| 1:03 pm on May 6, 2008 (gmt 0)|
I'm seeing the same effect with our site too. Following the launch of a redesign, we jumped up several of our #3-#10 places to many #1/#2 spots.
Since this w/e (3/4th May) some of our pages have now dropped from top 5 results to around 30/40!
I hope Google sort this out. The results they show now for our search terms are really quite rubbish and irrelevant.
| 1:46 pm on May 6, 2008 (gmt 0)|
I just thought I would make an update since my last post in this thread a few days ago where I noticed a big drop in traffic. I am noticing the traffic has very slowly been recovering the last few days, and today has had a nice jump back towards normal. I am thinking it was somehow related to the 2-3 days of lower than normal Googlebot activity that happened around April 22-24 for us. Googlebot seems to be back up to normal levels on our sites as well.
[edited by: Bigwebmaster at 1:47 pm (utc) on May 6, 2008]
| 3:35 pm on May 6, 2008 (gmt 0)|
Crawler speed dropping recently, switching results and so on are all linked to the new algorithm. It is a learning algorithm (AI anyone?) which is why Google are hush-hushing so much during this quite obviously HUGE software change.
The software itself is impressive for sure, but it has been given mandates that are causing the current slightly off-white results. Just take the chains off and let it go Google! I'm curious ...
| 8:45 pm on May 6, 2008 (gmt 0)|
Interesting observation internetheaven
Lets hope the AI is intelligent enough to spot AL (artificial links)
| 8:48 am on May 7, 2008 (gmt 0)|
|Lets hope the AI is intelligent enough to spot AL (artificial links) |
I think the AI move is to get away from things that can be artificially boosted such as links. Trying to create something that understands both the keyphrase being searched for and the intent of the user and matching it with a set of results. For example, mis-spellings have been dramatically adjusted over the past few months.
| 9:35 am on May 7, 2008 (gmt 0)|
Mundane after the talk of AI but I'm noticing that stem opposites are ranking.
For example search for green tablets and pages optimised for green tablet come top, search for green tablet and pages optimised for green tablets do it.
I guess this may [possibly] be because stems and synonyms add to topic weight but do not add to a sliding scale over optimisation penalty.
| 12:29 pm on May 7, 2008 (gmt 0)|
Simialr to affiliation's post...
I notice similar results with one of my biggest keywords. When I rank in top 5 for the plural I usually show at spot 10 or 11 for the singular. When I show in top 5 for the singular I drop out of top 5, usually around 8 or 9, for the plural. It seems to rotate day to day and sometimes throughout the day. This has been very consistant for about the last 6 or so months.
Yesterday is the first I can remember seeing both remain in the top 5 at the same time. Looking good today so far!
[edited by: MadeWillis at 12:31 pm (utc) on May 7, 2008]
| 1:09 pm on May 7, 2008 (gmt 0)|
HissingSid - I see the same thing in my area. End of December a site that wasn't really in the game suddenly floated to the top. Why? Because they'd spent hundreds on thousands of low rent links, blog spam, comment spam, off - topics, etc. They do now seem to be slipping and the old order is reasserting itself.
| 1:27 pm on May 7, 2008 (gmt 0)|
I guess the folks at Google wish the algo would do what they say it is doing.
Matt Cutts says words to the effect that they are going to come down hard on paid links and link spam, a few minutes later Google starts promoting sites that buy cheap links.
Perhaps they are using AI with an ironic sense of humour. It has got out of control and keeps taking the p*$$ out of the spam team.
| 3:19 pm on May 7, 2008 (gmt 0)|
I'm seeing the changes in singular, plural and ing versions. Google seems to be considering them the same now.
I guess the singular and plural don't matter so much but those two versions usually refer to the object while the ing version is about doing the activity. People are looking for quite different things in that case.
Hmmm, maybe it's more subtle than that. Maybe the algo looks at other words on a page that might relate it to the ing version. The concept that people might want to do the activity rather than buy the product.
| 5:58 pm on May 7, 2008 (gmt 0)|
I also noticed declining visitor-stats in the past week, but for the time being I'd say this has seasonal reasons: May, 1st is a day off in many countries and I suspect many people used this day plus the following weekend for a short holiday.
Did you compare last year's ups and downs? At least in my country these are the first days with some sunshine, so for many people time to fall in love and not stare at monitors.
Hope dies last;)?
| 7:06 pm on May 7, 2008 (gmt 0)|
I can tell my traffic has increased in the last few days. I also notice more pages being indexed when using the site:command in google.
| 9:24 pm on May 7, 2008 (gmt 0)|
Still seeing results jumping around like crazy here in the UK in my sector. Literally every few searches the results skip around I am even seeing one result jump from page 3 to 4th. Never seen this before on Google. Any ideas anyone? should I panic or should I just sit tight.
| 9:43 pm on May 7, 2008 (gmt 0)|
When do you think this results jumping up and down will settle finally?
In other words when do you think the storm will set down?
| 10:21 pm on May 7, 2008 (gmt 0)|
never, it's the new way - perhaps.
| 10:33 pm on May 7, 2008 (gmt 0)|
Literally every few searches the results skip around I am even seeing one result jump from page 3 to 4th.
It seems you are hitting different data centers (DCs). Different DCs receive different data at different times. To simplify, lets call it the data refreshes at Google DCs. Facinating, don't you think so ;-)
| 10:44 pm on May 7, 2008 (gmt 0)|
OK, we have waited weeks for search results to regain sanity. The results these days resemble looking at the spinning wheels of a slot machine (maybe we are have too many meetings in Vegas).
Can someone PLEASE, ask Matt to give us some kind of statement that truly relates to what we all are seeing?
So far, the silence has been very unsettling. Where is Google Guy when you need him?
| 10:57 pm on May 7, 2008 (gmt 0)|
Has anyone seen any hard consistent evidence that the public (ie. not SEOs and webmasters) are in anyway unhappy with the current Google search results.
I very much doubt it. I would bet if you carried out satisfaction results today compared with 6 months ago there would be no significant difference.
So basically why would Google care one little bit about webmasters being unhappy if their public is happy?
| 10:59 pm on May 7, 2008 (gmt 0)|
I think the up and down, round and round, constantly changing SERPs are probably here to stay or will be eventually. I also believe the public won't know it's happened.
It also fits with Googles stated aim to disrupt peoples business models.
| 11:04 pm on May 7, 2008 (gmt 0)|
> I'm seeing the changes in singular, plural and ing versions. Google seems to be considering them the same now.
Does Google really need some fancy algo to figure out most users who search for widget might not be led astray if the results are the same as widgets?
Good grief. Artificial intelligence. O-kay.
| 12:11 am on May 8, 2008 (gmt 0)|
I'm seeing some kind of stability in SERPs following Dewey weirdness after moving my site onto new servers. No idea if there's a connection. Also removed one possibly suspicious inbound link. Can't say, either, if it helped.
Supposedly it's okay to link from one of your sites to another if you aren't greedy (many crosslinks), but the link was on a 160x600 image, which we know is so prevalent as paid advertising. I wasn't convinced Google would ignore it and hold back hitting it with a paid-link penalty. I'm pretty sure Matt Cutts designed his algo to look at links from standard ad sizes.
| 3:43 am on May 8, 2008 (gmt 0)|
I am still seeing things jumping around
| 8:57 am on May 9, 2008 (gmt 0)|
For my .co.uk site the results have had big jumps over night. I have literally gone from a stable 2nd to 31st. The problem is in my sector we are being hit very hard by the credit crunch and then with google changing there results every few seconds running a online co. is almost impossible. I have been online for 10 years this year and its never been this hard.
| 10:47 am on May 9, 2008 (gmt 0)|
Btw, how does your site rank on the following 2 DCs?
| 10:50 am on May 9, 2008 (gmt 0)|
Is that on both Google.com and .co.uk or are you being hit hardest on just on .co.uk?
I think the overlay of the UK filter on what is already a moving target is very difficult to fathom.
| 11:10 am on May 9, 2008 (gmt 0)|
On both of those DC's even when I type in a unique text string I get scraper sites pop up in fornt of my pages. I have to say that it did this 12 days ago as well but then everything returned back to normal, fingers crossed it does this time as well.
Hi HissingSid we have been hit hard on both .co.uk and .com but some words have held up on .com that have been lost on .co.uk.
| 11:44 am on May 9, 2008 (gmt 0)|
Thanks for feedback. But do your site rank the same for the same keyword(s) on both the DCs I mentioned? I expect different ranking :-)
| 11:59 am on May 9, 2008 (gmt 0)|
Actually Reseller I am noticing different results on those DC's.
| 12:04 pm on May 9, 2008 (gmt 0)|
Thanks. Whats the position # on the serps?
| 1:28 pm on May 9, 2008 (gmt 0)|
The positions are still really bad for instance we were position 6, now on the 188.8.131.52 we are 80th and on 184.108.40.206 we are 140th. Crazy.
| 2:05 pm on May 9, 2008 (gmt 0)|
Searching those two IP's from South Africa, I see identical results on all main search terms. There are some minor fluctuations on a few secondary and tertiary terms. I suspect though, that the results I see on those IP's are different to what someone in the UK or USA or Australia would see for the same terms on the same IP's. All part of google's grand geotargeting obsession, which is only serving to kill online businesses whose core clientele is often from other geographic regions/countries.
To test that hypothesis, perhaps a moderator could authorise a two word search term which could then be tested on the two IP's by users from different countries, with the results posted here?
[edited by: Optimus at 2:42 pm (utc) on May 9, 2008]
| This 224 message thread spans 8 pages: < < 224 ( 1  3 4 5 6 7 8 ) > > |