Minor Panda update Oct 13 - per Matt Cutts

[sid786]

Matt Cutts responds on Twitter that they have applied a minor algorithmic update last late night.

Link: http://twitter.com/#!/mattcutts/status/124905069748559872 [twitter.com]

Have your websites been affected with this update? My site's traffic is stable, but I was expecting a positive bump.

[Mod's note: Fixed link so it would display, as the WebmasterWorld link redirect script will break it in most browsers. Copy and paste url into your browser if hyperlink doesn't go to Matt's tweet.]

[edited by: Robert_Charlton at 9:54 pm (utc) on Oct 14, 2011]
[edit reason] fixed link display [/edit]

[netmeg]

i summarize a lot of stuff about a specific topic, and i make it very easy , and fast to find the content people looking for...

My most popular sites do the exact same thing and have for years.

[Leosghost]

I'd bet that you don't "scrape" or use other peoples content without their advance permission though..:)

[kidekat]

I seems that 'big' brand ecommerce sites haven't been affected ie the company I work for had no idea this stuff had happened over the weekend.

[Whitey]

What I'm hearing and seeing is that the traffic released is not necessarily on quality keywords, more 2nd rate terms.

This is a very small sample and somewhat speculative. Is there anyone with analysis to support or reject these evolving rumour theories?

[Whitey]

This has been blindingly obvious from the start of Panda.

Anyone who has not been thinking that has not been paying attention.

@Leoghost - I'm a bit slow on things sometimes - but what we don't know at this stage is what the quality of traffic for Panda recovery folks is. It's still early days on the specifics within Panda - but we have a chance to understand if folks will share.

We need more analysis on keyword recovery to draw a better picture for the Panda / quality / keyword / traffic throttle - if indeed it exists, which i think it does.

Let's have some inputs - what are you seeing out there - it's hard to run blind.

[achean]

I've got a large, high traffic site, that was hit to the tune of 20% by Google in the latest update (originally hit by P1, etc). But, according to GA, I'm also seeing a meaningful uptick in direct traffic. Not quite enough to offset the loss completely, but substantial. There could be other factors at play, so I'm curious if anyone else is seeing a similar pattern?

[Whitey]

@achean - Glad to see you have an uptick - good news for you, let's hope it holdsand improves.

We need more analytical inputs to provide meaningful discussion though. I assume you believe it may have been a Panda release,

.... but what makes you think it wasn't?

What have you done?
What have you not done?
What do you intend to do ?

And maybe you'd have a look at your keyword analytics from Google and indicate how that compares pre Panda.

We need more data if you're willing to share.

[Leosghost]

Think about what "decision tree" means..

You won't get much ( if any ) traffic for keywords that panda has decide don't fit what it thinks your site is..and how it decides is based upon many factors, both signals within your pages and in the pages of who you are liked to and who you link to.

This is why a group of pages can affect how your whole site ranks for keywords that you think it should ( and panda has decided that it should not )..and panda is learning by re-running the cross related data at intervals..and mixing it with the usual algo variants..

All this has been "blindingly" obvious from the start ..to any one who read either Mr Pandas work ..or the writings of some such as SEObythe sea..But almost without exception posters here and elsewhere got hung up on what was affecting their "ranking" based upon solely what was on their site.." how come I have all these backlinks and I still saw my rankings for "widget" fall"..

Or they got hung up on the guys name and their association with the word and thus spent their time posting and reading smart ass comments about "not very cuddly" etc..

Or deciding to rant about their perception of push to adwords as a forced way out ( because they couldn't see what criteria Panda was using ..or didn't want to see ..they were "entitled" to traffic as they had always had it ) and therefore that must be why in their opinion the addition known as Panda was run out.

He is a programmer..as most at the plex are ..many are also specialised mathematicians..and the plex loves automation..

Panda is attempting to automate by machine learning to replicate the "decisions" that humans made about what sites are "about"..into which "categories" they "fit"..

So sites that do not give very clear signals ..such as sites that sell product ..but also run adsense or ads..are not clear to users ..not the ones who Google used to test and set the criteria ..pre panda ..so they lost traffic..any site whose purpose and thus category or "brand" was not obvious was affected..and whether the site was set up with the intention of selling a product ..if many of its pages are running ads ..it gets put in the "ad site" category ..and maybe in the "MFA" category ..and gets dinged by another part of the concurrently running other algo ..

Site "topic" or "theme" or raison d'ętre..needs to be real clear ..so clear that it could be taken in at a glance..that has appeared to be the way that even manual appraisal has been done for years with Google..quick glance ..you are in ..or out ..account "shut" or re-instated..spam ..or not..ecommerce ..or info..first, fast impressions always counted for Google's "customer support"..or "assessors"..they tag 6 pages from a site as spammy ..your site looks like that site to panda ..it calls you a spammy site..'til it learns different..or maybe it got it right..

Its machine learning ..and sometimes just like any machine learning it will veer off at tangents for a while ..they get too "off target" , it gets tweaked back..so it knows that "ford" probably means you are looking for a "car" or a "car manufacturer" and not "how to cross a river without using a bridge"..

And Google put a lot of store in "smarts"..remember when they hired people by putting mathematical puzzles on bill boards..you got it right you might get hired ..they didn't concern themselves with how many wrecks they may have caused because people were distracted..because the "smart ones" could figure it out instantly and keep driving safely, nor did they worry about if the boards were standing on pylons over the top of some homeless person's cardboard box home..

They wanted "smart"..that is what counts to Google..( and "fast thinkers".. because they apparently think that Larry perceives things like pages loading unusually fast ..he doesn't..two hundred milliseconds is nothing special..unless the rest are "stoners" )..thousands of sites might wither and die due to Panda? ..

So..there will be more to replace them..and they will ( eventually because it has to learn ..and Google take a long view..so "learning " might take a few years even ) be more targeted , more focused, and faster ( apparently no-one at the plex has the "balls" to tell Larry that it is his adsense and or analytics code that slows down many pages ..or maybe they do and he says ..well they'll ( the webmasters ) have to figure out a way to make the rest of the pages faster to make up for it ..or that newer CPUs will take care of it ..it is all javascript ;-) so mainly "bottlenecks" are put down to "client side" )..but more sites and more pages will come along to replace those that wither through lack of traffic..

That is like "evolution" for the web ..survival of the fittest..the dominant in their "niche" ..not the blackest, nor the spammiest..although both of those may from time to time benefit from the combination of panda and the other algo variables as the machine "learns"..meanwhile ..why should they care if some don't survive..and why should they cry if some sites decide to try to keep their place in the sun and their traffic via adwords..

Even when they compete in verticals with their own products ..in their thoughts and philosophy they are convinced, they are sure they can do it better, ( might be the case ..or not )..and the revenue can only make them stronger, and thus better equipped to do what they want to do..you, as adults "know" what is best for your children, what is best for kids in general, in the long run..because you are brighter and consider yourselves to know more than your kids..you are more intelligent than your pets..you have learned complex things like websites and economics and marketing and maybe even a little code..

You feel , nay ""know that you are smarter than many of the people who you interact with daily..on and off the web ..and those who visit your sites..

The people at Google ( and if the truth be known, at the other search engines and IT companies ) "feel" and "know" that feeling of knowing more and better and what is best ..in "spades"..

They will only "bend" a little, if they think either they have miscalculated or that you are as smart or smarter than they..complaining is not going to be taken as a sign of either..

And they may well be right..and their "way" will probably prevail..

There is plenty of "data" already available there in front of you ( more is not needed )..you need to look at it properly..from the point of view of the programmers ( Mr Panda included ) in the plex..it, "panda" is doing more or less what it was intended to do ..and learning as it goes..and deciding..always deciding..

[edited by: Leosghost at 3:54 am (utc) on Oct 18, 2011]

[achean]

@Whitey. Sorry, I should have been more clear. Overall, it was a downtick, and it corresponded with the latest Panda update on the 13th, with losses concentrated in traffic to long tail terms. So, I'm sure it's a Panda issue.

I posted because, in addition to the 20% loss in traffic referred by Google, I'm seeing a meaningful increase in 'Direct' navigation traffic according to Google Analytics. That's surprising, so I am curious to know if anyone else is seeing a similar pattern.

[tedster]

Panda is attempting to automate by machine learning to replicate the "decisions" that humans made about what sites are "about"..into which "categories" they "fit"..

Google has automated the creation and assignment of taxonomies for lot longer than Panda has been around. Why do you think it's also part of Panda?

[walkman]

Or deciding to rant about their perception of push to adwords as a forced way out ( because they couldn't see what criteria Panda was using ..or didn't want to see ..they were "entitled" to traffic as they had always had it ) and therefore that must be why in their opinion the addition known as Panda was run out.

There is plenty of "data" already available there in front of you ( more is not needed )..you need to look at it properly..from the point of view of the programmers ( Mr Panda included ) in the plex..it, "panda" is doing more or less what it was intended to do ..and learning as it goes..and deciding..always deciding..

Leosghost,
I am not going to comment on whether people that lost 60-90% of their traffic overnight with no warning or apparent way out are whiners or not. Or if Panda is fair, ready to go live or not, considering Google's market share.

But if I knew what you say you know, I would maybe give hint to a few people here and then head to the largest web consulting firms and tell them how to beat Panda. It's a multimillion idea, and people would gladly pay you. Don't feel bad, capitalize on it, you'll still find some time to try to teach us here. Everyone has ideas, but very few knew it exactly.

_{Personally I have seen 3 of my sites go from "Low quality" to "high quality" to "low quality" without me touching a damn thing in them and I can't get one out of Panda despite a gazillion changes and improvement in any metric I can think of.}

And Google put a lot of store in "smarts"..remember when they hired people by putting mathematical puzzles on bill boards..you got it right you might get hired ..they didn't concern themselves with how many wrecks they may have caused because people were distracted..because the "smart ones" could figure it out instantly and keep driving safely, nor did they worry about if the boards were standing on pylons over the top of some homeless person's cardboard box home..

Please stop perpetuating this myth, they are smart no doubt, but Microsoft research wipes the floor with them, and with many others. [academic.research.microsoft.com...] Unless you have an objective rank that shows that Googlers are smarter than other engineers at MSFT, Apple, Oracle or IBM for example. Try the citations and see. You can also filter by last 5 years, last 10 years and see. Granted MSFT and IBM work on many fields but it's not even close.

[Whitey]

@achean - that's cool - but as i see it you have 2 issues, you're obviously pandalized, and your're trying to work out why more direct traffic is coming to you.

Firstly, i guess you want to get out of the Panda fix - so as per above, what have you done / not done etc? You're here for a reason - Yes / No ?

Secondly, what do you mean "direct traffic" in terms of it's relationship to Panda. I'm assuming that the proportion of direct traffic would have increased, because of the Panda hit, but can't see ant reason why the no's would have gone up. Have a look at your analytics - what do you see when you dig deep ?

Is this really Panda related?

[disspy]

I think the terminology “Pandalized” shouldn’t be used any more. It seems, that the most of websites are affected with Panda somehow. Some in positive and some in negative spotlight. The negative connotation usually strikes first, so we always read some howling around the forums, and other who rub their palms just lurk and keep their master-ideas for themselves. However, this is the forum dedicated for people to share experience, and I decided to take more time for a longer post.

The fact is, the Panda became now the main part of Google’s SERPs algo which seems like Page Rank replacement, (Panda Rank, although still the same acronym).

From the webmaster’s point of view, it’s a difficult time, no doubt, but the actual idea of bringing the more quality to the web is not that bad. Google’s monopolistic position allows them to play gods, which would probably everyone do if we were in their (Larry/Page) position. They are in definition scrapers too, but in comparison to other scrapers, the most content-driven websites depend on them and their monetizing service as well. Only in case that ALL webmasters would boycott and isolate Google from crawling their websites, Google would probably fall down, but in 5- 10 years since they already own some of most visited websites too :) Do you guys still remember Alta Vista days or days without search engines? Some of you probably remember GeoCities or Lycos, but these times are over. Only the strongest survived. That’s a normal and natural selection. Nowadays we play the same survivor game, but with more spices and ingredients then ever before. Today the content-driven websites depend mainly on search engines (read Google). Although, you still have social web and other (white/blackhat) methods to bring visitors to your site, which is not that easy except your niche or site thematic is viral by itself.

Anyways, I read people share their efforts to get out of the Panda SandBox, so I just want to put my $0.02. Our high-traffic mainstream website “gone with the Panda” through couple of iterations. The agony started with Panda 2.3 – traffic down 70%, then one month later we recovered almost 80%. A month after, Panda 2.5 took away another 60% of traffic. Last Panda update (I don’t even take care about Panda versions/dates anymore) brought us back. Traffic increased for 30% and still increasing. Our mainstream SERP’s keyphrases came back and it seems we are on good path to get back (hopefully) to pre-Panda state. What we did? Well, we ara an agile team of 20+ people and we did a lot stuff during last 4 months in order to improve user experience and site visibility. First we asked ourselves, what we did wrong? We closely inspected and identified a lot of parts which have to be improved. Earlier, we put more attention on SEO and Google, and less on actual user experience. The complete SEO industry did the same thing. Now, Panda is just doing the opposite and SEO industry should (or must) follow Google’s path. Here is the short list of things we did and we still work on.

1.Improved long-click
2.Improved long-click again & again
3.Improved UX (design realignment –about 50 different views, new navigation, more user friendly interfaces, faster responses)
4.URL reconstruction (new URL map)
5.Subdominization (moving directories and low-quality content on subdomains)
6.Low-quality content decapitation (killed all low-quality sections/content)
7.Uniqueness improvement (duplicate content removal, scraped content removal)
8.Improved content authorization
9.Related links optimization
10.Server speed optimization (moved to new faster servers, bandwidth upgraded, more memory, network balancing reconstructed, reduced downtimes to minimum)
11.Site speed optimization (img/css/js minification, HTML optimization, cache optimization)
12.Mobile version optimization (faster page load, less JS, new easy-to-use template)
13.Content optimization (enrichment of authority content, UGC quality levelization)
14.Introduced A/B & Multivariate testing

Personally, I don’t believe in miracles, and still believe that all the hard work in right direction would paid off. In other words, for every disciplined effort there is a multiple reward for sure ;)

[chalkywhite]

For anyone hping to escape soon...read into this what you will but its not an automatic process, its as and when they can be bothered to run the next update..

[webmasterworld.com...]

[edited by: tedster at 4:43 pm (utc) on Oct 18, 2011]

[shankarmd]

My Site is Indexing well. But not showing in SERP from Oct 14th. all other scraper sites are getting first page result. So what is the Final Solution?

[Whitey]

However, this is the forum dedicated for people to share experience, and I decided to take more time for a longer post.

Glad to see you support the sharing experience. We need to encourage more to participate with data and facts that help each other.

Really liked your statement about the power of the user experience. Too many folks are chasing the algo, not enough said about what a site can do to improve the user experience. A great checklist and positive post to another set of steps in challenging times. Good luck with further improvements. Hope to hear you achieve 200% improvement on pre Panda :)

[snickles121]

I would wait at least a week before you do anything or panic. To me it seems like they hit a reset button and it will take time for the serps to even out again.

Note: That is what it looks like from my data, but who knows...

[snickles121]

This weekend made me nervous with low traffic, and I and my partner were brainstorming. I dont really want to point out tactical moves but this site and its members helped me out of a jam a few years ago along with tedster (big help).

So...our moves after 36 hrs of thinking this weekend was to get our site listed with google local, yahoo and bing (its a free business listing (they dont approve you until they send a pin number to your house), also you can hide your home address so you only have a business listing with a phone number. Note: we do run information sites but google say its ok to do this according to their terms. Our theory was that if google wants branding maybe it will help.

Second, we applied for a BBB listing. Hey it cant hurt.

Third, our linking structure from outside sources from tier 2 and 3 pages took a hit. So we made a plan in place to try to recover new links in that area. We use to have plenty of internal outside links but after panda it depleted. I might also point out is our main links were from free directories and so on.

Next, we applied for a yellow pages listing, hey you just need a credit card and phone number. Not like I cant screen my calls. And again branding.

Lastly, we issued a pr. It was just a free pr but it did show up it serps right away. Again trying to brand us.

Well, I hope this helps you guys, because I cant afford anymore hits in traffic. I will post back if any of this helps.

[rj87uk]

This is the literally first update in a long, long time that has had a negative effect on my websites which is currently around -35% traffic from Friday. I know its early into this update but thought I should be proactive gathering information.

From what I can gather Panda is to do with on page quality signals rather than SEO (keywords, links and such) am I correct in assuming that so far?

I am sure there will be threads here to do with website quality signals so does anyone have any favourites that you recommend me to read?

Lastly from reading these threads alone it looks like several people who lost rankings during the first or second Panda update has seen their websites recover if only slightly in this 'Panda 2.5' update and to get some more data can anyone who lost out on the first two updates of Panda post if they recovered some traffic in this latest update.

RJ

[snickles121]

I have to disagree, I think its linking structure. I also lost a good amount of traffic this update and all we tried do with our website over the summer was improve content. The only thing we didnt do was try to increase inbound quality links.

PS: What type of traffic are these people who recovered have. If you have 50 hits a day and got ten more so what. Were dealing with hi traffic numbers. I want to hear from the guy who has at least 1000 visitors a day. You cant run data at low numbers.

[disspy]

Quote snickles121:

I would wait at least a week before you do anything or panic. To me it seems like they hit a reset button and it will take time for the serps to even out again.

Optimizing a website for better user experience is not a panic attack. It should be one of the most important strategic goals for every webmaster.

Quote Whitey:

Really liked your statement about the power of the user experience. Too many folks are chasing the algo, not enough said about what a site can do to improve the user experience. A great checklist and positive post to another set of steps in challenging times. Good luck with further improvements. Hope to hear you achieve 200% improvement on pre Panda :)

Thanks for encouragement. Actually, we are on the way to complete the mission which is to deliver best possible user experience, with or without Panda. This doesn't include only layout but also the quality of content.

Quote snickles121:

What type of traffic are these people who recovered have. If you have 50 hits a day and got ten more so what. Were dealing with hi traffic numbers. I want to hear from the guy who has at least 1000 visitors a day. You cant run data at low numbers.

You think 1000 visits is a high-traffic website? Personally, I don't consider a high-traffic website under 500k a day!

[snickles121]

I have pages that was written and unique and not found on the net and had to go talk to people(###) in person to even get it. User experience is a joke, its all about money somewhere.

[rowtc2]

Here is an interesting article about how Google algo can work and Quality Raters Guidelines 2011 - article has a link to a .pdf page. I find it useful, in directions to look for improving a site. [potpiegirl.com ]

Panda may incorporate in a math algo the human rater's thinking.

[snickles121]

I meant no disrespect on numbers. Im just talking about accurate data. Im just trying to share accurate info. Trust me, I have kids to feed too and I feel the pain.

[netmeg]

User experience is not a joke.

[Whitey]

Trust me, I have kids to feed too and I feel the pain.

This is desperate stuff .... i see it all over the threads in people's dilema ... it's a great shame that Google doesn't improve it communication to assist, but there is a great community here, and some good people are lending their thoughts to help others ..... but you have to accept some responsibility to change your thinking, or you'll effect those you're trying to help. So this is not lightweight or selfish communication. It's important and folks with knowledge should join in more ...

On this very page ( #:4375846 ) you have some of the answers , but you and thousands are so very much up against it, that you don't see the wood for the trees and resist in denial. You are not alone, but as i say , you must be quick if you want to stay ahead.

Again, too many folks are chasing the algo with a fixed way of being and persitant complaints about Google - forget it it, it is what it is and not what you want - that will hold you back. Some folks have basked in the success of 10 or so years of manipulative techniques or cherished strongholds that make them vulnerable to change -it's fast becoming history. If you would only put your efforts into seeing what others see, you and thousands of others will be building better websites that do better on Google. Forget Google - Google will follow you.

Open your eyes and ears, remove your fears and look on some of the posts above. There's a great one not more than 4 or 5 back from this one - guess where ( #:4375846 ). And many more. On top of this get others to critique your site and suggest how it could become great in it's niche - that's brand , real brand.

Panda is a baby as Tedster puts it. The responses to it are similar - but there is a new opportunity, just let go of the past and step forward. User experience is very very important. It's definitely not a joke.

Thanks for sharing ... you came forward with your views .. hope what you read and hear with different eyes is now helpful.

Good luck.

We have to set the scene for a positive approach to the challenges. Those involved will spark new ideas.

[edited by: Whitey at 1:19 pm (utc) on Oct 18, 2011]

[Play_Bach]

huh?

[rlange]

Leosghost wrote:
[...] and panda is learning by re-running the cross related data at intervals [...]

Is it, though? If this we true, we'd expect to see gains and losses outside of manual algorithm tweaks by Google. I'm not sure there's been any evidence of that. There have been a small number of reports of changes outside of the typical Panda update windows, but that could easily be the result of "regular" algorithm updates.

--
Ryan

[Bill_H]

Tedster, Whitey:
I suspect that Google might have "adjusted" the Oct. 13th algo over the weekend and started to roll out the changes Monday AM.. We are seeing traffic returning to pre 13th levels across all our sites. It started Monday morning to pick up and just kept growing. Perhaps they have stopped throttling, who knows.

We ran our top 100 keyword positions on Monday and see virtually no change (always some slight up and downs, nothing out of the ordinary flux) from the previous Monday. Must have been some throttling.

Cheers,
Bill

[DirigoDev]

@Whitey - I agree 100% with your post above. Stop thinking about the engine and start thinking about the user. What real value do you bring to the equation? I once thought that original content was words arranged in a unique way written by a good copy editor. Not anymore. How can I build the most valuable article in the world on Topic x? This is how I’m thinking!

I’ve been reading “About Face 3: The Essentials of Interaction Design” by Alan Cooper. Also recently reread “The Inmates are running the Asylum.” Add value! Follow these texts and you’ll find yourself free from the clutches of the almighty Panda.

@snickles121 – Some local, PR, yellow page listing, BBB, and new external links are not going to fix your problem. This is not the recipe for a Panda fix. This stuff is foundation tactics.

IMO, the recipe can be found in answering these basic questions:

What is your value proposition?
How are you different?
How are you adding value to the WWW?
Why are people bouncing from your site?
How can you build a better Website?
What do your visitors want?
Why should users like or +1 your page?
Why should users contribute with Facebook comments or Disqus?
Does your site have authority?
Do visitors trust the information on your site?
How do different personas interact with your site?

One of my sites has recovered fully and if it holds, I’ll be rewarded with 500,000 additional SEO session starts this month. We put thousands of tech hours into our Panda fix on a site with ~2,400 pages. We have ~50 full time staff. We don’t know why or what worked. I have some ideas but I’m not going to share tactics because I’m not sure that they’re important. Everyone is looking for the magic bullet. I'm not sure that we're going to find one.

Like someone else said earlier, it would be hard to point out things that we didn’t change on our site. We deployed hundreds of changes – enough that I could write a book on our response and the senior management team’s evolution of thinking as it became clear that we were not going to recover soon or even at all.

I’d like to hear more high level discussion about what Panda responses/approaches worked.

As of 10:30 EST our traffic pattern is holding steady per the 10/14 update.

[edited by: DirigoDev at 3:13 pm (utc) on Oct 18, 2011]