Okay, we`re back. But only G. knows how long will our site be in the SERPs. Will see tomorrow :)
|If we check the datacenters using IPs in the URL, the browser shouldn't send the login cookies to google. So I think this should work as well. |
That's what I always thought, too - but dudibob reports seeing different results. I can't easily test this because I intentionally avoid doing searches while logged in.
There is one other possible explanation - Matt Cutts explained that even the same IP does not always point to the same data center.
wow..I never realized that. I get totally different results when I am logged into my gmail account. hmmm. Don't know how I feel about that one...
Is your Browse History turned on?
when I log into my gmail I get one set of results. When I log out I get another set of results..
Exactly. I thought it was important to highlight that fact since one of the phenomena we're talking about right now is rapidly changing SERPs. We should be talking only about observations while logged out - the rest would be tailored to each person and not of general value.
[edited by: tedster at 6:30 pm (utc) on April 22, 2008]
|And officially, Matt Cutts did not confirm this as a new algo... |
Hard to understand this, except that part of what I'm seeing appears to be be an ongoing link re-evaluation of the kind that has been going on for years, only now it's being increasingly applied to less competitive searches... perhaps factors and filters related to authority, affiliation, and trust.
The diversity thing is what appears to be new, but again, Matt has been mentioning for quite some time how Google likes to show "different kinds of sites" in the top results.
(I should add that I can't find any reference to anything Matt has said about diversity, but I know I've been referring to the idea of different kinds of sites for a while. It's most likely he said it at a conference. Maybe someone else can dig out something that's online).
For the record, my experience (that has gone on for weeks now) of ranking page 4 at night and ranking page 8 during the day happens regardless of whether I'm logged into a google account and also happens on multiple computers at multiple ip addresses across the country [I have reports from all over on this keyword and website pair].
I'd say that it's a red herring that the results are different when logged into google accounts and that this "warhol effect" is an overall SERPS phenomenon.
|Seems price comparison sites have now taken the lead over the actual ecommerce sites :( I hope this won't be the norm! |
Seems to be the norm in many areas now. Itís actually becoming difficult to purchase things off Google US. Even when you attach words like buy and purchase the comparison sites are driving the e-commerce sites way down. I've had to go to other engines to find seller sites I've forgotten the name of.
Today with Google I also noticed with four keywords exactly 9 of the top 10 sites were all Adsense but mine. Now I'm sure itís not that way in all areas (it would be nearly impossible) but this does seem greater than random. I've been expecting Google to turn a handsome profit, as they did, because of this. They also see the comparison sites as having pretty much unlimited "deep pockets" especially in the Adwords auction arena.
Google also in the last 3-4 days seems to be tinkering with lesser keywords in my areas as more sites vanish from the top results.
|So the logged in results are not so useful for understanding any changes in the ranking algo. |
It would be funny if they had an algo to sense for site owners interested in certain terms and then delivered bogus results to wind them up.
Come to think of it I think I've been got by this one ;)
Hissingsid- I have a sinking feeling that you might actually be on to something. That's kind of scary!
|It would be funny if they had an algo to sense for site owners interested in certain terms and then delivered bogus results to wind them up. |
Come to think of it I think I've been got by this one ;
I seem to have read somewhere - I hope someone can find it - that an algo enhancement a few quarters back was able to determine what KW(s) every site was trying to optimize for by looking at tweaks made. If the tweaks were either numerous or constant a "hold' type penalty would be applied...
I also remember in the same article reading that it could take anywhere from 6-9 months to work itself off...
As an aside.... I always thought that to have several "Rotating Algos" would be a great way to control spam since it would like trying to hit a moving target.
I think they have just introduced an element of randomness to the results.
I check for a particular phrase and our page shows at 37. I then go to my office in town (2 miles away) and the same page/ same phrase is 53. Same browser/ computer google version, not signed in etc. Everything the same but the results different - and stable all times of day for several days now.
|We should be talking only about observations while logged out - the rest would be tailored to each person and not of general value. |
I thought it is only a problem if you are signed up for "Web History". Or does Google track all our visits no matter what items we are signed up for?
|Or does Google track all our visits no matter what items we are signed up for? |
Google tracks everything...
One of the scary things about this monolith, IMO. I would not be surprised if being logged in or out is moot if you can be identified by IP address.
After last night when our site was back in the SERPs, today it is gone again.
celerityfm: is this still happening to your site?
"seem to have read somewhere - I hope someone can find it - that an algo enhancement a few quarters back was able to determine what KW(s) every site was trying to optimize for by looking at tweaks made. If the tweaks were either numerous or constant a "hold' type penalty would be applied..."
This would make sence as to why my one site dropped so bad but my other stayed constant...I have been tweaking a lot on the one and it suddenly has dropped and not much movement....
Oh, yes, Google is very much aware of competitive keywords. Its algo triangluates three major data streams: 1) keyword search frequency; 2) AdWords bidding rates; and 3) overoptimized sites. (It may also have many minor data streams to use.)
Google for months has had its own special algo for competitive single words and some of us suspect it applied that same algo to competitive phrases.
Random results at the top is somewhat reasonable in theory as a continuation of the Competitive Search Algo, but the only reasonable application of it is to scale results if Google wants to tie it into its new diversity idea.
I think there is a new Matrix of ranking. The matrix merges more data points than we're used to, so it's much more difficult to break down and figure out. You can't isolate the variables the same way.
Any PhDs here to figure it out for us?
The DCs look very calm. It could be Google Update Dewey is winding down!
Did anyone make any dramatic changes to their site before getting zapped by this Update/Test/Glitch/Effect?
I did. On the only site I have that I'm aware was hit.
Page Count: 200
Original Ranking: Top 10 for over six months for keyword "example."
Latest Ranking: None for example
Site Changes: Added "- Example.com" sitewide to all titles
e.g: The Quick Brown Fox Jumps Over The Lazy Dog - Example.com
The site has very good architecture, unique content, and unique titles.
It's very common to add the domain name or site name to the end of a page title. Why should Google penalize a site for doing that?
All other phrase ranking remains stable. At first I was slipping into and out of the top 10 after "Dewey" started to wreak havoc. Lately I'm not even slipping. Gone. I searched the top 1000 SERPs (pages 1 to 10 with a 100 results/page setting), and didn't see it.
It's ridiculous for Google to give no ranking to example.com for example.
[edited by: potentialgeek at 9:53 am (utc) on April 23, 2008]
I recovered Last night in G. We went out hard on Saturday night. My domain .co.uk. I didn't change anything it looks to me like geo algo test. I thin we are all in for a bumpy ride with G.
c41lum we are also being hit ahrd in .co.uk for terms we target the most, seeing 10, 20 and 100 position drops across a wide range of keywords for most of my sites, they come and go but are mostly gone.
frustrating because a) some of therms are not even that comeptitive and I have been bumped to page 3 b) we blatently have some of the best content on this subject.
I am not being friendly to our site either, many sites now ranking do not even have the pages of tutorials, guides and explanations we have authored. It is almost like we are being penalized for trying to be helpful and informative. One page sites and directories ranking above us instead.
Hmm, well, up until today, my site showed #2 for all searches regardless of whether or not I was logged in to my Google account. Today I see that if I am logged in, I'm still at #2. If I'm not logged in, I'm at #1 for everything. So I guess they figure my site wouldn't be quite as relevant to me as it might to someone else. Heh. I think I just got served.
|We're jumping around on .co.uk. We're seeing our site rank #2 for a one word search [manufacturer's name], and then it'll drop down to somewhere in the 20s or 30s. Then later in the day, back to #2, then down again. |
It's been doing this for more than a week now.
Quite a few other phrases are doing the same, whereas others are staying put. 'Tis all very strange.
Looking into this phenomenon a little further, and something very strange is happening.
We seem to be 'swapping' places with another site for the #2 spot and the #20-odd spot. One of us ranks #2, the other in the 20s, and then we'll swap places.
The sites are completely different, no common inbound links, they look different, structured differently, and have different content. They are actually a dealer of the [manufacturer] - whereas our site covers all the manufacturers in the market.
Why would it just be the two of us who are jumping around? The rest of the results are staying the same (give or take a few places obviously).
Oh - and it's only on this one phrase that we've noticed this happening. I'm investigating other phrases, but so far none have shown the same behaviour.
Those folks being hit on google.co.uk have you checked allinanchor:search term?
In my market it has always been the case that if you did an allinanchor search on google.com you got the exactly same (or almost exactly the same) as on .co.uk. At this precise moment this is consistently not happening across all the DCs you can check at mcdar.
I wonder if this is what is causing your problems.
whatever they are doing with the geo filter knobs, they need to revisit. Doing a search from google.com in Toronto produces 8 out of 10 sites from either .co.uk, or .au extensions. I've sent in numerous forms citing update dewey - but nothing's changed.
As for shuffling around - serps were changing daily up until yesterday and today - first 2 days in a row where the results haven't been different.
|It could be Google Update Dewey is winding down! |
Not in my sector, it's gone absolutely phrenic today with nearly all pages bouncing about with some of the most authoratative at 950.
We're in for some more crazy days yet until they get this under control since it plainly and clearly is not doing what they intended it to, and if it is then they're going to start losing users rapidly since several times this week I have had to go to Yahoo! and Live to find information that Google could not.
Sure they may believe they're gaining market share with more pages however that's because people can't find what they're searching for!
Mess, another fine mess you've gotten into Google:-((
Added: And the sites of mine that were in Google.com and then had in Google.co.za and then got them back into Google.com have returned to Google.co.za!
[edited by: HuskyPup at 2:59 pm (utc) on April 23, 2008]
I had been watching the datacenters for about the last 10 days. Our #1 keyword phrase had us in position 5 -7 at every datacenter I checked. Anytime I looked on www.google.com we were always in the 70.
This morning we are getting several reports around the country that www.google.com has us in #5-7. Something changed last night, and for the good. There was a lot of non-relevant sites returned before the update this morning, and I mean a lot.
ModernM - did you take any actions to get back to the first page?
delmiha - interestingly we have stayed on page 4 for our keyword of concern for the first time during the day, whereas usually it would have dropped to page 7 (is someone listening?) :) So JUST for today it has not repeated the night/day swapping. Will advise again soon.
potentialgeek - yes, we used to rank page 1 for our keyword for weeks. Then I think we got penalized due to a partner who had very aggressively placed links to us for that keyword throughout their site. We submitted a re-inclusion request 7 weeks ago and now this is happening. Who knows :(
Hrm. Now I'm at #7 for a very generic one-word keyword that I never thought I'd rank for without my location next to it, up from around #55. Yet my sitelinks are still gone. Sniff.
| This 147 message thread spans 5 pages: < < 147 ( 1  3 4 5 ) > > |