Forum Moderators: Robert Charlton & goodroi
Seems to be a lot of consensus that the shuffling is about links and link value. I am in a highly competitive industry and I definitely concur. I've spent the past three days doing in-depth backlink analysis on the competitor sites that jumped ahead of our site (pushing us to #11 from #6) and they exhibit obvious link building practices that Google supposedly frowns upon...mainly link purchases. I'm coming across a lot of run of sites. We have been steadily cleaning up our paid links, which seems to have been a mistake.
I've had this nagging thought in the back of my head that just gains more credence with each new Google update. I think Matt Cutts accomplishes through PR and dictum what Google is NOT able to accomplish algorithmically. You can only program a machine to do so much, evidence by the fact that we still don't have robot servants or cars that can drive themselves.
So how could the largest ad agency in the world (oops, I mean search engine) control the factors that they can't through algorithms? Why not create some sort of demi-god that respectable, white hat SEOs will flock to and follow without question? I think a lot of us have been duped and now the spammers and less-than-white hat SEOs are reaping the benefits.
Seeing as how many of us are seeing poor quality sites with poor quality backlinks beating out older, quality sites, is it too far-fetched to suggest that maybe Google had turned off a big portion of their algos that try to filter out paid links? Perhaps because after several months of launching a PR campaign against them, maybe they feel that enough sites have cleaned up those links? Or maybe because they only real filter they have the "Report Paid Links" database that they've been building?
PageRank isn't the biggest PR in SEO anymore, it's Press Relations and we all know what that's about...how to "spin" things.
[edited by: tedster at 6:37 pm (utc) on April 5, 2008]
Also the weight given to older links seems to be gone.
Things are still changing too much to analyse. We need to see how the results will settle in. But I still can't resist taking a peek at the data centers each day.
But I still can't resist taking a peek at the data centers each day.
Me too :-)
It looks like they ARE wiping off sites with a lot of back links from another site. However in our case it is completely whitehat and we still got banned.
We have one domain which we use exclusively for framed calls to external sites. i.e. all links to paid advertisers get a return link to our site added to the top in a frame. We use the go-between domain on a second server for hosting all these calls. We've been doing it for 15 years this way because of high server traffic demand. In the last week the PR on this go-between domain went to a white 0. We thought about naming the other server something like links.---.com to match our primary domain, along with all the inherent networking complexity of hosting subdomains on different physical servers, but now we're afraid to do that since it may incur the same penalty for our main domain.
Unfortunately, again because of server load issues, we also have large portions of our site spread across multiple domains and servers and highly interlinked by necessity (the busiest related pages are hosted on separate servers to help distribute the traffic). Guess those will be next to get whacked. Oh well, at this rate they ought to ban mapblast too, since we have almost as many links to them as well, bwahaha.
[edited by: lawman at 10:54 am (utc) on April 11, 2008]
Thanks!
[edited by: tedster at 2:45 am (utc) on April 12, 2008]
[edit reason] no specific keywords, please [/edit]
p/g
DEWEY: at one moment you're winning; at the next glance you're a footnote.
I have a bit of a problem that I can't seem to figure out. For the past two weeks Google has been shuffling my website from position number 4 to 14 for <a keyword>. In the morning it will be 14, in the afternoon it will be number 4, this changes about 3 times a day and like I said has been going on for about 2 weeks? Any reason why this would be happening or how I might try and establish some consistency with my placement?
I think Google is trying out different sites for qualifying in the top results. I was seeing the exact same thing as this going on for a few weeks and then all of a sudden I stayed there on the first page. The shuffling still continues with other sites. I now occupy either position #4 or #5 when previously my default position was #14. If google is doing it's shuffling I'm at #5. If not, then I'm at #4. The result of all this shuffling has had a good outcome for me. I've been in this new position for about 2 or 3 weeks now. This is for a one word primary keyword.
[edited by: tedster at 2:45 am (utc) on April 12, 2008]
[0089]...It may be discovered that there are periods when a document is more or less popular (i.e., has more or less traffic), such as during the summer months, on weekends, or during some other seasonal time period.The search engine may appropriately adjust its scoring of the document during and outside of these periods.
Google's History and Age Data Patent [webmasterworld.com]
That patent was from 2005, and I know such ranking shifts do happen for seasonal pages, or even for days of the week. With increased processing power, it's very conceivable that Google might test ranking shifts for different times of the day, too.
> I think Google is trying out different sites for qualifying in the top results.
But how does shuffling give it any data besides click-through-rate, or time on site, which are both very sketchy and unreliable?
It's either a software/hardware issue or combination of the two.
I'm not convinced yet there isn't a programming flaw. Google recently acknowledged a mistake wrt the positioning drop of -5. Its programmers at Adsense have made multiple programming mistakes in the last year that took them weeks to figure out.
I think there could even be a data storage/movement problem. Google has already admitted storage problems in the past. The amount of data it attempts to store isn't being reduced. If they're trying new ways to deal with it via data centers you could see some knarly results for weeks.
p/g
When I get bumped down in rank, the sites taking the top spots are ones I've never seen before. That, or there's more Wiki results.
Every time I think the results have settled down, a new site shows up in the results for a few hours or maybe a day.
thank you
If no, when will it settle down?
Erku - It's pointless to speculate on that, because it's not something we can predict. The most helpful observation I've seen with regard to your concern, though, is tedster's comment, back on April 2, in part one [webmasterworld.com] of this thread....
My emphasis added:
I think reseller has the right take here. This is more like an infrastructure change. As far as I can remember, we've never seen these rapid oscillations in ranking on the same IP address. I also can't believe that Google would settle for such instability over the long term.
That said, the infrastructure change looks like it's accompanied by new algo considerations, probably link related. The infrastructure change might in fact be enabling or facilitating these new algo considerations.
So, while the degree of instability is likely to dampen, the "everflux" effect (which is not new), of constant algo changes and refinements, is likely to remain with us.
It seems a lot of comments have been made here because some people waste time looking at datacenters for ranking differences. That's pointless and deceptive.
Whether there was a major or minor algo tweak, or new infrastructure that influenced existing data, what we have is a new basic set of results that has been stable for some time now.
If anything, the most notable thing here is the datacenters have become even more irrelevant. The "extra secret" sauce that is added to datacenter results before they hit Google.com is even more prone to effect the results than previously.
Whether there was a major or minor algo tweak, or new infrastructure that influenced existing data, what we have is a new basic set of results that has been stable for some time now.
Perhaps in your niche.
I know where I look that I am still seeing significant changes among the top 20. And I don't just mean shuffling of the same sites amongst those positions. I am still seeing entire sites appear and disappear.
I am happy to report that some of the heavily interlinked sites have been reigned in though.
If things are yet to be settled down there IS HOPE, because two of our sites that used to bring about 1000 visitors per hour each, not bring only 70 visitors per hour. I am talking about in average. Can you believe, from 1000 visitors per hour to drop to 70 visitors per hour.
But if the results have settled down already, then we have a reason to panic.
We are talking about very decent content sites of research nature. One is 4 years old updated daily, and the other 2 years old updated daily.
Perhaps in your niche.I know where I look that I am still seeing significant changes among the top 20.
Yes, it is very much niche-related. I'm seeing very different situations for different client niches, and for different searches within those niches. Some are completely stable. In some others, there are several sets of rotating results that are roughly predictable.
Some, though, appear to be wildly erratic, with rankings on pages 1 or 2 one day and page 7 or 8 another day. I haven't seen sites completely appear and disappear, but I have seen pages do so.
I am seeing that the results which I couldn't believe that Google would let stand are in fact changing... albeit I think they may still have some more changing to do.
The "extra secret" sauce that is added to datacenter results before they hit Google.com is even more prone to effect the results than previously.
I think there is a lot to what you are saying. But it still seems to me that data centers are an indicator of change to come. Also they can tell us where we might be without the secret sauce.
Too often people report that their traffic went up or down, but they don't know which keywords are involved, or which urls are showing the change in entry traffic. These are important pieces of information, and information that can generate immediate actions for us. But these days you can't even be sure that any given IP address actually points to the same data center from moment to moment.
So what actions can you take, even if you do notice that rankings have shifted a certain way at one IP address or another?
Data that is not actionable is not very useful, in my world. Time is too short. If you have the spare time and you enjoy the entertainment of watching the various IP addresses, then with the understanding in place that this is sort of a hobby, go for it!
In the past week alone I've been at #7, #8, #3 (!), and, today, nowhere. (For the top sector keyword; the rest are normal and stable). The craziness is only at google.com, not google.ca, google.co.uk, or google.com.au.
For the folks seeing erratic SERPs, is it for your entire site, or just certain keywords (e.g., most competitive)? Is it absolute or like the 950 penalty?
Google has been playing around with its top SERPs in the last months (-5 "penalty" for example), which Google didn't immediately recognize or correct.
It's easy to imagine this latest situation is more playing around with its algo for top SERPs for the most competitive search words. Which again Google hasn't quickly recognized and/or corrected.
I can see why people might think the shuffling is Google letting different sites get higher ranking to see how they do, but I don't see any real basis for this.
Google may be developing a revised algo for top SERPs in competitive areas based on link profiles.
To the point of stability, it really has no bearing on the users. It's really just us webmasters who are aware of shifting SERPs. And, presumably, the odd surfer who forgot to bookmark a site, and expects to find it in the same place in Google Search Results the next day.
p/g
To the point of stability, it really has no bearing on the users...presumably, the odd surfer who forgot to bookmark a site, and expects to find it in the same place in Google Search Results the next day.
This has, unfortunately, happened to me as a user a couple of times. The SERP's turned up a couple of real jewells in some research I was doing. My daughter got on my computer (NO! - She does know better, but the family computer was defragmenting/de-adwaring/de-etc. and unavailable) promptly shut down my browser and lost all my nicely arranged tabs (so she could check her myspace page :().
Went back to find those pages and they were nowhere to be found. And I mean nowhere in the top 500 or so.
I am still grumbling - at her and at Google...
I've seen relative stability in my sector with the top results still the same set of usual suspects. My site has moved up on some keywords, but then also moved down on others that were long established keywords which is more and more seeming like a - 5 or -6 penalty, again.
The strange stuff I'm seeing is just for one term, for other terms the SERPS are stable and OK in that they don't look strange.
More generally sites with on topic pages and one way outbound links seem to be doing particularly well after of the recent changes. It takes me back to Florida when I was reporting that sites that looked a bit like directories where doing particularly well.
I was recently reading "Winning Results with Google Adwords" by Andrew Goodman. He strongly suggests that the Florida update was all about forcing commercial web sites out of organic search and onto Adwords. I'm starting to wonder if what we are seeing is Florida II or another attempt by Google to push commercial sites out of organic search. This fits perfectly in my niche because the term that is being affected is the top $ Adwords term which on its own outperforms all of the other terms massively. For example in a period when I spent $4,000 on Adwords $3,500 was on this one 2 word term.
"Peter Norvig, a Vice President in charge of search quality ... implied the Google might attempt to make distinctions between what we might call 'informational commercial pages,' such as company histories, and 'solely commercial pages,' such as catalog pages. Left unsaid was their none the less clear message: commercially oriented pages are most suitable for the Adwords program. Those who want to reach customers should pay for targeted clicks and optimize their paid search campaigns for the best possible results." Winning Results with Google Adwords, Andrew Goodman pp 17 & 18.
There's an article on this stuff by Andrew Goodman here: [webpronews.com...]
It was written in 2003 but seems pertinent to today's SERPS.
I wonder if the recent changes are an implementation of an algorithm element that seeks to identify commercial "destination" sites by their linking patterns. But what happens isn't that non-commercial sites rise to the top for those terms rather what happens is a different kind of commercial site rises to the top.
In my own niche the current #1 is a site that has bought links all over the place with the keyword term in anchor text and which has outbound links to sites that pay affiliate fees to it. These look like clean, one way, outbound links to exactly on topic sites. As a result it has not only got the top slot but Google has given it site links.
Cheers
Sid
I am a bit concerned where it is going to stick, and why it is moving around so much. If anyone has any ideas to why this is happening and has been going on for weeks now i would be glad.
Luke
[edited by: tedster at 2:40 pm (utc) on April 14, 2008]
[edit reason] moved from another location [/edit]
Our number of pages in google have not dropped and i really cant see that we've lost much in the way of rankings although it does flux alot.
In the other engines particularly live and msn are gaining ground like never before with regards to sending visitors.
