Google's AdSense Farm Update Was a Re-ranking - NOT a Penalty

[TheMadScientist]

I know quite a few of us have said the changes in the SERPs related the AdSense Farm Update is not a penalty, and I've tried to explain it, but there are too many threads to get it in all of them and not everyone reads every post, so I'm going to go ahead and post this in it's own thread so people can link to it rather than trying to explain the difference if they would like.

PENALTY CHANGES - how they work

BEFORE PENALTIES
#1 = will get penalty
#2
#3 = will get penalty
#4
#5

Every result with no penalty just moves up, filling in the gaps that were opened.
They all stay in the same relationship with each other.

AFTER PENALTIES
new #1 = was #2
new #2 = was #4
new #3 = was #5
new #4 = was #6
new #5 = was #7

RE-RANKING CHANGES - how they work

BEFORE RE-RANKING
#1
#2
#3
#4
#5

All the results now get shuffled, some go up different amounts and some go down:

AFTER RE-RANKING
new #1 = was #3 [up 2]
new #2 = was #21 [up 19]
new #3 = was #2 [down 1]
new #4 = was #1 [down 3]
new #5 = was #11 [up 6]

So there's a big difference between a penalty change and a re-ranking.
If you only look at drops you don't see the bigger picture of what happened.

[edited by: tedster at 7:47 pm (utc) on Mar 1, 2011]

[tedster]

TheMadScientist proposed this post, and then I worked with "The Mad One"B (hazard pay applies) to make it as easy to understand as we could. I hope we succeeded. If there are still questions, please ask away.

[TheMadScientist]

Thanks for the help tedster ... I've had issues communicating in English since I stared writing PHP.

if($pages==='Lower results only move up to fill empty spots') {
$change='Penalties';
}
elseif($pages==='ReOrdered'
&& $pages==='some up'
&& $pages==='some down'
&& $summary_of_changes==='jumbled rankings from previous resultset'
) {
$change='ReRanking';
}

Makes complete sense to me, but I guess you're right there are quite a few who would probably scratch their head and think: 'wtf is he trying to say?' lol

Thanks for the Help!

[universetoday]

Great, now the tougher question. What was the rationale for the re-ranking? Has anyone discovered common characteristics between the sites that went up or down?

[TheMadScientist]

It was the new variable 'Quality' they added to the algo.

As far as I can tell, and from what I've seen it's got less to do with links and topicality than most seem to think. And from what I've looked at it seems to have quite a bit to do with 'Document Footprints' being evaluated.

[Lenny2]

Sounds about right to me...

[ken_b]

'Document Footprints'

Excuse my ignorance, but does that mean Page Templates?

[1script]

Since most people would come here after suspecting that they have had a penalty handed to them, it may be useful to rework the AFTER PENALTIES box to include some info about the penalized sites #1 and #3

BEFORE PENALTIES
#1 = will get penalty
#2
#3 = will get penalty
#4
#5

Every result with no penalty just moves up, filling in the gaps that were opened.
They all stay in the same relationship with each other.

AFTER PENALTIES
new #1 = was #2
new #2 = was #4
new #3 = was #5
new #4 = was #6
new #5 = was #7
new #54 = was #1
new #889 = was #3

Actual position after penalty depends on the type of penalty and fluctuates a lot due to lower SERPs being more affected by noise and everflux.

[Lenny2]

Thinking that it has to do with the amount of duplicate content on the page... I've seen a lot more craigslist / ebay listings (as an example); they have totally unique content (albeit shallow and low quality) but, what they don't have is a lot of a footprint or the same type of content that are on many pages. The percentage of unique content compared to duplicate content on page and on the site is very LOW; and it seems that the engine is awarding sites that have a LOW % of duplicate content...

This is perhaps a pretty simple explanation... probably would take less than a year to figure it out, but, anybody else have any thoughts?

[ponyboy96]

Thanks for pointing me over here. I'm starting to agree with this one more and more as I analyze rankings across several sites. In many cases the traffic shifts were so small you couldn't tell by looking at it, but after digging some more, it's there. Some up, some down, one big drop for one of my sites. Ranking shifts across the board, again some positive some negative. The changes are above and beyond the normal deviations that are seen.

[ponyboy96]

Lenny2 - I'm thinking this was a shift to more authority type sites. Well known brands and the like. That's the only thing I can think of. I also think the spam meter for duplicate content also got turned up a few pegs and other sites with more authority than you got the move up even when the copied someone else's content.

[TheMadScientist]

... does [footprint] mean Page Templates?

[ken_b]

I think it's the entire Footprint of the page (document), so does it translate to only 'template' no, not that limited, but include, yes, imo it may in some ways.

[Content_ed]

You're making some simplistic assumptions about the way penalties must work. If there's a penalty on the final score computed for any give phase from a site, it would yield results indistinguishable from what you describe. In other words, Google may have said,

We knock 10% off domain #1 for lack of trust
We knock 30% off domain #2 for lack of trust
We knock 50% off domain #3 for lack of trust
etc

Now, how does that affect the rankings of these sites for a given phrase? The answer is, it depends on what their scores were for the phrase was before the penalty (ie, how far they were separated), and on the other competition.

You may have examples where you track every domain involved for a particular phrase, before and after, so you could crunch the math and prove that it isn't a site-wide penalty. I don't have that capability.

But it's equally possible that a penalty is applied to one of the major factors in Google's algo, and that factor isn't always given equal weighting. I believe that would make it quite impossible to reverse engineer unless you had a huge sample.

Your reranking could easily be dependent on a flat penalty applied to something like the "uniqueness" part of the score, which Google may very well weigh differently for different searches.

[jk3210]

I suggest we stop using the term "penalty" unless one is truly suspected, and starting using the phrases "up-scored" and "down-scored."

[trakkerguy]

I'm agnostic (so far) about the factors Google used to implement this change. Lots of good thinking and analysis backing up all kinds of possibilities. One question I haven't seen thrown about yet is:

Why did they make this change all at once?

I can think of 2 reasons.

1. They wanted to send a message.
2. They HAD to do it all at once.

Lots of different question and conclusion arise depending on which of these answers you chose.

Unlike some other algo changes where they seemed surprised at the impact, this time they seemed to know or intend for this to be very noticeable. Else why did they publicize this so much in advance?

Was it all done this way to scare people away from the content farming profession? For PR reasons?

Or was there something about the tools they used to make this change which dictated a very noticeable change that they normally avoid (because they don't want us to know how the algo works).

[TheMadScientist]

@Content_ed

Unless you know Quality is a new factor in the algo, and some sites received a boost in scoring from the new factor ... I can't find the reference for it, but it's here somewhere.

Also, in your situation it would be possible for #6 to replace #5 but even though I guess it's possible, it's not likely for pages to 'come out of nowhere' like they did with this update. EG #342 would not likely end up in the top 10 from a penalty situation, but would from a re-ranking and this has happened in the SERPs I watch ... More than one set.

Pages came from out of nowhere into the top 10 in the SERPs I've watched and that's not a likely result of the former pages in the top 10 results being penalized.

I think the reference was in a tedster comment, but I'm really not sure. I KNOW it's here somewhere though. There were pages that got scored lower and pages that got scored higher by this one ... I've honestly scanned 20+ pages of threads looking for the reference so I could cite the source but I can't find the stinking thing! Sorry. This one was a re-ranking, not a penalty situation.

[netmeg]

I dunno, maybe it's just me, but it's looking more and more like what's been happening with AdWords over the years. The whole "Quality Score" idea (insofar as we can ascertain it from Google) is not an easy concept to wrap your head around, but once you *do* get a grip on it, it gets easier.

But in the AdWords forums, we've seen the same types of things for years that we're starting to see here - people getting severely dinged for QS, and it doesn't make sense to them because as far as they're concerned, their sites are great. And often, they are. But "great site" isn't all that goes into the equation. And Google won't tell you what goes into the equation (or they do it in such a fashion that really doesn't tell you anything) so all you can do is test and try stuff till you come up with something that works.

(Then they change the rules again)

Now I'm depressing myself.

But it does sound like the same type of evolution that AdWords has undergone to me.

[walkman]

Unlike some other algo changes where they seemed surprised at the impact, this time they seemed to know or intend for this to be very noticeable. Else why did they publicize this so much in advance?

Was it all done this way to scare people away from the content farming profession? For PR reasons?

At least 51% was PR, remember the articles written in the preceding months and weeks bashing google

[Content_ed]

Mad,

I don't watch a lot of results so I can't say, though I'm guessing that pages out of nowhere are probably showing up for pretty uncompetitive phrases where nobody really scores highs so a big move is no deal. It also seem to me that in every update pages rank from out of nowhere, simply because they have picked up a couple high quality links since the last redo.

I don't have any emotional investment in it being a penalty as opposed to a serious algo rework, I just don't see it as carved in stone one way or the other.

My main assumption to this point is that the issue is Google being unable to indentify where content originates and is therefore penalizing, or reranking, sites that copy a lot (article farms) and sites that get ripped off a lot because their content is attractive to the cut-and-paste crowd. Whether Google would express that as a "quality" figure or a "duplicate content score" is beyond me ken.

What I do know is that eHow, a big winner in all of this, contains uniquely bad content that nobody would be tempted to cut-and-paste:-)

I think back on what was my most popular page before the update, which drew close to a thousand visitors from search on a good day (like Mondays). It started out around four years ago drawing a couple people a day because it was linked into my site infrastructure. In a couple months that went to 30 or 40 a day. A year later it was 200 or 300, a year after that it was 500 or 600, since last fall it's been closing on 1,000.

The only thing that drove the page up in the rankings was that people liked it a lot so they linked it. They linked it from forums, where they often pasted the whole page in, and sometimes they pasted without a link. They linked it from professional organizations and from schools, or they just added a text attribution with my name. In fifteen years of publishing online, it was my second most popular page, it just happened to grab people the right way.

The update knocked it down a good 50% (the subject is US centric). No changes on my part, never any SEO of any color. Now Google could have reworked their algo a lot of ways that could do that. One way, in my mind, is a penalty for duplicate content, since that page and other pages from the site have been replicated like crazy.

To me, it looks similar to a single page duplicate content penalty I was hit with years ago, which cleared up when I got the high page rank duplicate taken off the website of a major corporation. But it could be other things. What it isn't is a bunch of new pages that address the problem better.

I'm not going to start messing with my pages in hopes of changing something to meet a change in Google. I don't produce a lot of pages, and this is an example of a page that took me several days to write after I gathered the information and knowledge needed. Webmasters who crank out a lot of pages are much more knowlegable about competing on Google than I am. The Internet isn't my primary business.

[TheMadScientist]

Yeah, I see what you're saying, and I get what you're saying about not caring wtf it's called. Really, I get that part.

The point of this thread though was to get some clarity on the semantics that more accurately describe the scale of the change, and also to help people understand looking for the cause of a 'penalty' is probably not the solution...

This was an 'old school' type change, like the ones that were discussed before any old ranking change for any reason was called a penalty ... Today 'penalty' is often used to describe any ranking change and looking for 'what you did wrong' (the answer to a penalty) doesn't seem like as good a solution for this situation as looking at what you need to do differently.

To fix a penalty situation you find the cause and remove, fix or adjust it. But this isn't that type of situation as far as I can tell. This is a large scale, long-term scoring change, so imo you're not going to find the link(s) causing the problem. You're going to have to change the way you do things ... IOW I think the fix is a bit (understatement of the year) more involved with this type of change.

I hope I'm drawing a distinction for people, because I doubt rankings effected by this are going to be fixed 'lightly' or easily. I think a 'fix' is going to be a rather involved process, and that's the distinction I'm trying to draw for people with this thread.

One huge difference that helps understand the distinction I'm trying to draw is with a penalty a re-inclusion request might help you out. I don't think there's much chance you're going to regain your rankings through one of those if your site was affected by the new algo, because the scoring system is different now ... To 're-include' your page(s) back at the old ranking levels they would have to change the algo back ... Not gonna happen, and that imo is a huge difference and important distinction to make.

BTW: I feel for you on the lost traffic...

[Lapizuli]

If what you say is true, then one of my rankings that fell is true to the penalty model. I.e.,

Before update:

Result 1
Result 2
Result 3
Result 4 (me)
Result 5

After update:
Result 1
Result 2
Result 3
Result 5
Result 4(me)

[TheMadScientist]

Or it got re-ranked below #5 due to a scoring change in the algorithm.

If people want to call this a penalty like they do every other time a page moves down in the rankings these days, then so be it, there's nothing to stop you, but personally if I thought of it as a penalty rather than a re-ranking and new scoring method I'd be chasing my tail looking for a way to fix it rather than looking at what I need to do differently...

[TheMadScientist]

This is a repeat of one of my posts from the now March update thread. Hopefully it makes a bit of sense...

Penalty = You did something wrong, fix it and get back to the top.

Re-Ranking = You're dealing with new criteria ... Think 'new search engine' that does things differently than Google did a week or two ago and to rank in the 'new search engine' you're going to have to find a way to fit the new criteria.

Think about the difference between Google and Bing.
They use different criteria to rank pages.

Ranking well in Google but not Bing does not mean you've been penalized by Bing.
They use different scoring methods...

Google today is not using the same criteria as Google a week or so ago. It's like trying to rank in a different search engine now ... There is no penalty to be lifted; there are new criterion in place.

[brinked]

today a website has bounced back hard, it was hit by this update by today it had its rankings returned and is ranking much higher for premier key phrases which IMO it really shouldnt. So this re ranking theory does make sense. I am going to wait another few weeks to make a final judgment of this.

[dickbaker]

TheMadScientist, I'm not arguing whether there's a penalty or not, or whether it's a quality score. I haven't been able to look closely enough to figure out what factors might be included in a quality score.

What I don't understand, and what I keep asking about, is how a page that ranked at, say, #7 could be sent to #70 in a single day. Other sites took the place of those on page one, sites that Google apparently thinks are better quality. I'm not arguing that (although none of us can prove conclusively why the one page was moved to #70 and the other pages appeared on page one).

My question is, why #70? Is every page from #1 to #69 of better quality than the page that, until last Thursday, ranked #7? That's a lot of pages, and a lot of real garbage.

I'm not disputing your theory about what's on page one. I'm questioning what Google sees in the pages between page one and the new resting spot for a demoted page, since I see--in every niche--some horrible content for the given search phrase.

[econman]

Interesting, useful thread. I hope we can keep it clean and somewhat narrowly focused -- or at least avoid too much overlap with the other 3 threads.

Toward that end I'll try to keep my own comments focused, and not broaden the topic too much.

I agree this isn't correctly described as a Penalty.

I think we should accept at face value Google's claim that they are trying to change their ranking system to down-rank, or suppress from view, content they think their users will perceive as being of "low quality"

For now, there is no indication they are trying to detect or boost the visibility of "high quality," except as an incidental side effect of pushing down "low quality."

That may explain why we don't hear about sites suddenly shooting up the SERPs.

I think we should assume this is a multi-faceted evaluation, which primarily focuses on machine-quantifiable attributes of each page and/or entire sites (still not sure which).

If it isn't a penalty, I don't think there is anything easy that can be done to a site to immediately regain prior rankings -- other than to focus your energies on increasing the quality of your pages, or adding fewer, higher quality pages in the future, rather than cranking out more low-quality pages.

Consider, for example, a site like ezinearticles which has admitted they have pages of inconsistent quality. If they were to identify and purge the worst 40% of their pages would they immediately regain their rankings and traffic? I doubt it. If Google's new algorithm is identifying the low quality pages, and ignoring them or pushing them into some sort of secondary index, then physically removing the pages from the site isn't going to accomplish much.

Over the long haul, however, by imposing higher editorial standards and encouraging their contributors to write fewer, longer, articles that don't simply regurgitate/repurpose/copy content published elsewhere, than perhaps they can improve their situation over time.

[TheMadScientist]

So this re ranking theory does make sense.

They make adjustments to every single major update after it rolls out ... Sometimes they even roll the changes back, adjust and re-roll out an updated version. I'll let you and others have a look at the old update threads to see if what I'm saying is at all true: [webmasterworld.com...]

Rollout and rollbacks used to be commonplace.
It could well be what you're seeing.

It could also be an adjustment to the scoring. This was a HUGE change and I've said in other threads I would give it time to settle down, because it used to be common knowledge they would roll and rollback and adjust updates for weeks after an initial implementation.

My question is, why #70?

That's like asking why was it #7 in the first place?
Was it really more relevant than say 1,293,452 other results?

That's the Multi-Million dollar question.

Remember 63 places different sounds like a bunch of places and in the results it is BUT when you're dealing with 1 million possible choices the change is only a difference in overall score of 0.00007-0.000007 = 0.000063.

In the overall scoring of pages it's not off by very much from where it was, and someone's page always ranks at #70 ... My guess is yours just happened to land there.

[edited by: TheMadScientist at 11:20 pm (utc) on Mar 1, 2011]

[Reno]

if I thought of it as a penalty rather than a re-ranking and new scoring method I'd be chasing my tail looking for a way to fix it rather than looking at what I need to do differently

You are exactly right. A penalty implies doing something wrong (in Google's eyes), whereas a re-ranking occurs as a result of a re-evaluation, which simply means they've formed a different opinion.

Here's a simple way to think of it: You do your best singing act on American Idol and the judges score you an "8"; then the next week you do the exact same number with the exact same level of talent and a different set of judges score you a "5".

That's what has happened with this update ~ the revised algo is like a new set of judges and it sees things differently. There's a bunch of people who are going to try to "fix" their sites, but since they did nothing wrong to begin with, there is nothing to fix. The re-evaluation has simply lowered Google's "opinion" of you, and until we know what they mean by "quality", we are chasing after shadows.

.....................

[brinked]

TMS, exactly, and that is why I am going to wait a few more weeks before making any kind of opinion on this update.

You are 100% correct in that when google makes an update, its not just one thing they are changing to their algo, that is what a lot of people have a hard time understanding, they are all looking for that 1 common denominator.

[Content_ed]

Mad,

So I just checked Analytics for the last five days, vs the same period previous week. Sorry I didn't look before. Of the top 100 pages, all but 3 showed double digit losses, between 20% and 50%. Of the three pages that showed gains, the gains were less than 5%, and may not have come from Google (I didn't bother checking).

The 100 pages cover eight different subject areas, though four are interelated enough to share some common navigation. Ages vary from ten years to one year. They represent roughly half of the non-blog pages on the site. Some are pure text, some are text and graphics, some are text and video. All have attracted external links to various degrees.

Based on this, I don't believe Google is reranking all of the pages based on some individual, page-by-page metric. I don't have another practical explanation, other than a site-wide penalty. I will try a "site reconsideration" through Webmater Console - thanks for reminding me it was there.