| 8:38 pm on Sep 22, 2007 (gmt 0)|
Just an update about the spam sites with cn domains.
I just did a search using a common phrase that people used to find my site.
Out of 420,000 pages it looks like this:
#1 spam site
#2 spam site
#3 spam site
#4 relatively related site
#5 my site
#6 spam site
#7 spam site
#8 used to be site on #1-#3
#9 spam site
#10 spam site
and it's the same on all following pages...
Scary stuff that GOOG doesn't seem to be able to get rid of. It seems like these sites are (like I said in another post) too close to the kernel in their now outdated algo, so if they get rid of these sites, they get rid of authority sites also. This problem doesn't seem to be "fixable" with a little tweak to the algo anymore. Needs substantial infrastructure changes!
| 1:43 am on Sep 23, 2007 (gmt 0)|
|I haven't seen any thread about how great the search results are for the searches you do, but if you want to start one for people who like what they see then it would be great as a complimentary to people complaining. Serious! |
Actually, Google's results are pretty good for most of the searches that I do (and no, I don't just search on my own topics). They aren't perfect, but they're better than they were a few years ago when affiliate spam cluttered the SERPs.
In any case, my satisfaction or your dissatisfaction isn't the the topic of this thread, which is about "September 2007 Google SERP Changes." I'd have to say that I haven't noticed any big SERP changes this month in my own searches. I can think of improvements that I'd like to see, and the jury is still out on the value of Google Universal Search as far as I'm concerned, but if there are widespread major changes in Google's search results this month, I have yet to see them. If you've seen something different, that probably just goes to show that one man's prosciutto is another man's Spam.
| 1:52 am on Sep 23, 2007 (gmt 0)|
|but if there are widespread major changes in Google's search results this month, I have yet to see them |
No comment on what that means.
For those who are blind to the major changes.
(with a specific query)
(independent indicator of algo changes for over 2 years)
and oh yeah. this thread.
BTW - my sites are ranking just fine. But they always do.
And yet still, miraculously, I, everyone is this thread, and a non-human/computer SERPs program are able to notice various changes among many SERPs. Very odd.
[edited by: whitenight at 2:01 am (utc) on Sep. 23, 2007]
| 2:26 am on Sep 23, 2007 (gmt 0)|
Folks, let's keep the this thread away from "generic editorializing", whether positive or negative. As it says in the Google Forum Charter [webmasterworld.com]:
|- Off Topic - |
WebmasterWorld is, first and foremost, THE place for webmarketing professionals. Opinion and editorial comment, including google-is-great, google-is-bad rants may have its place somewhere on the web. But the noise level it creates here makes it difficult to filter through threads for information of real value.
Generic editorialising, whether pro or con, may be removed without notice. We're not here just to vent, we're here to help each other understand.
I will be cleaning out the general comments from this thread, so that those who are looking to understand Google's changes can pick out useful information more easily.
So let's focus here on intelligent observations and professional analysis, and not personal opinion about Google - thanks.
| 2:39 am on Sep 23, 2007 (gmt 0)|
I definitely see noticeable changes for single keyword searches, but the wild and crazy results that were popping in and out seem to have disappeared. Things have settled down to something not bizarrely different from what we had 3 weeks ago.
| 4:38 am on Sep 23, 2007 (gmt 0)|
Here is a look at a single word SERP that has undergone significant changes - some good, some questionable, some bad. No CN spam.
G Pgs Indexed 865
Y Links 1130
Y linkdomain 1490
MS Pgs Indexed 2272
Alexa Rank 1117133
Web Arch Age Dec 03 1998
Delicious index 6
Comments - Good solid authority site on the subject. Hired gun SEO. Fiar amount of "gray" play regarding links.
Y Links 10
Delicious index 0
Comments - See #1. This is really what makes the site an authority. A worthy indent, unless one would like to see #1 gone and replaced with this.
G Pgs Indexed 3310000
Y Links 399
Y linkdomain 45500000
MS pgs Indexed 2023449
Alexa rank 9
Web Arch Aged Nov 27 2002
Delicious links 5
Comments - How can you compete with Wiki?
G Pgs Indexed 29
Y links 302
Y linkdomain 411
MS Pgs Indexed 180
Alexa Rank 1404764
Web Arch Age Dec 15 2001
Delicious links 12
Comments - A lead generation site. Poor to non-existent content. Splash home page. 3 pages of actual content. The rest is a psuedo-directory of sorts. Benefits from paid links and blogs.
Delicious links 0
Comments - see #4
G Pgs Indexed 168
Y links 10900
Y linkdomain 12000
MS Pgs Indexed 531
Alexa Rank n/a
Web Arch Age Apr 02 2002
Delicious Links 0
Comments -? This one came from nowhere. Majority of links are run of site c@#p directories owned by the site owner. It is listed in the "directory" at #4.
G Pgs Indexed 753
Y links 2120
Y linkdomain 2440
MS pgs Indexed 4971
Alexa Rank 1187953
Web Arch Age Sep 18 2000
Delicious Links 14
Comments - US industry association's site, so should rank as an authority. Benefits from many .edu and .gov links.
G Pgs Indexed 232000
Y links 618
Y linkdomain 10600
MS Pgs Indexed 121410
Alexa Rank 296338
Web Arch Age Sep 26 2000
Delicious Links 7
Comments - Closest thing to web 2.0 in this niche. Thousands of pages of user generated content. Kind of a cross between Craigs List and You Tube, and many of the content contributors link in from their own content. Authority -? It is certainly a major "gathering place" for the niche.
Y Links 133
Delicious Links 0
Comments - See #8
G Pgs Indexed 448
Y links 204
Y linkdomain 241
MS Pgs Indexed 437
Alexa Rank 2594120
Web Arch Age Oct 22 1996
Delicious Links 0
Comments - This is an odd one. Owner purchased "widget.com" several years ago (for an mazingly low price) which mirrors all the content. G can't decide to list one or the other or both. In fact a few days ago both appeared in the top 10. Another psuedo-directory, a few pages of meaningful content and an e-store. FWIW, widget.com has 3 X's the links, many older legacy edu and gove links from prior to the transfer of ownership.
G Pgs Indexed 1660000
Y Links 10
Y linkdomain 2220000
MS Pgs Indexed 68453
Alexa Rank 341
Web Arch Age Dec 31 1996
Delicious Links 0
Comments - IMHO Answers needs to be shot and buried. If any mere mortal did what Answers.com does we would be branded as scrapers and/or affiliate marketeers, and would be forever banned from the index.
G Pgs Indexed 48
Y Links 26
Y linkdomain 49
MS Pgs Indexed 201
Alexa Rank 3623112
Web Arch Age Mar 23 2004
Delicious Links 0
Comments - This one would rank well in the worst of the web, as far as layout is concerned. G cache gives a page size of 44k. My DSL line is still choking after 4 minutes of downloading. Cloaking? If someone can figure out why this site is top 20 I sure would like to know it.
G Pgs Indexed 1490
Y Links 89
Y linkdomain 49900
MS Pgs Indexed 1278
Alexa Rank 15954
Web Arch Age Oct 18 1996
Delicious Links 2
Comments - One of the original articles about these widgets on the web. Terribly dated after 11 years. Was top 10 a couple of days ago. Haven't seen it in the SERPS since Florida.
G Pgs Indexed 388
Y Links 90
Y linkdomain 115
MS Pgs Indexed 160
Alexa Rank 4056957
Web Arch Age Aug 31 2000
Delicious Links 0
Comments - Claims to be the largest widget outlet in the US. A few pages of content. Many pages of product description (mainly marketing hype). This is another site that has arisen from nowhere and was top 10 within the past 5 days.
G Pgs Indexed 17
Y Links 119
Y linkdomain 148
MS Pgs Indexed 85
Alexa Rank 6352660
Web Arch Age Jan 10 1997
Delicious Links 0
Comments - Whoa! Where did this one come from? Forum links and DMOZ clone links account for most of the 148 in DMOZ. The domain has changed hands 3 times since 1997. Very different focus each time. Suspect legacy links that don't show right now.
G Pgs Indexed 27
Y Links 2
Y linkdomain 4
MS Pgs Indexed 70
Alexa Rank n/a
Web Arch Age Mar 30 2001
Delicious Links 0
Comments - This is a mirror of #15. All dup content (see comments under #10 for another example of G struggling with mirrored sites/dup content with this update)
G Pgs Indexed 97
Y links 104
Y linkdomain 136
Ms Pgs Indexed 522
Alexa Rank n/a
Web Arch Age Feb 16 2003
Delicious Links 0
Comments - nice site. Easy to navigate. Many pages of content, though some is a bit skimpy in terms of value.
G Pgs Indexed 12
Y Links 37
Y linkdomain 39
MS Pgs Indexed 11
Alexa Rank n/a
Web Arch Aged Mar 25 2004
Delicious Links 0
Comments -? Perhaps it is because the footer of each page is stuffed with a list of keywords?
G Pgs Indexed 126
Y Links 625
Y linkdomain 716
Ms Pgs Indexed 688
Alexa Rank 3092291
Web Arch Age May 16 2001
Delicisous Links 7
Comments - This site ranked top 10 for 5 years. Sometimes as high as #2. It has perhaps gotten too diffuse in its focus? Should still be considered an authority in that it is more or less a DIY guide to widgets, quite unlike any other in the top 20.
The only external events that I am aware of that would have a significant impact on this niche is that a new DMOZ editor saw fit to rearrange the entire category about a year ago. Many, many authoritative sites were buried in subcats, as if all sites are created equal and the only thing that mattered was region. I suspect that the G Directory update recently has had an impact on these SERPS.
Only one hyphenated URL appears in the top 20, and it has fallen 18 places.
#6 suggests that ROS links may count for something, at least a the moment. The shear number of links seems to be the primary reason for the ranking.
Beyond that, it almost looks like there is a random element introduced into this algo. 13-20 are filled with sites that could have been plucked from a hat.
<edit> fixed formatting</edit>
| 6:44 am on Sep 23, 2007 (gmt 0)|
So you don't think on page, semantic webs, anchor text, KW density and KW prominence are important then.
| 8:51 am on Sep 23, 2007 (gmt 0)|
I think I may have an idea why some sites go down. I was surfing Google for my own sites when I found a warning attached to one of the pages..just said beware. When I clicked on beware it popped up with: This site may put spyware or a virus on your machine. (paraphrased)
I was absolutely flabbergasted! The only thing that ever did that was Their very on adsense ad, not Two months ago, that was doing it. I reported the ad and quick put it in my filter. I found the ad by using the preview tool. And that during their quality shouts.
I was so shocked that I lost the URL to the serps page so don't know how to prove I saw it and where.
I have rocked down quite fast to number 53 from number 6.
I have always tried to keep a check on my ads from other companies also, a little freaky behaviour and zap it is gone.
| 9:04 am on Sep 23, 2007 (gmt 0)|
Well people love to say on forums:
Don't automate (link building)= low quality.
Don't automate (content generation)= low quality.
Don't automate (ad serving)= low quality.
| 12:31 pm on Sep 23, 2007 (gmt 0)|
Is it possible Google tests its new algo/changes on one or a few industries before letting it loose on everything? How else to explain why some folks see huge changes while others see nothing?
| 3:19 pm on Sep 23, 2007 (gmt 0)|
|So you don't think on page, semantic webs, anchor text, KW density and KW prominence are important then. |
I don't see much of a correlation when I look at allinanchor.
Density is all over the place. Ranges from 0% - 4.42% on the plural term widgets and from 1.98% - 9.52% on the singular term widget.
8 of the top 20 do not even have the plural term widgets (which was the search term) in their title tag, using the the singular widget instead.
Very little evidence of proper use of <hx> tags or "proper" document structure.
I do see a lot of stuffing in alt tags on a couple of the sites, and one with the old black of keywords at the bottom of each page.
So no, I don't see that on-page factors are playing a huge role in this update. Semantics? Certainly, given that there are sites in the top 10 that do not contain the exact search term (widgets) but do have the singular form of the term (widget).
| 3:24 pm on Sep 23, 2007 (gmt 0)|
|Is it possible Google tests its new algo/changes on one or a few industries before letting it loose on everything? How else to explain why some folks see huge changes while others see nothing? |
I have suspected something like this for a long time (2 years?).
Changes often appear to propogate through the serps by sector, and it seems to take anywhere from a week to a month.
It would be interesting to be able to map changes by sector and compare against something like DMOZ to see if there was a correlation.
I suspect that the recent G Directory update has played a far loarger role in the current changes than most are giving credit for.
| 6:34 pm on Sep 23, 2007 (gmt 0)|
I strongly suspect that they test on some industries before others. I generally see changes in my niche 1-2 weeks before it is reported here. The changes I see are never quite as dramatic as they are for others though.
| 7:23 pm on Sep 23, 2007 (gmt 0)|
I have been playing around a little bit to get a search that would remove the spam sites and the only thing I could come up with is the usual "-" :(
Anyway, these are some searches I did:
widgets in Ontario 246,000
widgets in Ontario/ -.cn/*.html 226,000
widgets in Ontario/ -.cn 126,000
widgets for sale in michigan -.cn/ -.cn/*.html 999,000
widgets for sale in michigan 1,200,000
Still there and still going strong. A lot of them occupy top positions :(
| 1:09 am on Sep 24, 2007 (gmt 0)|
One of my visitors to my site came from google.com using this search:
widget manufacturers in houston texas
This search in Google shows 967 results.
I did the same search in Google.co.uk using the Web search and got 14,500 results.... If you use Show only result from UK you get 14....
How come they are so different? Shouldn't they be almost the same in Google as in the Web Search in Google UK?
Google.se shows 14,400 for the same search
At all other Google country domains they ask "Did you mean midget manufacturers in houston texas" but not in Google.com
| 1:51 am on Sep 24, 2007 (gmt 0)|
Have you ever seen Antonioni's movie Blow Up? The photographer thinks he might have taken a photo with a crime off in the distance - but the more he scrutinizes the detail and blows up things, the more it becomes a collection of grainy dots. Kindly, this is what I think you're doing here: looking at Google so closely that you're getting lost in all the grainy details.
With hundreds of thousands of servers and data migrating all over the place all the time, there are often reporting features that mean very little, except to say "that's just the way it is right now." My guess is that even when the results say 14,000 pages, you could not click all the way to page 99.
Those page numbers have never been very good. Because of the way that Google shards their data, the figures are often difficult to calculate, and different data servers can show quite a variation. If those anomalies persist for several months, that *might* mean something you can make use of.
| 2:14 am on Sep 24, 2007 (gmt 0)|
You might be totally right here but I think, at least for the search I did, that there is something else going on.
The sample I gave about the widgets, is not really good at all but I couldn't come up with something else that would fit, and I know that the one I choose was not good enough :(
I think that in the search with only 900 or so pages is a result of something else, and that is that they have categorized that search in the .com but not yet worldwide. That is why they only have a small result set here and so much more overseas (spelling?) and have to ask if that was what you really searched for....
It seems that the results are better if you don't use google.com, not so much spam sites and sites are almost where they used to be....
| 7:52 am on Sep 24, 2007 (gmt 0)|
For my problem term allinanchor:problem term returns exactly the same 1 - 10 on Google.co.uk and on .com the places move just one slot up or down.
Previously I was #1 on both allinanchor and standard searches now I'm at #5. This change is only for that one term. For a 3 word search that includes the problem term I'm still #1 on both allinanchor and standard search.
The two word term is more semantically ambiguous than the 3 word term and other unaffected 2 word terms.
So is it something with allinanchor or semantics or the way the semantic algo is applied to anchor text?
Or none of the above?
| 1:06 pm on Sep 24, 2007 (gmt 0)|
I really don't understand what they are trying to accomplish at this point.
Not all is irrelevant of course, but there are countless searches returning the worst that that internet has to offer.
My SERPS often look like this in the top 20 of Google.com:
1. 3 or 4 true authorities, or niche authorities to which the relevancy factors have been cranked up to the maximum. This guarantees a certain % of relevancy I assume.
2. Up to 9-10 sites that are in fact based on blogs or forums. That's up to 45% of annoyance with self promoting and paraphrasing non-sense that does not help any consumer at all. People faking comments, talking to themselves often in order to end up sellin their own stuff. Just annoying, looks like the blog issues search engines all had in 2005.
3. Foreign language sites from languages I barely recognize, these have only pluged in usually 1 occurence in a page or their domain name - who cares I can't freaking understand what they are talking about!
I am not talking only JP or CN, but also German, Spanish and others that look like martian to me.
4. LOW, very LOW and cheap quality sites. A competition of total crappiness that continuously pops up from nowhere! The crary part is that sometimes they seem to stick.
I've analyzed these sites: cr.ppy free directory or directory network backlinks, flat and often grammatically content (also often to the limit of copiright infrigement).
NOTE: those sites are often blogs, approx 50%.
5. Up to 5 links out of 20 from the same company. Including subdomains of course
6. Empty pages or almost empty or sites that stopped doing business for about 1 year +
7 Companies that have installed a blog on another domain: both sites rank almost equaly in top 20.
8. I usually find 1 site "This site may harm your computer" anywhere from pos. 1 to 20. Then if it may harm my computer just remove it, there are so many cr.ps around at this time.
- Enough now with widget.com and variations, is this 2007 or 1999? Usually anything that's #*$!widget.com is suspicious, great branding hey! -
- Subdomains widget.example.com are back on top, I thought we were done with this for a long time. But when example.com is powerful it seems to tranfer some juice to subdomains.
- How can a set of sites with 30-50 backlinks from unrelated and obviously shady on site SEO can replace sites in business and relevant for years and including backlinks only from relevant sources with no SEO overboard? I am no Google but at my level of understanding this doesn't make sense, even MSN gets more relevant than that.
One of the big difference between these sites and the previous Relevant sites is that new sites have not suffer 3-3 years of scrapping.
- How come sites I believe have been banned or filtered for overoptimizing/buying bulk quantities of textlinks get now a clean slate while they practices have usually not changed at all?
- Why so many blogs or forums? Is this all a freash content problem or indexing issue? Why so many old static pages are not even indexed anymore? And why their over optimization and super low inbound and outbound link profile gets unfiltered?
- Why a site that STATES clearly on their PAGE TITLE that they sell widgets in Ohio becomes relevant to "national" type of search term (I'm in CA). Why tiny local players are relevant unless their have their office accross my street?
- Please shoot about.com and answers.com straight in the head.
I (no kidding) currently already paid the price of this thing: ordered a product for my laptop from a site. Company is in India and won't deliver. If I've searched my stuff on co.uk I would not have ordered from them.... That's just my mistake, it was only $12 so I don't care but it shows how much consumers can be mislead in some industries.
With my little experience I have never ever seen such cr.p SERP's since the Florida update, n-e-v-e-r.
I know the industries we are involved very well, let's say I am an expert in research, and I can tell you that except authorities all regular online+offline players got their position hijacked.
It's not the case on all industries we are involved in. Only thos where there is a lot of google noise with blogs and forums it seems but I don't have a definitive answer.
[edited by: tedster at 7:46 am (utc) on Sep. 25, 2007]
[edit reason] member's request [/edit]
| 4:33 pm on Sep 25, 2007 (gmt 0)|
Many of the SERPs I have been watching are being peppered with websites that have mass amounts of low PR paid links. Totally irrelevant footer and sidebar links in mass amounts seem to be very effective at the moment.
I thought Google would be better at spotting this type of blatant link gaming, but for the moment, garbage links seem to be working better than ever.
Maybe Google is trying to flush out these sites by getting link spam reports from the site's they are knocking down.
I have nothing against a nice relevant paid link placed here and there, but to see a site dominate with tens of thousands of worthless links drives me up a wall. I guess these links arent so worthless anymore.
On one hand, I want to fight back with the same strategy; on the other hand, I just have to believe Google will figure out how to spot websites that have thousands upon thousands of irrelevant links in footer and sidebar blocks of paid links.
| 10:36 pm on Sep 25, 2007 (gmt 0)|
The code is:
|According to |
212B34 FF0E EFBC8E D Ideographic variant full stop
This does appear to be a google attack and not a china attack. Notice the difference between a site search for one of these sites:
| 11:20 pm on Sep 25, 2007 (gmt 0)|
Thanks for that. Normally we do not allow specific Google searches to be published in posts here, but we will make an "educational exception" in this case.
I can certainly see why this kind of domain variant might have Google scrambling to fix some deeply embedded routines throughout their system.
| 11:40 pm on Sep 25, 2007 (gmt 0)|
It does seem to be certain commercial terms only that are being played with. Amazing the irelevent result pages are showing very relevent sponsored listings, talk about double standards.
Interestingly, I have been out of the google loop for a while, some sites I de-seo'd (ironic statement) for post florida and then left alone have made it back up the listings.
| 4:53 am on Sep 26, 2007 (gmt 0)|
|Many of the SERPs I have been watching are being peppered with websites that have mass amounts of low PR paid links. Totally irrelevant footer and sidebar links in mass amounts seem to be very effective at the moment. |
Yes ever since 09/13 from what I can tell, and even worst than that I have countless examples of sites with 'only' a few dozen 'horribly cheap' backlinks that now compete on top with authorities, while regular industry leaders are pushed back to page 2/3/4+.
Very few backlinks of extremely low quality suffice to compete and stabilize on top of some industries.
|Maybe Google is trying to flush out these sites by getting link spam reports from the site's they are knocking down. |
I don't think so. What are we going to say? This site sucks, it's irelevant? The quality of sites that have popped up out of nowhere is the issue but they don't actually all spam (some do of course).
You can't even report these results as a knowledgeable webmaster, too much to report on just quality. Can't even report foreign language stuff, it's not spam - it's just Google's weird irrelevancy at the moment.
I don't think that end users would report anything at all. They are just going to trust what Google shows - too bad for them, good luck with the Better Business Bureau and their bank for all the money that will go to scammers here and especially abroad.
|Interestingly, I have been out of the google loop for a while, some sites I de-seo'd (ironic statement) for post florida and then left alone have made it back up the listings. |
Same thing I've noticed. Along with the cheap sites are back all those filtered for overoptimizing....whether they still over optimize or not.
The whole thing also pulls up tiny local players from all over the world. Logically if some site Specifies their location on the Homepage title (ei. city, country, state) then unless the user specifies (or ip location shows) that he want such local results LOGIC would dictate that these are not shown. But that's just logic and logic has nothing to do with some serps Google shows at this time.
Such a plague.
| 10:45 am on Sep 26, 2007 (gmt 0)|
Is Google broken? I just got several 950'ed sites back. It won't last but it's nice to see anyway.
| 4:53 pm on Sep 26, 2007 (gmt 0)|
I just did a search in a friend's industry where he has been top 3 for the main keyword for 3 or 4 years. I notice he is down to number 5 for his main keyword phrase. So I checked the backlinks of the new number 1 website in this industry.
surprise: they have thousands and thousands of Chinese backlinks - I could not spot one real link for this site.
As far as I can tell, the backlinks are all Chinese guestbook and blog spam. Perhaps the nofollow is not being used as much in China, Russia etc... and the loophole is being exploited?
| 2:49 am on Sep 27, 2007 (gmt 0)|
|...the backlinks are all Chinese guestbook and blog spam. |
The question of nationality came up in another thread as well... are these in fact Chinese sites, or are they simply .cn domains?
| 2:51 am on Sep 27, 2007 (gmt 0)|
seems to me there seems to be a word penalty on certain commercial terms. words being anything sale related - buy, buy online, sale, for sale etc etc.
maybe even sites with links to paypal and payment proccesors are copping it.
| 3:06 am on Sep 27, 2007 (gmt 0)|
a search term, form example, of "buy widgets" is only bringing up pages that link to purchasing sites and have page with description such as blogs and forums and not the actual purchasing sites themselves.
| 3:41 am on Sep 27, 2007 (gmt 0)|
|are these in fact Chinese sites, or are they simply .cn domains? |
A few are .cn english machine generated websites but most are pretty much pure Chinese text - except for the links.
I am also seeing a huge Russian network of sites boosting some other very new dot COMs into the top 10. At the moment, it's a pure numbers game, get 50,000+ links from anywhere, it will put your site over the top. It has been a long time since I have seen blatant, irrelavant, spammy paid links work so well.
I am thinking this is temporary, if not, there goes my whole link development strategy.
"Keep your links relevant, paid links are evil" - ha
| 12:02 am on Sep 27, 2007 (gmt 0)|
I've been noticing that the shopping engines such as bizrate, nextag, pricegrabber, and shopping are taking up a substantial portion of page 1 search results. I don't remember the situation being as drastic as this. Historically the engines were not ranked this high and they were heavy users of sponsored links. They rely a lot less on sponsored links these days. What do you make out of the shopping engines taking over page 1 of product results?
[edited by: tedster at 3:46 am (utc) on Sep. 27, 2007]
[edit reason] moved from another location [/edit]
| This 142 message thread spans 5 pages: < < 142 ( 1 2  4 5 ) > > |