Google Updates and SERP Changes - February 2011 part 2

[boirun03]

< continued from [webmasterworld.com...] >

But, I believe that Google now prefers sites with a much higher percentage of "valuable" pages. If you don't meet the percentage value determination, whatever that is, you get whacked.

I tend to agree with Fred. This is what I am finding as well.

[edited by: tedster at 8:00 pm (utc) on Feb 26, 2011]

[econman]

more hard data on the sites that lost out in the Farm update

Great find!

TripAdvisor wasn't hit (or, at least not badly enough to be listed), yet their sister site Travelpod dropped 91% by one measure and 68% by another measure (from 41,774 keywords to 13,261 keywords).

Not sure how the writing quality compares, but they are both based upon user generated content.

[maxmoritz]

I do not see this as having anything to do with writing quality, although of course I could be wrong. The sites that dropped in my niche all had the same quality/level of writing as the sites that remained. It may be a combination of things, but it's definitely not as simple as some judge of writing quality.

It looks to me like something was added to the algo that devalued sites rather than a changed the ranking order. None of the SERPs I look at changed with sites moving up and down and all around. Sites got devalued. Other sites moved up as a result of them moving down. Does anyone else have this view?

[TheMadScientist]

The sites that dropped in my niche all had the same quality/level of writing as the sites that remained.

Good golly I didn't read that right... lol

[edited by: TheMadScientist at 6:08 pm (utc) on Feb 28, 2011]

[Whoa]

Wow. So many excellent sites that I often rely on. That's crazy.

I guess beauty is in the eye of the (algorithm) beholder.

In my space, resources for entrepreneurs and small biz owners, there are some really good sites on this list that have been decimated.

I know StackOverflow complained about gravity gone wrong when they ranked lower than scrapers, and that motivated Google's cleanup, but this is just wrong -- so many good sites, not scrapers at all, just obliterated from the search results. Crazy. We can only hope that reason returns. There must be a way to eliminate the bad guys without killing so many good guys in the process.

[dickbaker]

It looks to me like something was added to the algo that devalued sites rather than a changed the ranking order. None of the SERPs I look at changed with sites moving up and down and all around. Sites got devalued. Other sites moved up as a result of them moving down. Does anyone else have this view?

No, I think it's a combination of devaluing sites to some extent, and then devaluing (changing rankings for) pages in particular.

I've looked at my site as well as others for which I have stats. Every page on an affected site moved down, but some moved down a little, and some moved down a lot. For example, moving from #4 to #12 while another page moved from #5 to #49.

There are pages that dropped eight spots that seem to be as strong as the pages that dropped 44 spots. There's other factors that play into the difference in the levels of drop, though. I'm seeing some of those factors, but I need to dig deeper to find more.

[TheMadScientist]

Does anyone else have this view?

Yeah, now that I read what you're saying correctly ... Could be a 'near duplication and origination of information' type of assessment ... EG There are already 10 of these we have which seem to be published first, so we'll discount the rest that have a high degree of similarity.

[chrisv1963]

Wow. So many excellent sites that I often rely on. That's crazy.

It almost looks like a Google engineer messed up the database or something.

[Industry]

tedster...thank you for the link to the 300...exactly what I was looking for...saw the Top 25 but this is perfect....how close do you think this data reflects real life....I read how they got to the numbers....but you know what they same about numbers....

[tedster]

My first guess here (and it is a guess) is that the Farm Update leans too heavily on the Scraper Update of a few weeks ago. And the Scraper Update does not yet do a very good job at identifying the original source of content, especially for extensively republished content.

[zoltan]

Many of these site are UGC, for example in B2B google thinks that if we have Alibaba (not affected), we do not need other portals / marketplaces? Affected: tradekey, globalsources, thomasnet, business, all big names and except the first one these are old 10+ years websites. Hmm...

[maxmoritz]

I've looked at my site as well as others for which I have stats. Every page on an affected site moved down, but some moved down a little, and some moved down a lot. For example, moving from #4 to #12 while another page moved from #5 to #49.

There are pages that dropped eight spots that seem to be as strong as the pages that dropped 44 spots. There's other factors that play into the difference in the levels of drop, though. I'm seeing some of those factors, but I need to dig deeper to find more.

I've had exactly the same thing happen with my site. I agree with you here, but I still think it's a case of sites being devalued as a whole (to a different degree with different pages, sure) rather than the algo being changed and one site being hit across the board by chance.

Yeah, now that I read what you're saying correctly ... Could be a 'near duplication and origination of information' type of assessment ... EG There are already 10 of these we have which seem to be published first, so we'll discount the rest that have a high degree of similarity.

Interesting. If that was the case a test to get out of the hole could be re-writing content in an attempt to be substantially different. When content is similar perhaps link authority could be the deciding factor.

However, I don't see why the massive page drops would have occurred, where quality sites have ended up far below many sites of much lessor quality...by every metric.

[Whoa]

Long ago, I did some coding that ran an index of dissimilarity calculation on census data to determine which cities were the most segregated in the country. It boiled a city down to a single number, where 0 was perfectly integrated and 100 was perfectly segregated. I thought that was so cool to boil down a complex issue such as racial segregation into a single number.

Google's doing something similar in this new algo, I believe. It simhashes pages on your site to determine whether they are similar to each other. If they are very similar, those pages don't need to all be in the main index. So let's drop them into some secondary index, and only use them if we really, really need to.

More importantly, the new algo does a simhash on your entire site to see if it is similar to the scraper sites that Google (and we all) hate.

If your simhashed number is similar to a scraper site, well, then, by golly, you are probably a scraper site and should be punished. (Actually, it's probably not a single simhash, it's probaby many different ones concatenated together.)

So, if you got whacked, you are similar to a scraper site in some respects, even if you are a good site. That's my theory. So, think about the attributes of a scraper site and then be the opposite of those bad boys. What does a scraper site not do, because they are robo-created and have crap content, that is a sign of quality? -- figure out the tough signs of quality that a scraper could never do and do more of that. My two cents, anyway.

[ortelius]

It looks to me like something was added to the algo that devalued sites rather than a changed the ranking order. None of the SERPs I look at changed with sites moving up and down and all around. Sites got devalued. Other sites moved up as a result of them moving down. Does anyone else have this view?

That's how it looks to me also. That seems to be how it plays out in my niche. As for the different drops in rankings, such as some pages on a site -5, some -50. Could it be that the algo seeks to reduce Google traffic by X amount across a site? This would result in a wide variety of drops for different pages.

[TheMadScientist]

Hmmm... I can say there definitely seems to be something to be said for uniqueness of content with this update, which I personally don't mind too much.

From what I'm looking at anyway.

[maxmoritz]

My first guess here (and it is a guess) is that the Farm Update leans too heavily on the Scraper Update of a few weeks ago. And the Scraper Update does not yet do a very good job at identifying the original source of content, especially for extensively republished content.

But my site and the site's of other members here (from what I've read) didn't have "republished content". The content on my site is entirely unique.

if you got whacked, you are similar to a scraper site in some respects, even if you are a good site...think about the attributes of a scraper site and then be the opposite of those bad boys. What does a scraper site not do, because they are robo-created and have crap content, that is a sign of quality? -- figure out the tough signs of quality that a scraper could never do and do more of that. My two cents, anyway.

I'm going to test that. It's a nice theory that would only make a site better. While the content on many e-commerce sites including my own, and some "info sites" may be unique, it could look to a machine like spun content from other sites like it. If the content on several sites isn't unique enough, maybe some of them get hammered (although that does seem way too heavy handed).

[Lenny2]

Well Overstock and JCpenny are not on that list... so black hat link buying is not a factor for this update. (IMO)

I am all screwed up because the content scrapers argument seems to be the most enticing idea in the forums... yet, there are so many sites that are content scrapers ranking ahead of us with our content (or copied content from the MFG).

Seems like a lot of OLD websites... so maybe it has to do something with the "old" coding on our old sites? I for one have been "scared" to change much on our site over the years for fear of screwing something up for the G bot.

Did anybody catch, is this list: [spreadsheets.google.com...] all of the websites that were negatively affected or just the most negatively affected? And if they are the "most" negatively affected it is it a complete list or is this just a sampling? As hard as it is to take today's reality, if our site is only one of 300 it doesn't seem like that much collateral damage and probably a relatively easy thing for Google to sweep under the rug.

[TheMadScientist]

The content on my site is entirely unique.

I don't know your site, but...
I'm absolutely NOT trying to pick on anyone or say their site is 'spammy' - Not at All.

IMO There's more than 'one thing' going on...

Here are some thoughts:
1.) Throw a query dependent reading level in to the rankings, like QDF, only different (QDRL).

2.) Add an 'origination & similarity' filter.

3.) Do some 'other things' with document classification and behavior.

The content on your site (pages affected) may not fit a QDRL (query determines reading level) type filter or assessment.

The content on your site may be totally unique, but you might not have been the 1st of say 10 to publish (or more importantly: get spidered) something similar (topically and informationally) out of all the pages GoogleBot has spidered.

The content (documents affected) may fit a classifier pattern of a 'spammy site' based on phrase frequency, related phrases, and 'natural language' factors. +(User Behavior and/or Links)?

The point is: imo fixing the rankings of the (sites) pages affected by this change is going to have to be on a site-by-site and document-by-document basis, after we've seen a complete settle-down in the results for a bit to make sure Google-side adjustments aren't still rolling.

[Joshmc]

On the sites I monitor via the "Site:" I have seen a 67% average decrease in their index. These are all large ecommerce sites.

[tedster]

Here's the methodology behind that Systrix data. They monitor 1,000,000 high traffic keywords. They took US rankings the day before (Feb 22) and the day after (Feb 24) the new algorithm rolled out.

This list is the top 300 sites who lost the most keyword positions. It is total positions lost, not weighted by potential traffic loss.

The 300 list is worth a study, IMO, even without any traffic weight in the stats. Losers include prnewswire, autobytel, daniweb - that's quite a diverse group.

[maxmoritz]

The content on your site (pages affected) may not fit a QDRL (query determines reading level) type filter or assessment.

The content on your site may be totally unique, but you might not have been the 1st of say 10 to publish (or more importantly: get spidered) something similar (topically and informationally) out of all the pages GoogleBot has spidered.

The content (documents affected) may fit a classifier pattern of a 'spammy site' based on phrase frequency, related phrases, and 'natural language' factors. +(User Behavior and/or Links)?

Interesting theories, and they may well be part of it. It would seem very difficult/impossible to do this with any degree of accuracy across all sites on the web, which may be part of why we're seeing so many sites hit that "shouldn't" have been.

[Whoa]

Agree completely with you, MadScientist. Nice post.

A website's content can be unique and yet you can be similar to a spammy site. If a similarity attribute is, for example, the ratio of ad space on your site to text space, and your ratio is similar to that of a spam site, it raises the probability that you are a spam site. There are many "tells" that suggest a site is spammy. Even if you are unique and good, you may inadvertently have enough of those tells to convince an algorithm that you are spammy. It has nothing to do with whether your content is unique and whether Google crawled your unique content before it crawled that same content on a scraper site. It's similar to Bayesian spam filters -- your email can be unique but it can still get flagged as spam based on its characteristics.

[tedster]

It's similar to Bayesian spam filters -- your email can be unique but it can still get flagged as spam based on its characteristics.

That's a very good way to think about this update. And since the actual algorithm took over a year to develop, it's probably a lot more complex than a Bayesian filter.

[chrisv1963]

And since the actual algorithm took over a year to develop

Did they seriously need a year to develop this crap?
A few minutes ago I found a Blogspot blogger that copied my text, hotlinked to my images (5 images!) ... and is now ranking higher than my website, the original.

[Lenny2]

HOT OFF THE PRESS!

A major competitor of ours, a major website with lots of copied content... NOT ON THE RECENT LIST, has fallen from the first page of google for a large % of key terms.

Anybody else noticing any movement? Sadly, I can't say we have necessarily moved up :(.

And I certainly feel for the competitor. But the biggest move I've noticed since the new algo launch.... and maybe an indication that more is to come.

This all happened within the last 1 hour.

[Rhonie]

The following is common sense to me on how Google may have changed their algorithm. In no way shape or form that this may be correct, but if I was ahead of programming for Google it would run about like this for several reasons.
Now I am not a pro, mathematician or anything close, but this is what I have about figured out:
If you are a retailer it is very important that your description of a product is not the same in another site, you need to adjust that, and you need to link to quality sites for reviews of the product you are selling for credibility.
Now the first = 1 if someone would enter the exact name of your site. No deduction for that, but you would get a deduction and a possible sandbox if your site was like Seers.com, and your content in the page anyway resembled Sears.com. Also, if your site contained too many keywords with the name of Sears.com, and had content not related to Sears.com, or any main keywords that Sears uses, perhaps like Craftsman or Kmart.
Now news sites that are big that have published content are not punished, or sites with a huge amount of visitors either. There may be an exception, or inclusion in the script for sites that keep up a high number of hits per page written in the algorithm, the script may be written for this to change on a daily basis.
Lets see if I have this about right?
In no way I assume the point system is correct either, but again I am just using this is a layman's point system to make it easier to understand.
The second "you" was placed intentionally just to show that many sites may be inserting or changing to another word in between the article content, I am sure their are many like I, my, our, etc...
The following is just my example of algorithm for Googlebot:
Possible points for sand box = -10,000 points
Ranking on page for keywords top = 1
Penalty for points in word sequence are is follows:
if = 1
you = 2
use = 3
the = 4
same = 5
words = 6
in = 7
an = 8
article = 9
you = 10
after = 11
the = 12
first = 13
hit = 14
time = 15
and = 16
date = 17
than = 18
your = 19
site = 20
points = 21
will = 22
be = 23
deducted = 24
this = 25
many = 26
points = 27
total = 28
deduction = 29
of = 30
points = 465
Highest return on page due to wording sequence is on page 4 3/5th.
If total matches other = page post to newest first hit time/date ahead for wording sequence. Comparative analysis for other sites = below or above, number of sites listing the exact keywords.
Grade for age of > one year, < one year.
Add 5,000 to sites that have exact wording content using first find time/date = 0 points for posted ad within 24 hours, credit points for sites that have more than 5,000 new visits per day per page.

Now you all need to think outside of the box. You must assume that Google is a company that needs to make money, in advertising, statistics, along with many other ways and this is done by organizing sites to the best possible stats that interest their users. The problem with this, if it is figured out completely, they will just change the algorithm again.

Now another way to look at this is Craigslist. Why does Craigslist give away free postings? I would hope you all know. This is so they get a lot of hits to the site, and the content remains fresh to get a high Google rating, so they can charge to post paid ads in major cities, with a minuscule effort in advertising.

I would like for any of you to respond that either may agree or disagree with this? I know it is along way from exact, but this is about what I am understanding.

[Rhonie]

chrisv1963, Do you run your own server or are you on shared hosting? Think about it.

[Dead_Elvis]

chrisv1963, a little creative .htaccess work should take care of those hot linked images for you. And at least give you something to laugh about :)

[vordmeister]

Just an observation here:

I searched for something a little out of the ordinary today and as usual google did not let me see results for it. They just showed me a load of rubbish. Bing was a little better, but I ended up taking better advice from a forum.

It made me wonder where we are going here. Search engines were handy 5 or 10 years ago, but I'm struggling to find out how to use them these days.

Has it got to the point where the spammers have made search engine algorithms look stupid? If search engines rely entirely on automated systems why is it a surprise that humans who write rubbish but are good at SEO are ahead of you?

(My own sites have improved ranking in the latest update, and they are worth it of course, but I would rather I had decent search results when I try to search for something.)

We've got thousands of mostly negative posts here about the latest update. I'll bet they would have been outweighed by positive posts had the job been done by a human.

[edited by: vordmeister at 8:56 pm (utc) on Feb 28, 2011]

[ckissi]

@chrisv1963 , how many pages are indexed for that Blogspot blogger ? It seems to me that smaller sites win because they're simply smaller, so not too much content there (not a content farm). Maybe we loose because of this and smaller niche scraper sites wins from this.

[AlyssaS]

there are so many sites that are content scrapers ranking ahead of us with our content (or copied content from the MFG).

Do you know how they are scraping you? Most common and garden scrapers use your RSS feed - set this to summary or short and they just pick up the first few sentences, which is "fair use", rather than your full articles. Is there any reason why you need your feed to be set to "full"?

I've never had my content scraped, because I've limited the feed. Of course having URLs scraped so that some bot can come back later to post junk comments is another thing altogether. They only way to stop them is to ban IP addresses - they usually use proxies - and many use the same proxies, so once you find one ban it.