Google Updates and SERP Changes - April 2010

[wmxpert]

< continued from [webmasterworld.com...] >

We are in the same situation, since March 15, 2010 our traffic is continuously declining with the average of 6% every week. In current week it gets worst, approx. 15% decline!

Overall, 30% to 35% traffic has been reduced compare to peak week of January 2010.

"site:" is also showing huge drop in index pages, though it only provide the avg. not actual.

Don't know what's going on, Is it really the effect of Caffeine roll-out?

[edited by: tedster at 9:46 pm (utc) on Apr 1, 2010]

[panicbutton]

"No way Google would have got off the ground if this was their organic offering in the beginning." !?!?!

I beg to differ.

Pre-google I recall that most SE results were so woeful that users were happy to spend 10 minutes digging through Yahoos! hierarchical mess to find what they wanted. Google solved that brilliantly and they still dish up the right results nearly all the time. Frankly, they do an awesome job of serving pretty damn good results in the face of constant attempts to fiddle the results by people like you and me.

[dusky]

Some deep crawling going on, particularly on the sites that blipped on the 15th, I mean massive, even dwarfing yandex's crawl rate by 100s of %s, you can hear the CPUs grinding and squeaking, well you can imagine them to be at least. I had to increase the smarty cache timeout for one of the sites to make sure the CPUs aren't overwhelmed even though on this particular server, they are dual quad core with plenty of memory with well optimized sustems and software!

Is G* trying one of my servers patience?, hey G* "A you taoking to me?"

Anyone else being deeply machine gunned by Gbot the last 24h!

[dusky]

systems I mean, I should spell check, but I meant the de niro catch phrase, I like it that way!

[Jeffhyde]

Ok... Let's keep the discussion on what exactly the current situation with the SERPS is and this new update, instead of the conspiracy theories on the reasoning behind it.

What we do know:
An update to the algorithm has been put in place. If it is "Caffeine" it is dedicated to speed and accuracy. In reality it is simply putting more focus on certain attributes or less focus on other. What are these attributes?

What we need to figure out is:
What is the main difference in the algorithm that is causing leading websites to crash and dismal websites to prosper?
(from my personal experience with these results is a large amount of spammy, highbacklink, low content websites are being pushed to the top)
Is it fresh content?
Is it the quality of backlinks?
Is it the quantity of the backlinks?
Are websites with backlinks from untrustworthy sources being punished?
Is it the loading time of the website?

I think a good way to gain a perspective on this is to check the number 1 site for a number of different niches and compare the new SERPS to the old SERPS. We can do this by comparing chrome to firefox which offer both versions of the results. I have noticed chrome gives the new results, firefox delivers the old results. We should also be comparing the PR from the two serp results.

[1script]

@dusky:

Is G* trying one of my servers patience?

I see that from time to time. They come down hard on one site as if checking whether people are asleep at the wheel. It is often a harbinger of change although, sadly, not always for the better :)

Yes, I see some heightened Gbot activity on some of my sites killed on the 15th of March. Approx 7-10 times of what was the norm after the 15th. However, I have more activity still on sites that were not affected, so it could just be the regular exercise.

Another note about these Gbot binges: when they come, they pummel left and right throwing all sorts of garbage URLs along with the actual working ones. Not necessarily made-up garbage that would make me think it's an attach masked as Gbot but some of the URLs that actually existed months or years before and were either 410-ed, 301-ed or umm... sort of abandoned (as in oh, #*$!x, I forgot to fix that!). On a normal day Gbot mostly asks for good, recent, clean URLs. On a binge day like this they rub in every mistake I ever made in an .htaccess, linking structure or leaving temp files out in the open. Watch out for those spikes!

[1script]

Sorry, finger twitch:

would make me think it's an attach

attack, I hope it was clear in the context.

[dusky]

1script, I forgot to say, you're right about all sorts of old, deleted, redirected URLs and even the bogus non-existent URLs, seen them do it before and like you said, helps one to fix what's forgotten to fix. I managed to find a major security vulnerability (vis-a-vis SEO and SERPs) and fixed it, thanks to gbot, it kept asking for a non-existent URLs and was getting a 200 found header. The problem was in my rewrite rules, imagine if a bot was appending thousands of /?q=whatever-they-like-here as query strings and get a 200 header, which means toolbar users and the people G* can follow will lead to G* finding non-existent URLs giving header found responses, hence put a temp ban on indexing, phfeww thanks G*!

Back to the update, neither chrome, nor firefox or other browser giving any different results at the moment, all seem to give identical results (FF, IE, GC, Safari and Opera) on G*.com and have been for more than a day, not logged in any one of them or have PS by the way!

[Kelowna]

I have noticed chrome gives the new results, firefox delivers the old results.

I dont know about that but for the last few days aol is showing different serps than comcast with comcast being closer to the old results. In some cases the serps are getting close to the same so it is getting harder to tell. I think from watching this unfold that the aol type results may be the final product that the gorg is after.

[cangoou]

Hm. I have a lot of sites which where doing well in the last weeks but are now back where they were in january. What would you do? Just sit and wait and hope that the good positions were based on Caffeine and come back when (if?) G gets things managed? Or increasing the search for links (and therefore dropping the work for content - you can only do one thing at a time) because what we see is about to last for some time?

If Gs intention was to confuse us, they did very well ;-)

[c41lum]

Yeah Im see different results in Chorme compared to firefox and Safari.

[internetheaven]

... getting a bit sick of this flux ... surely Google's market share is going to be down again to Bing from this? How long are users going to put up with seeing different results throughout the day? All week I've been rotating from fresh good results to fresh spam results to old authority results (Feb/March) to even older naff results (Dec/Jan) and then another set of results that I can't even define that are somewhere in between all four ...

That's not to mention the endless, and often useless: "we're not showing you results for what you typed in, here are the results for what you actually meant ..."

There are so many words, local dialects, catchphrases, company names, musician/band names etc. etc. -- Google really thinks that it knows everything? What was wrong with "did you mean?"

[bowdeni]

(first post)

Just thought I'd add my frustration with the search results over the past couple of weeks. Yes I've receiving different results in different browsers, but I don't believe the browser used determines which index/algo is used.

As with 'internetheaven' above, I'm getting some older results from Feb/March, and throughout the day one my clients will rank 6th, 16th, and 24th for a competitive term.

Looking forward to Google settling down and serving universal results to everyone. It's making reporting a little more problematic!

[member22]

I agree with everyone can google make an announcement of what is going on and when everything is going to be settled. For the last 3 months I have been getting up every morning hoping for something to settle and the only news I got mid March is that we are weeks away from it... anything new...

[edited by: tedster at 4:56 pm (utc) on Apr 22, 2010]

[bwnbwn]

My thoughts for what they are worth.

Over the years it has been G's method IMO of checking how when certian filters/algo rankings effect the serps these filters/factors are added or taken out, and what effect each or combinations have on the serps.
Live testing in the serps of their effects when added or taken out.

I have a hunch this is what is going on now. Why sites are popping in and out, why spam sites are popping into the serps were they don't belong, and why established white hat sites maybe getting wacked only to come back then go away.

Do I have proof no, but from watching this happen over and over again especially when there is a major change I can only assume what is happening. It does make sense to test each filter/algo ranking factor to see what effect this will have on Caffeine then study the serps and revert back to make tweeks.

[dusky]

Yes, I second that, I wrote here on this thread while back when they started doing this and I suspected that's what's going on, testing their new caffeine algo live, I called it the Rock-n-Roll / rock and roll effect, I understood it to be rocking the SERPs, knocking major white hat sites sending them to the rank abyss. Consequently, low quality and content sites are pushed up for few days or even few weeks for some sites with 1m+ pages, then rolling back the clean results once they did their tests which would've meant additions and subtractions to SERPs. Problem is, they are still doing that on a much larger scale now to the point where you can't do a cache flow forecast or a business plan / strategy based on website rank and traffic, here today and gone tomorrow!

[bwnbwn]

dusky now that you said that yes I do remember that post. I think it was about the time we were getting all the updates on each and every ip Google owns where they could see Caffeine results and it was buried so fast it just didn't sink in.

[gford]

dusky - I am seeing a massive increase in gbot activity as well and been hit with a big downturn in traffic to the one site that is now getting massively crawled.

The sites that lost no gtraffic have seen average to a bit less than average gbot visits.

Something is amiss and the hourly flux on search results is puzzling and frustrating.

[member22]

"Over the years it has been G's method IMO of checking how when certian filters/algo rankings effect the serps these filters/factors are added or taken out, and what effect each or combinations have on the serps. "

I agree with that it seems that is what they are doing but what is bizzard for one keyword I now rank 1 st and I was 5 th before and for the other keyword I rank on 10 th page and was on the 3 rd one before. How can that be see that it is the same site and same expression + one word ?

I guess the filters still needs some tweaking so that rank 1st everywhere ;)

[gouri]

Is anyone seeing a big difference in rankings for long-tail keywords on the Caffeine datacenter vs. Google.com?

It seems as if there is some sort of filter or a somewhat different algorithm being used on Google.com.

[internetheaven]

It seems as if there is some sort of filter or a somewhat different algorithm being used on Google.com

Everything is ... like ... different dude. Everything! Google has gone all ... like ... cosmic. It's ... like ... there, but not there. You know?

*cough*

You need to back off with the control, man! Google search results should be free to express themselves however they want.

Cafffffffeine. Ca ... feen. Ca-ca, ca-ca, ca-caffeine! Moo-ca-ca, moo-ca-ca, moo-ca-ca, moo-ca-ca, moo-ca-ca-ca-caffeine!

............. caffeine ........

Peace and love to all.

[gouri]

The reason I posted is I am seeing this more for long-tail keywords than keywords.

I was wondering if anyone else has noticed something similar.

[drall]

Site that took a traffic hit from all this mess and hasnt seen the deepcrawl since Jan 12th has been getting drilled by gbot for the last few days. Actually surpassed slurp and ms bot combined.

Maybe they pushed a big part of the new caff setup out and have released the deepcrawl again to start building data back up?

[kd454]

Everything is ... like ... different dude. Everything! Google has gone all ... like ... cosmic. It's ... like ... there, but not there. You know?

*cough*

You need to back off with the control, man! Google search results should be free to express themselves however they want.

Cafffffffeine. Ca ... feen. Ca-ca, ca-ca, ca-caffeine! Moo-ca-ca, moo-ca-ca, moo-ca-ca, moo-ca-ca, moo-ca-ca-ca-caffeine!

............. caffeine ........

Peace and love to all.

That was quite funny IMO.

[whitenight]

For the UNOFFICIAL record, I'm calling Caffeine
UNOFFICIALLY launched.

Why?

** It's on 80% of all DCs.
The remaining 20% have been getting Caff datasets flickering for the past 24-48 hours.
The "old" SERPs and Algo are now the minority.

** At the time, I couldn't be bothered to post and then argue with the masses,
but the INFRASTRUCTURE (NOT datasets) has been fully live and operational for ALL DCs for about a week and half now.

since testing the INFRASTRUCTURE is far more complicated than observing datasets, you'll forgive me if i don't give any practical examples... other than pointing out i've given past examples of how ANYONE can do this on their own in the earliest Caff threads.


** Contrary to many posters persistent arguments, the Caff Datasets are using the same ALGO that was featured on sandbox.gorg.com

Will it be "tweaked"? Of course. At some point.



** Should one expect current differences between the ALGO in Caff and the "old" ALGO to be "fixed" in the future?

I would NOT bet on it.

Yes, it MAY change. But sitting around waiting for whatever changes one thinks
Gorg SHOULD re-implement into the new SERPs is not a wise idea, IMO.

As always, I give the warning that the SERPs don't lie, they can NOT lie.
Liked or disliked, they are what they are.
And the current SERPs ARE the "Truth" as far as rankings go.


Let the moaning/celebrating begin.

Big Daddy is dead. Long live Caff. :)

[arizonadude]

I was wondering where you went whitenight!

I agree that the beast has been unleashed. I've been seeing it for the last week almost everytimg I've checked it.

And yes, long live Caff!

[gouri]

As always, I give the warning that the SERPs don't lie, they can NOT lie.
Liked or disliked, they are what they are.
And the current SERPs ARE the "Truth" as far as rankings go.

Whitenight,

That was great information in your post.

-----------

For keywords and long-tail keywords that are showing some differences in rankings between the Caffeine datacenter and Google.com, can we expect the rankings on Google.com to become more like what we see on the Caffeine datacenter as the Caffeine datasets finish the tweaking that they might be undergoing now that the infrastructure is in place?

[edited by: gouri at 1:22 am (utc) on Apr 23, 2010]

[whitenight]

For keywords and long-tail keywords that are showing some differences in rankings between the Caffeine datacenter and Google.com, can we expect the rankings on Google.com to become more like what we see on the Caffeine datacenter

Yes, exactly.

Now (actually 2-3 months ago) is the time to figure out WHY the rankings are different.

Then get yourself "liked" by the new Algo.
ie. what factors are re-weighted in Caff and to what degree.

hint - understanding this will also affect the "non" long-tail, competitive keywords.
It's just easier to see/test with long-tail for obvious reasons.

[walkman]

** It's on 80% of all DCs.
The remaining 20% have been getting Caff datasets flickering for the past 24-48 hours.

I agree that it's on almost all but strangely I have checked on google.com, www2 and www3 with several googledance tools and in my own area and I don't see it there. Maybe those 20% of DCs without the caffeine are getting more traffic?

[cangoou]

Hi whitenight, good to hear from you. I still see 2 different sets of data here in Europe which have been switched 3 times this week.

Then get yourself "liked" by the new Algo.

It still seems to me that keyword-variaty is not so important anymore. Sites which ranked good in the "old" SERPs for "widget" and which got more links like "green widget", "good widgets", "cheap widgets" are kicked back while sites which got straight "widget"-links are preferred now. Can anybody confirm this? Seems like a step back somehow if it's true...

[internetheaven]

For the UNOFFICIAL record, I'm calling Caffeine
UNOFFICIALLY launched.

Why?

** It's on 80% of all DCs.

But here in the UK users are only seeing it about 10% of the time. I officially reject your official launch status. It's like Google has gone: 5 .. 4 .. 3 .. 2 ... oh, wait ... 10 .. 9 .. 8 .. 7 .. 6 ... hang, on ... 10 .. 9 .. 8 .. 7 ... actually, no ... 10 .. 9 .. with UK Caffeine rollout.