Forum Moderators: Robert Charlton & goodroi
AlexK, that's a different domain. But the point is very well taken. You've found a pretty obscure query (~295 results) that the keyword stuffing spammers like to target. I'll check this out in more detail.
[edited by: Brett_Tabke at 1:03 am (utc) on Nov. 12, 2005]
Was 13 now down down and down....
No clue why.
I do not recip links all 1 ways..
GG and MC seem to be right there when things go well. But when they don't, it's Google policy to be quiet. I understand that. It's just that getting only half the news is more misleading than getting no news. So thanks for your help guys, I just wish it would be more consistent.
I think you don't get it.
When things "go well" it means SEO's are happy. Therefore the anti-SEO team comes over.
When everyone here is mad and whining it means their thing worked.
GG and MC have the mission to destroy you and you are helping them out.
<state> widget forum directory - used to to be #1 now #30
<state> widget forum - used to be #1 now #4
<state> widget - used to be #35 now #8
widget - not found now #233
Also it seems internal pages are showing the same type of change with the exception that almost all of them display the homepage first with an indent to the particular page. Before, the indented page was listed by itself - not indented. (Is this making sense?)
So as I see it, I am getting better ranking for fewer words in the phrase but slightly worse ranking for more words in the phrase. Might just be the new algo, might just be where the chips fell, I don't know but I find it interesting.
By the way, referrals from msn and Y have gone up dramitically in the past 3 weeks.
[edited by: webdude at 5:34 pm (utc) on Nov. 8, 2005]
So thanks for your help guys, I just wish it would be more consistent.
Let me put myself in Google's shoes for a second here:
Why, in heaven's name, would I help you?
Don't you think google has you right where they want you?
I run over 50 sites and only one has been adversely effected. All others are up. My attitude is to just delve in and figure out the prob with that one site and get it corrected. G builds algos for there purpose, not ours. It is our job to react accordingly. However, that reaction can't be addressed until we are sure the new results are in fact stable.
Therefore....if this update was complete....I would think GG would say so, so we can get on with what we need to do.
G builds algos for there purpose, not ours.
Exactly.
It also means they will not help you crack their algo.
Every single word that changes on any of the data centers gets a comment here. With that kind of free feedback I'd build a Google myself....
I'm just waiting this out, and not reading this thread much anymore. There's nothing to do but wait.
Perhaps they build indexes for short phrases first and after that those for more complicated patterns?
Ankhenaton - I don't know if that is a good idea or not, I think google eventually does see the 301 and fix things - but what do I know google doesn't like my sites either!
I am just keeping busy, lol. Probably pointless. I did not really see why this would be necessary in the first place since ServerAlias is kinda Apache standard.
How hard is it to detect this?
If content www.example.com = content example.com likelihood of same site = 99.999999999999999
If google can't above this it's really a sorry state of affairs.
[edited by: Ankhenaton at 5:57 pm (utc) on Nov. 8, 2005]
I belive J3 is off from public view.
I wonder why 'GoogleGuy' didn't respond to my post. Interesting.For everyone who has asked though, here is the broad outline of the methodology we used to determine key site inclusion quality.
-----------------------------------------
1) Each person selected up to 10 topics each to study2) In groups of 3 people (quality check point), a search term was identified and agreed for each topic
3) For each topic, Google, MSN and Yahoo were searched using the identified search term. The top 25 results for each were recorded. For Google we used the most up to date 'Jagger' DC, partly because I felt it would add value to this forum.
4) All the sites identified for each topic were then aggregated into a single list. Each list, covering a single topic, typically had 30 to 50 sites included
5) The lists were then exchanged with different groups of 3 people (quality check point). Each group were then charged with viewing each site and selecting the 5 most important/useful sites from each topic list on that particular topic.
6) At this point, we had the 5 key sources in each topic area, with no search engine bias attached in the selection process. We had reduced the impact of subjectivity by passing the topics from group to group. The next phase was therefore to establish the degree of key site exclusion in each search engine.
7) The topics were once again placed with different groups of 3 people. Each group now searched each search engine on the search term(s) for each topic.
8) If any of the 5 key sites were missing from the top 20 returns for a search engine, that search engine was marked 'red' for that search term.
9) This exercise was performed for all topics and search terms.
10) The number of red marks for each search engine was calculated.
------------------------------------------This process gives a relatively objective measure of quality site exclusion (and thus inclusion). The search engine with most red marks clearly has most quality sites buried or missing. As I stated earlier, this should be important to search engines because a researchers search experience and success is determined as much by what they DO find as what they don't see ('spam').
On this measure MSN was the clear winner, with very few quality sites excluded from visibility. Google was by far the worst performer.
As a research entity ourselves this is clearly important to us. If important data sources are filtered out of the returns haphazardly, it makes the engine almost a liability to use in some instances. It does also confirm our 'feelings' over the last few weeks. We have in fact changed our default search from Google as a result of this. Whether we change it back will be determined by a re-run of the above sometime in the future. It hardly looks likely though.
Finally, browsing the current results in the Jagger 3 DC, for me, there seems to be a clear problem with identification of quality DATA, as opposed to trustworthy sites. Just because a particular site is deemed to be important, doesn't mean that a single page that simply references a particular topic is of any value whatsoever. In some of the niche topic areas I explore, the DC returns are full of them. It seems that all Amazon, IBM, HP or any large internet centric institution needs to do is create a page on a topic and they rank highly. This is generally a very poor effort.
Regardless of my opinion, I hope that the above exercise is of interest.
Seriously. The gravity of this issue is immense. Demonstrably, scientifically, 'Jagger3' had a substantial number of important data sources missing, and I am not talking about commercial selling sites.
The problem is that J2 is also poor in comparison with competitor search facilities. The trend has been going on for some time: apparent focus on what shouldn't be in the returns, at the expense of consideration and control of what SHOULD be there.
To be open, the whole scenario seems to smack of very poor quality control. Rather than creating a robust scalable sampling regime, using sampling theory, they seem to be testing via haphazard input from subjective webmasters. I was astounded to discover this when I came here. It doesn't square with a multiple billion dollar corporation, or should I say, one which has a robust control over r&d.
Maybe they will roll back even further, and actually address both parts of the equation (inclusion and well as exclusion). I wouldn't be betting any money on it though, or if I was one of you guys, my livelihood.
He said "STARTING yesterday"(last Friday), not "temporarily visible at". 66.102.9.104 is the only DC that I can see that still has a very minor flux still going on at. Out of 20 keywords/phrases I checked across DCs, only 66.102.9.104 showed any kind of change, and that was, out of 20 searches... on ONE of them out of 20, I saw one site appearing in the top 10 that didn't appear in the other DCs. One site, out of 20 searches and 200 top 10 listings. By the way, the site was a commercial site among mostly informational sites.
I'm taking screen shots of everything, that way, if something does change, I'll have something exact to compare to and can maybe figure out why the change occured... that is, if more changes(besides very minor) do occur.
Agree partly.
Regarding livelihood. The volatility of funding institutions could be possibly your next quality research topic. Would they let you do that? ;)
In the words of my colleague on his recent research proposal was: "I just had to put in these buzzwords". Also the nepotism and cronyism there is as bad as Google's search result, imo. Well at least in science, as probably everywhere else, democratic principles of power sharing are failling, imo.
People having to ponder to the funding institutions every turn and twists and people hanging on Googles wierdo twists and turns are in an equally volantile situation.
Dr. Anhks word of the week
Which seems to suggest changed values on anchor text.
Yes, I would agree with you there, JJ ... many of my sites are not returning as strong results for variants on anchor text. I noticed that last night.
I have been away from the Jagger watch for a couple of weeks, courtesty of Hurricane Wilma, and I must say I am grateful I didn't have to suffer through Jagger 2 and 3 ... and whatever it is I was seening last night and this morning is a clear improvment over J1 results ... even better, one of my sites is high in the SERPs for the first time ever ... and I never could understand why it wasn't before ... as all other sites I have in the same niche were steadily rising in SERPs ...
I do appreciate all your posts ... but I am especially interested in the sort of analysis that seeks to determine whether the algos have changed for particular elements, such as anchor text.
Thanks to all.
Cheers, MJ
Google's problem as I see it, after reading trust rank, is that they focus too much on likelyhoods and stats:
MIT links to you site, you must good pages, whether you have stuff about engineering or Teletubies.
You got a ROS link, so you must've cheated so we'll send you to #300, instead of just ignoring the suscpicious links unless they can be checked.
Too broad of a brush, and very easy to fall way off the radar IMO. We're not talking about from #2 to #5, we're talking going to 10 visitors a day.
He said "STARTING yesterday"(last Friday), not "temporarily visible at". 66.102.9.104 is the only DC that I can see that still has a very minor flux still going on at. Out of 20 keywords/phrases I checked across DCs, only 66.102.9.104 showed any kind of change, and that was, out of 20 searches... on ONE of them out of 20, I saw one site appearing in the top 10 that didn't appear in the other DCs. One site, out of 20 searches and 200 top 10 listings. By the way, the site was a commercial site among mostly informational sites.
He did indeed, and at the time results on 9 looked radically different to 7 (to me anyway).
Then all of a sudden they vanished into thin air and were replaced with very similar 7 data. Since then they've been fluxing.
This is the 2nd time 9 results have vanished now and the first time they were re-built after what looked like an error? (quote: results that crazy people had made).
I suspect a similar event is taking place again as we speak, however I could be completely wrong.
Either way... it's good news here.
I wouldn't be betting any money on it though, or if I was one of you guys, my livelihood.
Google is an economic jugger naught the likes of which this country has not seen in a long time. I for one am enthusiastic about trying to be attached to that in any way I can. There’s never been such a clear and equal opportunity for economically, business minded people to make money, with no large investment; just their wits and some internet savvy. Try building some sites and get in the game; if you do you will never look at the fundamentals of economic opportunity the same way again, quality search results or not.
Then it returns all the pages. This has been happening off and on for the past 3 days. That is what it is showing right now. Definitely something going on on that DC.
I have just visited the DCs. And as many of you have noticed, there is a very strong flux at the moment.
And here is some food for the thought:
I don´t think that one can point out at the moment for sure any specific DC/Dcs which reflect Jagger3 updated serps. Not even 66.102.9.104 or 66.102.7.104
[edited by: reseller at 6:44 pm (utc) on Nov. 8, 2005]
Just another note. On this 3 word phrase when I search 66.102.9.104, it shows a return of 2,560,000 pages, but only 14 results are shown.
webdude - I saw this too. This DC only showed a fraction (less than 100 results) for a popular query (5 million results plus) I was looking at. This was also repeatable. Not sure what day I saw this - might have been Sunday.
