coffee and cigarettes consumed during such evenings. And most important, the social value of such evenings is very very rewarding.
However, we all know that our analysis wouldn't change anything in any ongoing or ended game. And we wouldn't have any influence on the players in such tournaments. We do it because we are passionate Chess players on all levels and ratings: We Live..We Learn..We Share..We make friends...
The Passion of Google Datacenters Watching
It isn't only a passion but also discipline to observe, analyze and posting remarks about the DCs in general or specific DCs in particular. And patience and focus is the name of the game.
In fact watching and analyzing changes on the DCs of high importance, IMO. Because it tells us much about what do we expect tomorrow's serps to look like, and might take actions regarding our sites in good time.
And as you might have noticed, watching google datacenters is a very educating process. Take a look at some posts on different parts of this thread and you shall see important topics as canonical issues, supplemental issues, 301 redirect etc. explained in details.
IMO, this thread reflects the huge high quality resources this great WebmasterWorld community has.
Some of us are addicted DCs Watchers and we do that with passion and there is no doubt about our spirit of sharing our observations with the rest of our kind fellow members.
We Live.We Learn..We Share..We Make Friends..
Wish you all a great day and a happy DCs watching..
| 8:35 am on Apr 29, 2006 (gmt 0)|
Well said reseller. :-)
I suppose the real reason people watch DCs is to be first in on the action when something (finally, lol) happens.
| 1:38 pm on Apr 30, 2006 (gmt 0)|
reseller, if you are taking a nap after lunch, please take a cup of cappuccino. My default google.com (126.96.36.199) is showing the results of those wonderful DCs which were ranking sandboxed sites high up the SERPs. IMO that set of results will spread across, with the usual flux.
| 1:58 pm on Apr 30, 2006 (gmt 0)|
Honestly, yes. I took a nap and on my way to take a Cappuccino. You must be a Psychic :-)
I recall GoogleGuy once mentioned "The Sharp eyes of McMohan" ;-)
So you could be right that 188.8.131.52 will spread across the DCs.
| 4:51 pm on Apr 30, 2006 (gmt 0)|
Why does [184.108.40.206...] show such a great difference in results? Our results is simply great on this DC. Should I keep an eye on the DC, do you think G will launch this DC's results eventually?
| 5:26 am on May 1, 2006 (gmt 0)|
|I recall GoogleGuy once mentioned "The Sharp eyes of McMohan" |
Thanks reseller. You have a good memory. Guess, I just happened to notice changes before anyone did those couple of times. Since I watch more than one site, it is easier to spot a change in SERPs across the board.
By the way, 220.127.116.11 was showing the results of 18.104.22.168 when I posted, but now it is no more there. Google must have been tracking the user behaviour on a safe Sunday.
Anyway, have a nice week ahead :-)
| 6:03 am on May 1, 2006 (gmt 0)|
Good morning McMohan
Good morning Folks
Rumour has it that Matt Cutts and his WebSpam Team have been working the whole weekend on cleaning spam of the following two DCs sets. And thats also explain why Inigo hasn't posted on his blog since.
Talking about Matt's blog, I see few folks don't like Matt talking about whitehats and blackhats. It shouldn't be difficult to guess why :-)
Wish you all a great day and most successful week.
| 12:08 pm on May 1, 2006 (gmt 0)|
I still think that 22.214.171.124 is still the cleanest DC out thier, correct me if I am wrong.
| 1:06 pm on May 1, 2006 (gmt 0)|
I don't keep track of the indivdual datacenters but keep a running tally, for statistical purposes. For the past few weeks, their have been two clear groupings of results for our keywords. One in the range 10-15 and the other in the range of 45-54.
I can tell you that the better results have moved from 2 to 3 datacenters, to 12 datacenters as of today, and my traffic is increasing to prove it.
I don't think we should concentrate on individual datacenters, as they change quite frequently but on the big picture.
One thing we have never concentrated on has been recip. links. We simply refused to play the game, and have depend on natural linking.
We have been hurt drastically since Florida, due to off page factors, and it appears that the era of recip linking to get good search results, is nearing its end.
Its good to see that Google is finally rewarding sites again for content and not for silly offpage factors.
| 1:28 pm on May 1, 2006 (gmt 0)|
For some reason though G is not indexing regularly our site. The have indexed 24 pages so far out of 400 pages. In all G dc's we show 24 regular pages and the rest approx. 600 of them in supplemental results. However in 126.96.36.199 if I do a search for our site: only 24 pages come up and no supplemental results. Can someone help me in getting our site indexed? We have been indexed regularly until now and had no problems but unfurtunately all of our pages vanished.
| 3:37 pm on May 1, 2006 (gmt 0)|
For Monday May 1, here is the results I have for the New Google Data Set. This dataset was first seen in 188.8.131.52 & 104 a few weeks ago, and have progessively gained strengh, and is now on 12 datacenters.
I imagine, over the next few weeks, this dataset will replace the old serps, at the current rate of conversion.
For good or bad, I believe this is what you are going to get.
| 4:35 pm on May 1, 2006 (gmt 0)|
Although ranking has not changed yet, I am seeing a cleanup on supplemental results on this datacenter. Has anybody else seen this? I am hoping that this is the beginning of the supplemental refresh that Matt indicated would happen near term at the end of March.
| 4:52 pm on May 1, 2006 (gmt 0)|
I see a change in rankings as well as having all of our supplemental results removed.
Does the fact that we have pages in supplemental results keep back G from adding our regular url's to thier database. Ever since we have the s.r. issue G has removed most of our existing url's and has not put it up again.
Is it G problem or is thier anything that I need to do to get this problem resolved?
| 4:53 pm on May 1, 2006 (gmt 0)|
I see a great change in rankings to the better and having our supplemental results removed.
| 5:23 pm on May 1, 2006 (gmt 0)|
The stuff at [184.108.40.206...] is the "experimental" stuff that had been on [220.127.116.11...] for several weeks beforehand.
This is the data set that for a search that should return just 20 results (19 of which should be for supplemental URLs), now returns more than 900 supplemental results, 880 of which do not exactly match the query that was typed in...
| 7:22 pm on May 1, 2006 (gmt 0)|
FWIW I can confirm something interesting is definitely going on WRT this group of data centers:
I just checked a 3 word phrase that we used to name a site launched about a year ago (e.g. prefixwidgeting-name.com). All of these words are moderately competitive.
Google says that prefix widgeting name matches 38.1 million documents.
Google says that prefix widgeting matches 402 million documents.
Google says that name matches 3,090 million documents.
The 3 word phrase does not generate a very large number of searches, but is useful as a benchmark for how well the site is doing in the various search engines.
The site has been in the top 5 in MSN for the 3 word phrase since last July, and has recently been solidly in position # 1 on MSN for the 3 word phrase.
The site has received a growing trickle of traffic for less competitive "long tail" search terms from Google. The site has fluctuated in Google for the 3 word phrase -- sometimes around position 172-238, sometimes dropping completely out of sight, once briefly surfacing at #14 then dropping firmly out of sight.
On the above data centers the site is suddenly in position #7 for this 3 word keyphrase; on the other data centers it remains firmly out of sight (not in the top 200).
The remainder of the top 10 SERPs are roughly the same on all data centers, so if it is an algorithm change, it isn't dramatically shuffling the other results for this particular 3 word phrase.
I'm not sure if our site is benefiting from a change in the algo's being used on these particular data centers -- perhaps something related to the use of keywords in the domain name -- or perhaps this is just another temporary "test" of the site, but the difference between this group of data centers and the others is quite apparent and consistent, so there is definitely something different about them.
| 8:07 pm on May 1, 2006 (gmt 0)|
Brett Tabke & Matt Cutts Might Have Something To Announce Soon!
Good evening Folks
There is something in the air for sure!
It seems something good is happening as you might have noticed of the reports of the Heroes Google Datacenters Watchers. Guys you are GREAT!
And I was just about to write; Rumour has it that...
No. Instead its a feeling I have that Brett & Matt might have something important to announce soon..
[edited by: reseller at 8:10 pm (utc) on May 1, 2006]
| 8:09 pm on May 1, 2006 (gmt 0)|
|Brett Tabke & Matt Cutts Might Have Something To Announce Soon! |
The end of the "Google Data Center Watch" thread?
Just kidding with ya reseller :)
... may the watching be with you!
| 8:24 pm on May 1, 2006 (gmt 0)|
"Just kidding with ya reseller :) "
You are most welcome, Peter :-)
| 8:40 pm on May 1, 2006 (gmt 0)|
While 18.104.22.168 does show significantly less supplementals, other datacenters like 22.214.171.124 and 126.96.36.199 are showing an opposite canonical mega-meltdown, with www 301's that have been in place for two years being removed for non-www Supplementals that have been redirected since 2004.
If things go according to form, the positives will soon be swallowed up by the corrupt data that Google simply is incapable of making go away.
Ignoring 301's is sadly par for Google's course, but removing the wrong URL after TWO YEARS is simply mind boggling incompetence.
| 10:17 pm on May 1, 2006 (gmt 0)|
This was a pretty important find by g1smd and very worthy of it's own thread.
I have been studying this for the better part of a couple hours and from log files it is pretty surprising. It is clearly changing the pattern of referrals pretty radically.
I was looking on saturday trying to figure out what had changed. As soon as I read this thread - I knew exactly what had changed. (thanks!)
| 11:03 pm on May 1, 2006 (gmt 0)|
g1smd:"Google begins to look more and more like archive.org every day."
I've just done a quick 'site:' search on 188.8.131.52.
The point of indexing a page as it was on 23 Jan 2005, three months before a complete site redesign and 301 redirection, even as a supplemental, is....... what exactly?
Googleguy? Cutts? Anyone?
How is this data of use to any searcher (other than the occasional Internet historian)?
| 11:27 pm on May 1, 2006 (gmt 0)|
I'm giving up on Google. It'll drive me to the funny farm if I continue to bang my head against the wall trying to optimise for this sad excuse for a SE. They just killed 99.9% of my site and I'm having to go back to working a day job!
My advice: do NOT rely on google unless you want trouble.
| 11:32 pm on May 1, 2006 (gmt 0)|
Brett - you mention about the nature of referrals changing, what I am intrigued about is in what way, i.e. more relevant, better quality?
I say that because I am finding a large decrease in relevance but better conversion to adsense!
| 11:50 pm on May 1, 2006 (gmt 0)|
Hmmm, been out for the evening, and now I get back in, I see that [184.108.40.206...] has dropped all the supplemental results for some of my test searches (I suspect a hand-edited SERP for those). This almost looks like the Supplemental Cleanup that I have waited 2 years for (but which I have already seen in place twice before in the last couple of months - but each time it only lasted a few hours before reverting back to a mess).
This time the cleanup is "too perfect" on those searches. For many other searches that DC now shows a massive extra number of results, many of which are not an exact match to the query, even for "searches in quotes"!
Brett: the behaviour I saw in #1 only lasted a few days. At the moment there are only a few DCs that still do that action. For most (at the moment) searches, both the "new" and the "old" searches are showing a modern cache; but as always (like for the last 2 years or more), an "old" search still shows old content in the snippet (content that is no longer on the real page or in the Google cache of the page), and a "new" search only shows new content in the snippet. The "old" search is marked supplemental, and the "new" search is not.
| 12:02 am on May 2, 2006 (gmt 0)|
So,if you alter a page, Google will return that page for the current content but it will also return that page if you search for the previous version of the content. Before today, you could only see a modern copy of the cache. Now, you get to see either a new copy or the old copy depending on exactly what you searched for.
Have I got this right then? Google are returning results based on old content, when they are aware the content has been changed and sending users to pages that may not be relevant anymore?
| 12:10 am on May 2, 2006 (gmt 0)|
It might be useful if I repeat a post here, from earlier in the day in some other thread.
There are many types of Supplemental Result:
- There is the non-www Supplemental result where the www version of the same page is showing as a fully indexed normal result. This type is splitting PR and almost always having an effect on ranking. Some pages of the www site will be supplemental and some will be normal, and some pages of the non-www site will be supplemental and some will be normal. The cure is to add the 301 redirect, so that all the non-www pages are dropped.
- There is the Supplemental result that shows for a page that is now 404, or where the domain has expired. These are usually just clutter in the SERPs, especially for those where Google no longer has a cached copy of the page. Some do, some don't.
- There is the supplemental result that represents the previous version of the content on the page, where the current content appears as a normal result. That is, have a page with a phone number on it, and then alter that phone number. Do a search a few weeks later for both numbers. If you look for the new phone number, you see a normal result with the phone number in the snippet, and the cache also shows the new phone number. If you then look for the old phone number, you get the same page returned in the SERPs, but now with the old phone number in the snippet, but still pointing to a more modern cache that only shows the new phone number. This result appears as Supplemental. These are not affecting PR, but for a site where contact details, prices, etc have changed you have ample opportunity for surfers to be confused. For a few days, some DCs have been pointing "old" search results to an old copy of the cache for that page, while a "new" search shows a new cache.
- There are the supplemental results that show when you have a site where many pages are classed as duplicate content of other pages on the same site. This occurs when the title and/or meta description and/or on-page content is too similar. The way to fix this is to make sure that every page has a unique title and description, and each page has different content. If you have a load of similar pages, use the <meta name="robots" content="noindex"> tag to keep all but one of them out of the search results.
- There may be other types, but there are often times when a site is hit by a combination of the above types.
| This 112 message thread spans 4 pages: 112 (  2 3 4 ) > > |