Forum Moderators: Robert Charlton & goodroi
I copied your post as I agree it is not on topic and I was hoping more people would have seen this thread, but it dissapeared rather quickly to page 2. Sorry for not staying on topic with the other thread.
I posted this yesterday and maybe people that are dealing with this can provide input as it is really a different issue all together.
"Its early in the game, but 100% full recovery."
Off topic from this thread, but I see some September 22/December 27 things fixed, while others got hurt more.
Yet another screwup, but it should make some people happy to see some pages recover. I had three pages still effected by September 22 that weren't healed on December 27th. Two (totally trivial ones) healed today, while the other (very important) one was joined in the screwedup box by another (very important) one. Cool, not.
-------------------------------------
I tanked on September 22 as well, then on J3 recovered until December 27th. Tanked in serps from that point until yesterday and recovered.
Hmm, I know that I have been making a lot of changes since december 27th, basic cleaning of links and content. I would love to say that is what did it, or maybe it was my endless month of suibmitting spam and dissatisfied reults, but I have noticed a trend with the index. For me it seems that these "data refreshes" are kind of the "under the hood" serp changes that Matt was saying that over time would be implimented on big daddy.
One of the reasons I was posting in the BD threads, but it is clearly a different issue.
I should also point out that I also did a 301 on a site and that might have something to do with it, but I am not convinced about that yet.
JuniorOptimizer, my site suffered the same fate as you, down on Sept 22nd, recovered for Nov/Dec and then down again on Dec 27th. Almost 3 months later and traffic charts are still flatlined.
In the back of my mind I have to keep telling myself that all of this is temporary due to BigDaddy and that once this is in place we will all see a more stable set of SERPs.
I took a cool $1,500 per month hit on this one, but like a cockroach I've adapted and found the money elsewhere. If and when this ever comes back, it's like a bonus. The site went from being the most relevant one to being a pariah as far as Google was concerned :)
This one is one of the more mysterious Googlepoofs (TM) since there seems no real reason for it, nor a reason why one thing versus another recovers.
===
In terms of the thread title though, there for sure is no "rollback". It's just a change that effected many things (with Sept 22 in common) differently.
Interesting, but I would have to see it as a rolback. The reason I say this is that I have historical data that is showing me that what I see this morning is mirroring the days prior to September 22, prior to December 27th and today. The only difference is the result count from the index.
Maybe it is a rollback with a twist? meaning it rolled back, but also included an index increase that keeps the newer pages below page 1, then maybe a gradual serp change over time would include those newley added pages to a more appropriate ranking?
Just speculation of course. :)
However, after dropping in on the must-get-a-life (AKA "Bigdaddy") thread -- will that ever end? -- and checking some of the data centers that have been mentioned, I still see some of them with the pre-reshuffle results. So maybe for some reason you've been getting directed to one of those datacenters, 300m.
Good point.
However, in the big daddy thread someone asked me that and i told them that it was all datacenters. Not just one.
see below;
maha Which DC are you looking at?
>Its early in the game, but 100% full recovery.
300m All of them. See some may recall that on December 27th all datacenters underwent a data refresh. Matt Cutts posted this on his blog and stated that it was unrelated to big daddy. With that being said, there was very little in the way of discussion about it because everyone has been focusing on big daddy. This morning at least for the keywords I track, everything on all data centers has changed.
All of them = several big daddy, several other non big daddy Ip addresses, any data center I have checked.
I think there is something wrong with Googles page rank calculation in my sector. You can't have 404's and malicious redirects ranking well for allinanchor.
I have watched my site disappear and reappear on these specified dates for about 1.5 years. Until recently, my whole site would be penalized. But since Oct 2005, it has only been individual pages that drop, then return a month or two later to their previous position.
Recent dates where I have seen this same activity are 9/22/05, 10/15/05 (seemed like an adjustment to 9/22), 12/27/05, and now 3/8/06
I don't mind it so much when it affects individual pages, but when it drops the entire site it is rough.
eventually some SERP's are close to 'not good at all', I have a hard time understanding niche authority sites (that's my standpoint) that drop ....Yep mr GG please explain me this one.
if this is a roll back there are definitely some casualties and i don't see the improvement AT ALL..
my 2 cents
I think there is something wrong with Googles page rank calculation in my sector.
Indeed - that is what I have been thinking is one of the underlying issues that has resulted in the problems Google has had for a while.
Canonical issues etc dont help this calculation.
However, IF <RK> values as discussed on a recent thread here is indeed PR then in Big Daddy DCs those <RK> calculation looks like Google maybe correctly calculating PR again - after a long term of absence or problems for some sites.
Not sure that this re-calculation has been applied to the serps as yet though.
Steveb - did you look at the <rk> thread - what are your thoughts on this value, do you think it is a lot of Hooey, and not of interest?
I can agree with that concept. That was without a doubt one of the most informative threads that I have read in a long time. Maybe the rk is the PR and they could not apply it to the data centers because of the new infastructure? And they have finally been able to resolve what ever was the issue and push it out on a test run to see what happens?
I am watching a site that has been PR0 for a while, which shows <RK> 0 on Non Big Daddy DCs and <RK> 5 on Big Daddy DCs - and no changes of rank anywhere....
Although of course sites maybe at different stages of a possible change/fix/update etc.
Visitors and rankings dipped in August, tanked in September (21-22), stayed exactly there and are back since yesterday, almost exactly where they were before.
Supplemental results (caused by non-ww whcih appeared in August I suspect) are still there on all DCs. But rankings have returned despite them.
I was hit hard on 9/23 and then destroyed on 12/27. Went from top 3 position on all kw's to pages 2 - 9.
As of yesterday morning, I am back, and in some cases, in better position.
300M - To me, this is not a roll-back as I have actually increased my position ranking on a few terms, and see newcomers to page 1 that were certainly not there on 12/27.
My opinion is that G "fixed" a penalty filter that was harming good sites, and it has to do with cannonical issues. As of Monday night, a site:www.domain.com command had my home page nowhere to be found. Come Tuesday morning, my SERPS return and the site: command has my home page first.
I have 75,000 links, so I don't think new links did anything here. I did clean up some source code and resolved some "spammy" cross network anchor text strategies. Perhaps I was tripping a filter there. Who knows!
I'm not really seeing any improvements today. I'm seeing 1 page in particular that dropped has returned to #1 (an .html page) but pretty much every other page I've checked (all .php) is nowhere in the top 100 where I used to be.
I'll be keeping my eyes on this for the next couple of days.
But I have to say that I saw something similar last spring. And now, as then, I can see no pattern to what has recovered and what has not.
This business just seems to keep getting more and more random, as far as I can see. In fact I often wonder if that isn't just what Google intends.