Google Algo Change - Human Behavior Theories

[browsee]

Just checked several losers from Feb 24th and Apr 11th, sites with dark/shady colors are affected. This is in line with their statement in Wired.

"We used our standard evaluation system that we've developed, where we basically sent out documents to outside testers. Then we asked the raters questions like: "Would you be comfortable giving this site your credit card? Would you be comfortable giving medicine prescribed by this site to your kids?"

So they tried to convert user behavior into an algorithm. I've been trying to understand how we can correlate human behavior to our website design.

Based from my research I've found that there is a relation between the Google algo change and Maslow's 'hierarchy of needs' theory in psychology.

I found this article describing Human Behavioral Theories That Can Be Applied to Web Design
[sixrevisions.com...]

Important factors in order of importance.

Accessibility: The website can be found and used by all people.
Stability: The website is consistent and trustworthy.
Usability: The website is user-friendly.
Reliability: The website is consistently available, without downtime.
Functionality: The website offers content, tools and services users value.
Flexibility: The website adapts to needs and wants of users.

Most of the Google engineers are PHDs, they tend to follow academics rather than reality. I am very confident that they included some of these behavioral theories in this algo update.

[walkman]

Tedster, I meant for the plex people, not those in temp jobs doing ratings

[whatson]

My site that got affected is all black background, with a lot of white text. It was partly to do with the fact it was for a nightlife theme, so black bg was fitting. However, I have another site without a black bg, and black text on white, that has not been affected whatsoever.
I just find it hard (unfair) to believe that such a penalty would be dealt. Anyway, I am going to re-do the colors, and will post my results.

[Dan01]

Facts, we don't have any besides an admitted statement that Google is trying an algorithmic approach to mimic human judgment of trust.

Yes. We all seem to be trying to backward-engineer what Panda did. Besides, trust, they are looking for site quality. At can be very difficult for a computer to determine. They must have made a breakthrough a few months ago, but like you, I want to know more of the FACTS (the signals, criteria, ratios, etc). Open facts like that would improve the quality of the web in general.

[GuyFromKlingon]

Larry Peige says he wants Google back where it started. Let's help him out, just add this to your robots.txt.

User-Agent: Googlebot
Disallow: /

[TheMadScientist]

..most who haven't been hit are not posting on the subject..or have given up posting on it

That's exactly why I didn't hop in at the beginning of the thread.

I think a good look at the SERPs will show the same as my working knowledge of / history with sites NOT hit, which refute every point in the original post ... I can go right down the original list, type a query, see one of the sites I'm referring to still in the top 5 and gaining and KNOW the OP sounds good, but it's not the answer.

[Sorry if I sound like I'm being harsh Browsee, but I really did that.]

One of the interesting things about people not recovering, imo, is an SEO would try to reverse engineer the changes by making one at a time, but I doubt anyone except an SEO or someone 'trying to game' the system would do that ... Also, it's not one 'factor' that changed, it's multiple factors, so if you change one at a time, you will still not 'pin down' the changes, you will have a series of changes you personally made and which ones actually made a difference may not be inclusive of all the changes ... If you make 10 changes over time, it could easily be 1, 3, 8 that made a difference when combined (especially if there's a delay involved) with the others having no effect.

Personally, on this one, I would not recommend 'one change at a time' for recovery. I would make larger scale changes that are 'new site' or 'new focus' oriented, rather than the micro 'figure it out' changes normally recommended.

[heisje]

Yandex, Baidu and Bing are not serious Google rivals for most of the world.

Be assured that Yandex, Baidu and Bing are striving to change this state of affairs, and they may even succeed at that. They also seem to have a lot of webmaster support in this endeavour - something that won't hurt either.


.

[katmondu]

And I've noticed a lot of people saying they are getting kind of bizarre rankings lately, I think that is the random testing going on before things shake out.

Whatever is happening, it better go away soon. I searched < search terms removed > because I'm having an issue with my air handler and i got an entire page full of gross results related to pregnancy, abortion and/or sex.

By the way - since the panda update the best results I've found for the queries I make have been from ehow.com - I used to hate those results, and now I look forward to them because they are at least the most relevant in the sea of crap Google has placed before me.

[edited by: tedster at 12:34 am (utc) on Apr 18, 2011]
[edit reason] removed specific search terms [/edit]

[walkman]

katmondu searche for < search terms removed > and got "1]< search terms removed >[/1]"

Hahahha. Content is King. Panda is the Prince

[edited by: tedster at 12:36 am (utc) on Apr 18, 2011]

[tedster]

We have many threads here generally discussing and criticizing Panda. This thread is intended to focus specifically on one aspect: "Human Behavior Theories." So please everyone, let's hold to that theme.

[wyweb]

Um... noob question here. Please be gentle.

When did Panda actually roll out? Are some sites just now getting hit by this?

There's a reason for asking.

I just took a major hit but I'm not thinking this is Panda. I'm thinking it's this direct advertising thing I've got going.

We're talking about 50% reduction across the entire network. And since the entire network isn't involved in direct ads, I'm wondering if this is some sort of delayed Panda crap.

[driller41]

I did not realise how advert heavy some of these sites are until I viewed them through IE.

I had been using Firefox with AddBlockPlus installed when scanning this thread so I got a totally unreal picture.

When I viewed the same sites with IE the number of adds became obvious, even Ehow is plastered up like a tart on the town.

[netmeg]

I think the main dates for Panda hits were Feb 24 and April 12, with incremental shifts on other dates.

[wyweb]

4/12 fits perfectly.

Man oh man....

[kartiksh]

"We used our standard evaluation system that we've developed, where we basically sent out documents to outside testers. Then we asked the raters questions like: "Would you be comfortable giving this site your credit card? Would you be comfortable giving medicine prescribed by this site to your kids?"

Is this the reason, my site of about 10 years old now getting hit badly? over the last 10 years the traffic remain steady to this site as well as the original content keep on increasing. Now that after panda update, first time in 10 years the site is having 50% of visits compared to all these 10 years average and this is consistent. Is this because most of those who may have tested said "they will not give credit card information to my site?" as that is something obvious for my site. However it has lot of trust from .edu, org and gov since its inception. The color and text are very much inline of what is expected by Panda.

I am unsure whether this common human behaviour or judgement is good for sites of my kind. I am experimenting, and will post here if thing changes to the noticeable level.

[kartiksh]

A small finding which may have impacted my site traffic : I searched deep inside analytics and then to found that there were good number of visitors coming from eHow and ongoing through those pages I found they were using “rewrite” of our original content and then gave us credit at bottom as reference. Going thorough other threads I noted they were hit and so some of my visitors were lost which were “referenced” to my content. :-)

[koan]

over the last 10 years the traffic remain steady to this site as well as the original content keep on increasing. Now that after panda update, first time in 10 years the site is having 50% of visits compared to all these 10 years average and this is consistent.

Yup, welcome to Panda hell. Read all you can. Time for a site redesign and hope for the best.

[Bewenched]

I found they were using “rewrite” of our original content and then gave us credit at bottom as reference.

They have done it to our stuff as well even though we have a clearly defined copyright statement.

Oh and the links to you from ehow have a rel=nofollow so you get NO link credit either.

[Meenakshi Mosaic]

Can someone brief on the analysis came after this update?

[Meenakshi Mosaic]

Thanks crobb305. it helped to find out a direction to the overall conversation.

[tedster]

Everyone I know is still involved in the analysis, even after four months. So there's nothing either brief or conclusive to share. I think there's no option but to read the many threads here about the Panda Update.

[potentialgeek]

When there is so little progress after four months by a large community of webmasters, you have to wonder if Panda represents a paradigm shift.

Google has said they used their standard procedure for quality control at the beginning, but Mr. Panda is a different animal. He, as was noted earlier, is a Decision Tree expert.

Cutts has said a lot of data is used to calculate the Panda Rank - so much so that they only do the updates occasionally - like with PageRank. (Think back to the monthly Google Dances of yesteryear.)

Panda could be the first time the Decision Tree Paradigm is used significantly in the Google algo.

The tragic irony is that Panda is supposedly trying to think like humans, but none of us can figure it out yet.

How do we think like a machine trying to think like a human?

"Panda is trying to mimic our understanding of whether a site is a good experience or not. Usability can be a key part of that. They haven’t written code to detect the usability of a site, but if you make a site more usable, that’s good to do anyway." --Matt Cutts

[edited by: potentialgeek at 8:04 pm (utc) on Jun 28, 2011]

[freejung]

How do we think like a machine trying to think like a human?

The simple solution to that would be to simply think like a human, and hope that the machine gets better at it.

[dickbaker]

The simple solution to that would be to simply think like a human, and hope that the machine gets better at it.

I hope the machine gets better than the voice recognition systems used on so many phone systems.

[freejung]

I hope the machine gets better than the voice recognition systems used on so many phone systems.

Well, the machine is quite a bit bigger, but then so is the voice it's listening to. So far the signal appears somewhat garbled.

[Errioxa]

I think Panda uses algorithms that finds patterns in the sites scored low.

Panda is using the human to learn

[potentialgeek]

After looking at various rankings for my sites and competitors, I'm getting the sense that the Decision Tree is anchoring off the first part of the search phrase (more so than usual).

I'm seeing sites with higher ranking for "widgets" get simultaneous higher ranking for "widgets red" as well as "widgets blue" etc.

I see this on my own site after Panda 2.2 when I recovered from Panda 2.0. I had seen something very similar on competitor sites after Panda 2.0 (they didn't get Pandalized).

It seems as if the same type of engineering for the brand-detection algo is being applied to Panda. The brand-detection algo was added a year or two ago and it tries to determine when users are seeking a brand's main website.

At first I thought Panda was assuming various sites which went higher in rankings after the Panda Update were brands. But now it seems more likely Google simply applied the same type of engineering for brand detection to Panda.

I know for example a search word which is for a generic thing. There is no brand in the world with that word only.

The other possibility is that Google is now using the brand filter more liberally, i.e., the concept but not in the commercial sense, like Coke or Pepsi.

I've seen sites that are top 3 in rankings for keyword X now, since Panda, get top 3 rankings for just about every related and similar search with keyword X. The word isn't branded and the top 2 sites aren't brands, either. It's as if Google assumes they are brands.

There's either a flaw in the brand-detection filter, or the idea is being applied as a "brand-like filter."

Remember a few years back, when, if you had great ranking for keyword X, you could get top/great ranking for so many different searches with that one word? Then they changed the algo, so that didn't happen? It feels like a reversion.

I've always wondered about the brand-detection filter since it was first introduced. How does Google decide what's a brand and what isn't? Unless it has a list of brands it entered manually, it must be done by algo.

What are the official brand signals? Which are unofficial? Which are clear? Which are noisy?

[tedster]

I think "brand detection" is a kind of semantic processing, first and most of all - possibly needing some reinforcing signals for a total positive. If I'm right, then occasionally a non-brand phrase might get a false "brand" identification.

[HuskyPup]

a non-brand phrase might get a false "brand" identification.

Yep, hence the rise of directory sites especially so within my industry.

Google's assuming that 10-20 listings from several suppliers for a requested product = authority, when, in fact, for probably 99.99% of searchers they are totally useless since they only apply to quantity, wholesale purchasers.

I have seen this massive shift on some of my sites where, for many products, I used to rank in the top 3 to it-doesn't-matter-land but at the same time the products for which I am world-leading have held firm.