New Panda Update - September 27

[BrodyDodes]

Google rolled out its most recent iteration of the Panda algorithm on Wednesday 9/28/11 (or Tuesday according to who you ask). Saw the first recovery of one of my punished sites since the first iteration back in February. Anyone else see recoveries?


[webpronews.com...]

When asked if an iteration of Panda was implemented this week, a Google spokesperson told us, “yes.” She also provided the following statement:

“We’re continuing to iterate on our Panda algorithm as part of our commitment to returning high-quality sites to Google users. This most recent update is one of the roughly 500 changes we make to our ranking algorithms each year.”

If you’ve followed the Google Panda update saga throughout the year, you may recall Dani Horowitz’s story. She runs an IT discussion community called Daniweb, and it was hit hard by the Panda update, but she made a lot of changes, and gradually started to build back some Google cred

[viggen]

I dont buy that whole bounce rate thing, as I have several pages where Google sends me for years the wrong visitors and it had zero effect so far on ranking. One page in particular ranks Nr.1 for one term with 97% bounce rate and time on page is like under 2 seconds, it is unchanged around 500 visits a day.

Its Googles fault, I never optimized this page for anything, its a book review for an old book its written by a professor who was an authority on this subject. This review is cited by universities, newspapers like NY Times, by institutions like Stanford University etc... from the 500 visitors a day 3 to 5 actually look for information on that book, the rest for something else that sounds similar but has a total different meaning;

now again I am wondering; Google sends me for years the wrong visitors and now i should be punished for it?

As I mentioned before, I stopped caring for Google a while ago and do my own thign with my sites as i see fit. The more i tread Google like an unwanted child the more it begs for attention...


cheers
viggen

[zeus]

dunivan - a user can also be longer on a site be cause he can not find what hes looking for, but still think it must be on that site.

[Dave_Hybrid]

People are in denial. Google can and will use real user data in the algo. Don't fight it, it is/will happen.

"A user can browse a site for a long time because they think what they want is there but never find it."

"A user can bounce in 10 secs after finding exactly what they want."

So what...

Everyone's examples of A user might do x, is true, there are always exceptions.

However when you are google and have access to data you cannot comprehend things stand out... not the exceptions... the rules.

Not to mention everyone is simplifying things. Who in there right mind would use bounce on its own as a signal. But when you combine time on site, bounce, pages per visit, new or returning visit and countless other things you get a more complete picture.

Really, can we move on, this is getting a little old.

[edited by: Dave_Hybrid at 7:54 pm (utc) on Oct 10, 2011]

[londrum]

do we really know that bounce rate is a big factor? everyone is talking about it, but maybe it's just a red herring.
of course, it's bound to be part of the algo in some form, but would they really base so much on it?

[walkman]

People are in denial. Google can and will use real user data in the algo. Don't fight it, it is/will happen.

It has been happening for 5+ years, at least, everyone knows it. Bing uses it extensively as well but for keyword /page which is the correct use. If Bing sends you a visitor for a keyword and they return right away they will make note that your page might not be the best match for that keyword. But on the other hand they will keep sending you referrals if the users is 'happy.' I get dozens of referrals for 2 of my pages for an extremely popular keyword from Bing.

No one is denying, we're questioning what and how Google might be using such data to penalize entire sites. Decade old sites all of the sudden lost as much at 80% of traffic.

[aristotle]

walkman- your site isn't the only one that gets a lot of mis-matched long-tail traffic from Google. It happens to all of my sites, and probably most others as well. Instead of complaining about it, you should take advantage of it by creating pages that grab peoples' attention regardless of what brought them to the page.

[potentialgeek]

My main site gets revenue by display ads. I've been focusing lately on user experience and increasing average pageviews. So far I've been able to almost double average pageviews for the site. If and when I recover from Panda, I could make twice the revenue every day.

I prefer to invest my time in work that produces instant results with extra revenue every day than all the guesswork on Panda that has led to little value for most webmasters.

I've only found two pages on my main site that weren't Pandalized. Have we had a thread on pages that dodged the bullet? And what clues they could provide to the Panda algo? It seems to me they could provide the best data on recovery.

Page #1 - Long with lots of details, like a good article, good inbound links.

Page #2 - Longer than most pages on site, has a research chart, a few images, and a relevant embedded video at the end.

My sites that didn't get Pandalized had long pages. Some of my long pages on my main site that got hit the worst have text that is too similar (looks spammy). It's probably not unique enough. (It also has a lot of Javascript, maybe too much.)

I need to create a new page soon. I have to make sure it looks like an article, it's long, unique, and not repetitive.

I think of Panda as the English Professor Update. I can't submit anything that the prof has seen before. It can't have plagiarism. It can't be short. It needs to be well written. It can't simply state the obvious. Something about it must be original - data and/or opinion - preferably both.

Google's list of questions on quality control seem to mirror what a professor would want. If you were to submit your Pandalized site to an English professor would it be accepted?

P.S. I think we have to remember that previous algos, which may still be operating, factor changes made to pages. Too many changes to fix Panda could backfire.

[danijelzi]

@potentialgeek
The situation you described above is exactly like mine, with a site with display ads. The pages not affected by panda on my site are those with longer texts, some with shorter texts but with embedded videos, and pages with more originality and personal opinion.
The worst hit are those with less text, but also longer which are similar to other pages on the site, no matter how detailed they are.

What I couldn't fix I simply deleted (404), other stuff got expanded and added media - images, videos - and on September 28th I saw the first signs of Google traffic recovery since February.

[suggy]

@danijelzi

Yes, derivative content is definitely a no-no under panda. No point rewriting same material for slighltly different search term. But, I guess that goes to the heart of what Google was trying to tackle - writing for search engines?!

[Hissingsid]

There are more and more clues that point to semantic vector analysis being an important part of Panda.

Semantic vector analysis can be used to generate lists of potentially relevant terms based on the search term entered. Hence increase in number of long tail terms.

Semantic vector analysis can be used to efficiently store semantic classification data of sites and pages.

Semantic vector analysis can easily identify use of language that is not natural, duplicate pages etc.

Add to this a rules based content policy that benefits certain types of content plus user experience data analysis and you have a system that learns. Google's semantic understanding of what users want when they search for certain terms grows over time.

Has Panda been applied to all languages?

[Dave_Hybrid]

Just got a non canned email from an adsense rep asking why i no longer display adsense on my site and how they can help. This will be interesting.

[Dave_Hybrid]

Well... seems like a waste of time, maybe some good will come of it...

Rep: Why have i removed adsense?

Me: Because you've taken away all my traffic!

Rep: Oh, lets take a look. <Long chat about Panda and how it seems im a false positive>

Me: So what can we do here?

Rep: I can help you add a border and bigger font to your ads to see if revenue increases.

Me: LOL. Thanks but bye.

One interesting thing was, he categorically said ads had no direct effect on search ranking. None.

Although he couldn't say for sure that they dont have an indirect effect, ie 20 ads causing a high bounce/low time on site/low page view per visit/low return visits etc.

He did ask some interesting questions, how will i survive without an adsense income (not talking a few hundred bucks here), what my future plans are etc. I simply said im using savings and I will eventually get a job.

I explained it is no different from setting up shop in a town, then the local authority stop people coming down my street and refuse to let me move premises. Google dominates this market, they let me build a business based on their referrals and then took them away.

He said he is hearing this from lots of publishers and will pass the info along. I'm not sure what good it will do.

[edited by: Dave_Hybrid at 10:08 am (utc) on Oct 11, 2011]

[pexcornel]

at Google the left hand has no idea what the right hand does.
I don't even care about their traffic, I suggest everyone to do the same, tracking every move that Google does is a waste of time.

[Dave_Hybrid]

It's hard not to care about Google when you lose 100k a year. I'm guessing you're not in that boat. But many of us are. Doing what Google likes definitely hasn't been a waste of time for me for the last 6 years. Although atm anything i do seems futile.

[walkman]

Has Panda been applied to all languages?

Other than Chinese and Korean I think.

at Google the left hand has no idea what the right hand does.
I don't even care about their traffic, I suggest everyone to do the same, tracking every move that Google does is a waste of time.

I thought so too once, but they seem to always do well for themselves. We have to care for the traffic since it's 70% (when counting AOL and an few other portals) and almost all people use search engines to find info. When you meant is that we need new search engines :).

It's hard not to care about Google when you lose 100k a year. I'm guessing you're not in that boat. But many of us are. Doing what Google likes definitely hasn't been a waste of time for me for the last 6 years. Although atm anything i do seems futile.

Been in this boat for 7+ months. The trend is even more worrying as independent webmaster are being squeezed out little by little and that's by design IMO. The only safe bet is to design a site that Google cannot penalize, Panda or no Panda, like Cultofmac, Daniweb, Android Police and the likes. Otherwise the commercial side will be full of huge brands that paid their way there, either by bribing people for links /G+ or paying to build the 'trust' Google loves so much.

[Dave_Hybrid]

Sucks, doesnt it, im trying not to let it get to me but it's making me very ill.

Watching that kind of money wash away is soul destroying.

Im not in the commercial space, im an info site just like the examples you made.

I'm still hit, hard.

[walkman]

Im not in the commercial space, im an info site just like the examples you made. I'm still hit, hard.

Panda and their massive brand bias even resurrected ZDNet, Cnet, PcMag and other once left for dead tech brands. They are on page one for pretty much any tech related search, so you add them and those manually exempted from the Panda rules and your drop may be explained. Level playing field and all

[Rasputin]

Thats the ironic thing - since I'm getting on a bit I always expected to be outpaced by technology, and tried to have a great looking site site with ajax photo galleries etc - and now I find that many queries are returning sites that haven't been touched for 10 years and are the simplest html sites with tiny graphics you can imagine (travel niche), while we languish invisibly on page 5.

Earlier today I was in a site where a butterfly graphic followed my mouse pointer around the page (until a pop-up window crashed the browser) which is something I haven't seen for a long time...

I guess the conclusion is that we've been trying too hard. Ditch the databases, ajax, scripts and CMS and return to simple and very quick loading hand coded websites, even if the information is horribly inadequate and out of date.

[aristotle]

guess the conclusion is that we've been trying too hard. Ditch the databases, ajax, scripts and CMS and return to simple and very quick loading hand coded websites, even if the information is horribly inadequate and out of date.

I totally disagree with this. Just because a site is hand-coded, doesn't mean that the content is poor or incomplete. And a new article isn't necessarily better than an old one. It's the same with a lot of things. Many old books are far better than most recently-written ones. And many old movies are far better than most recent releases. Don't condemn a site just because it's old or hand-coded. You also don't seem to realize that some new sites are still being hand-coded.

[dunivan]

Speculation says google didn't like their own update: [blog.searchmetrics.com...]

anyone hit by the latest iteration of panda show recovery lately?

[Simsi]

(MikeNoLastName) - Therefore they can only evaluate what they CAN see and that is: whether a person RETURNS to G and SELECTS _+ANOTHER+_ SEARCH RESULT USING THE SAME SEARCH TERMS!

I think this is just the tip of the iceberg. If Google sees user behaviour as key to understanding a site's value it has to know the user. Intimately.

It wouldn't surprise me to learn that Google knows a hell of a lot more about our behaviour than we think. In fact, I wouldn't mind betting there is a seperate algo that specifically profiles each user (anonymously or otherwise), an algo that is - or will become - comparible in complexity to the actual ranking algo itself.

There is so much they can see besides that point above. By classifying searches and sites they could analyse and profile every seach we do to build up historical behavioural patterns for each individual, They could probably tell if we are shopping, researching or seeking something and then compare how we normally do each of those.

For example, if we "normally" exhibit behaviour - for example like backtracking when shopping for price comparisons maybe - then future sessions can be measured against normality.

If you put your mind to it, you can start to come up with a ton of "signals" that we provide every time we hit Google, from the computer we use to the way we use it and how we interact with the web. Even our spelling and grammar could be useful for G in profiling individuals.

[walkman]

Speculation says google didn't like their own update: [blog.searchmetrics.com...]

anyone hit by the latest iteration of panda show recovery lately?

I actually lost some of the small boost I finally got on Sept 28th. It was Google's let's-get-DaniWeb-off-the-news update so it was for a good cause. Naturally, as a good netizen, I gladly accepted it. Today looks I lost a bit more, but I don't know why since I don't follow the twitter messages exchanged among the cool crowd.

The rest of webmasters and business owners are suffering in silence for now. Amazing how the 'scientists' at Google find data, depending on what type of press they get. Must be those 4.0 GPAs ;)

[falsepositive]

I'm not sure if it's seasonal -- and I've expressed this in another thread on Panda & Seasonality. My 2 sites (Pandalized and non-Pandalized in the same niche) are up. The Pandalized site is up more than the non-Pandalized site but both are doing better. Again, not sure if seasonal or Panda-related (I don't really analyze keywords and don't do hardcore SEO) or maybe both!

[Rasputin]

@ aristotle,

From the statement 'very quick loading hand coded websites, even if the information is horribly inadequate and out of date'

You can't conclude that I was suggesting that for all hand coded sites 'because a site is hand-coded the content is poor or incomplete'.

I neither intended that conclusion or believe it to be the case.

I was suggesting that many old sites appear to have some characteristic that many newer more sophisticated sites lack, that outweighs the frequent lack of content on the older sites and enables them to rank unnaturally highly.

Whether it's link age, domain age, content:code ratio, faster loading due to lack of scripts and graphics, lack of adverts or one of many other possible factors I can't say.

[aristotle]

Rasputin Apparently I misunderstood you. I thought you were associating hand-coded pages with poor content, as if they frequently occur together.

But now I disagree with your new statement: "that outweighs the frequent lack of content on the older sites and enables them to rank unnaturally highly." I don't think there is a "frequent lack of content" on older sites. If a site survives long enough to be called "old", that often means that it has good content. Conversely, a new site with poor content often disappears within a short time.
Quite often old wine is better than new wine.

[Zivush]

Quite often old wine is better than new wine

It is true, but if you take testers/tasters, you get for every wine different opinions.
So how an algorithm can judge a site accurately?
What makes one page better than others on the same keywords - Social media signals, BR, time on site, site's credibility, the author name?
Of course, a high traffic site gets more social signals (maybe its BR is better too) but does it make its content better than any small site written by an area expert.
Hard to tell.

[Hissingsid]

Are brands being boosted because their brand names are semantically very close to the generic?

For example Amazon ~= Books

I noticed a few weeks ago when I was searching for bits for my motorcycle if I put in the model number of my bike along with say mirror (searching for new mirror glass) I saw in SERPS other motorcycle mirrors. Google new that the unique model name was a motor cycle. This has now changes and more sites are ranking with that model number in them and there are no other motorcycle brands in the SERPS so something changed.

Did anyone else see anything similar?

[diberry]

Did anyone notice a change in Google search traffic on October 11? None of my sites have been previously affected by Panda or any of its updates, but on Oct. 11, one of them had a 47% spike in Google search traffic. Could that be Panda, or is it something else?

ETA: Sorry I can't delete this. More research revealed that the bulk of the search spike came from traffic going to one particular page. One of the keyphrases had suddenly doubled in traffic (and I was already #1 for that phrase), so it must just be that suddenly a lot more people than usual are searching for that phrase. Must've been in the news or something.

[walkman]

ETA: Sorry I can't delete this. More research revealed that the bulk of the search spike came from traffic going to one particular page. One of the keyphrases had suddenly doubled in traffic (and I was already #1 for that phrase), so it must just be that suddenly a lot more people than usual are searching for that phrase. Must've been in the news or something.

Or maybe less ads showed up?

But as far as I can tell, Google has reverted the Sept 28th update. My best non-Pandalized site got a 20% boost to record traffic and my Pandalized site lost that boost it got on Sept 28 so all the changes equal one big, fat zero.

[diberry]

There are no ads on that term, and never have been. I guess it just got popular suddenly.

Sorry to hear about your big fat zero.