But you can't monetize a site without traffic either.

The small sites (1000 url) are not penalized, although they usually have a lot of ads, high bounce rate and bad design... the keywords ranking remains the same.

Maybe we should observe them carefully

This is correct.

My best big site with a lot of good content got pandalized.
One of my small and very crappy sites gained a lot of traffic since Panda (and again with the September 27th update).

My conclusion is that Google lost control over the serps. The algo got so messed up that they no longer know what is happening.

We were hit on the first Panda... lost about 75% of our traffic... we were on the original top 300 list of losers. Have not really been affected since in any iteration of Panda. I really feel for those that recovered only to lose again... We let go of 7 employees... I know how tough it is to deal with. Anyway about panda here is what we have done:

1. Sped up the site (still too slow... but relatively faster than we were before)
2. Got rid of paid links
3. Google No followed all content on the site that did not represent the products with enough information... (improved quality of content)
4. Google no followed about 1,000 additional niche category pages with low quality/duplicate content.
5. Added features like: reviews; comments
6. Created blogs with high quality niche information linking back to our site.
7. Added Facebook and Google +1 (BY THE WAY GOOGLE + HAS RECEIVED ZERO RESPONSE FROM OUR USERS... Facebook has received minimal response. IT'S COMPLETELY A WASTE OF TIME. AND I'D HAVE TO COME UP WITH A VERY CHEESY GIMMICK TO GET TRACTION WITH IT IN MY NICHE...I THINK THAT UNLESS YOU ARE IN A NERDY VERTICAL YOU WILL FIND IT IS IMPOSSIBLE TO NATURALLY BUILD THESE METRICS.)
8. Updating the site with new/revised content on a daily basis.

It's probably a good thing that I didn't recover... because if I had and then lost my traffic in this latest iteration I'd probably have given my 29 year old heart an attack... making my 8 month old boy fatherless.... (I'm obviously not going to sue Google for any future health problems I may come up with for dealing with their search engine :)

Overall our key terms (thousands of them) are IN GENERAL improving VERY VERY VERY VERY small bits at a time... our webmaster tools show a net positive in POSITIVES versus NEGATIVE... I do believe the POSITIVE has to do with the enhanced content on the site... and the slow recognition of some of the UGC (in conjunction with quality content) initiatives I mentioned above.

Personally I'm resolved to just doing as all respectable webmasters have bowed down to before me: Creating a site/content etc. that user want to see/use and screw looking at GAnaltyics. If the site is meaningfully better than the next best thing, it will be found. At least my heart will be beating to the sound of my own drum... rather than worrying about what the new CEO of Google may or may not do.

Having said that... if anybody has unlocked the secret code to the Panda, may you survive the next one!

Shatner :

I made some improvements after Feb 27, but stopped at the end of April. And made no further improvements since then. I recovered in June, and then was re-pandalized last week.

So I don't really think that's a factor.

I think the fact you did nothing is a major factor. You broke the threshold, time went by and you did nothing and slipped below the threshold. You waited to see if it worked and did nothing more. That's the problem.

And i suspect others that rose and fell in a hole also did nothing major since they broke Panda the first time .... but let's see what testimony turns up to challenge that.

btw - i'm sure there's false positives going on as well - this is probably a big work in progress by Google. I wish they'd break their silence and understand folks and families are hurting out there that needn't - giving them half a lead with vague directins appears as if if it's a game. It's not a good corporate look - secret or no business secret. Better guidelines are requried. Folks don't understand and they seem to know that - so why tell them half the truth knowing the implications of it's actions and the levels of many webmaster's dependency. It's not a game for many - it's their means to live.

My best big site with a lot of good content got pandalized.

chrisv1963 - Let's have some numbers , nobody will have a clue unless we get an idea of the URL / content ratios. What is the number of quality pages versus total pages that are good quality ? Approximately, what's the quantity of quality content on each of those pages? Have you really done enough to signal to Google your entire site is entirely one of quality?

It would really help if those contributing to the discussions were a bit tighter on facts, since we can't see the sites concerned, otherwise it's just a useless stab in the dark. We need more precision to have a sensible analysis. ( btw - thanks for sharing chrisv1963 - we need more like you )

Lenny2- So you didn't recover. Most folks ( on very few reports - so few are willing to share for the benefit of others ) that got out are talking major surgery to content / URL ratios , blocking pages with high bounce rates etc etc - doesn't look like you put a focus on that.

To the best of my knowledge , links are not a part of this - except "potentially" in the context of earlier posts in this thread, where you assist users to source alternate quality content sources.

Can you confirm my understanding of your situation ?

More folks need to share with specifics ( including Google ) , otherwise the learning and quality improvements will not happen and a lot of folks will go out of business. Some will enjoy that, but it's a better world when you commune decently for common benefits and trade ideas.

It shames me to see so many webmasters without a clue over nearly 8 months , accepting their fate like lambs to the slaughter. Why accept it, moan and groan - really, you deserve the best and you deserve success and Google needs to respect that to.

[edited by: Whitey at 8:09 am (utc) on Oct 3, 2011]

@Whitey,
I appreciate your work here however till now couldn't get a clue on a set of activities that might lead to a recovery and think that it is a waste of time.
May I ask you - are you a Google employee looking for evidences? I don't think there's a logic behind these Panda iterations.
Artificial things wouldn't work. I don't work for Google.

If G continues to provide awful search results (what happen right now) Bing eventually win the game. Bing is better than Google when it comes to bounce rate and time on site. Much accurate.

are you a Google employee looking for evidences to the disaster you've made

I wouldn't have the knowledge to unlock the front door at the Plex. But you guys do and so does Google - just share some basic facts.

What I have noticed, is linking may be ane of the demotion factors, in one set of results I watch the #1 has gone, linking was a factor in the site being there. In one of my own sites, I have been trying the link game, also have been demoted.

Another demotion was a site that I had included in Twitter, minimal followers, so what I would take out of that is if you put a social aspect to your site and get little followers, G perhaps think site is not popular.

The same for +1 etc, not enough +1's and your site in G's eyes may be unpopular. Best bet is not to go down the +1 or social route, at least it's not letting G see how popular or unpopular you are

The sites I have that maintained or improved their positions are sites that I have did none of the above, but sites that I have delved in the above have all been demoted.

I would like to make an observation based on our experiences with panda penalties and recoveries with several websites. It is something which appears not to have been identified by commentators, and it is this:

In formulating its algorithm for Panda, Google appears to be emulating its policies for manual penalties. That is, an algorithmic (Panda) penalty appears to be applied for an initial and specific period of time based on the nature and severity of the breach. Unlike what many commentators have suggested, the end of each penalty period bears no relationship to the date of the next Panda update.

This accords with our own experience, and it fits exactly with the explanation provided by Matt Cutts in his video at [youtube.com...] It’s worth watching twice. Carefully note the comments at 1:02, and the Freudian slip at 1:36.

It makes every bit of sense that Panda penalties would follow a template similar to that of manual penalties. After all, the concept behind Panda is to automate the penalty process, and to avoid the requirement for impossibly-large amounts of manual intervention. As we have seen with so many other aspects of Google’s processes, Google is all for and is all about algorithmic solutions.

As with a manual penalty, if the breach is maintained then the penalty period is likely to be lengthened. A (perceived) mild breach which is rectified promptly is likely to have its algorithmic penalty removed within a short period of time (as we have experienced). Manual or algorithmic, these are simply ‘time-outs’, as Matt Cutts explains.

The Panda phenomenon is still relatively very new, and while we are now seeing more of the sharp end of the initiative, the coming months (and it may take many of them) will reveal more of the pattern of Panda and the solutions to avoid it or to escape it.

damn. i was hit by this one.

two top blogs being wipeout. :( both of them are number one till last week.

its a news portal blog the contents came from other sources but i put the link source of the article.

i have 4 kinds of these blogs so far two were hit.

i dont think the issue is the content could be sitewide links.

@ seoNOOB,
From the sites I've seen affected it is much more likely to be the contents that is the problem - although sitewide links might not help your rankings I don't personally think they are part of panda.

I would guess that if most of your content comes from other sources that is much more likely to be the problem, unless you are substantially rewriting every article and adding your own opinions etc - sites that re-post news or aticles from elsewhere are pretty much the type of site that google were targetting with panda.

Welcome, Headmaster, and thanks for sharing that. If true, your insight is quite useful. One question though...

the end of each penalty period bears no relationship to the date of the next Panda update.

Why, then, do we seem to have the largest number of recovery reports at the time of each update? Yes, there have been some who have gained traffic at other times, but they seem to be the exception.

Perhaps the standard penalty times tend to be measured in months, so that the release date tends to roughly correspond to a Panda run?

Why, then, do we seem to have the largest number of recovery reports at the time of each update?

My belief is that the answer to this is sites are released in bulk because it is the measurement criteria that change with a Panda run, not that sites that are scored against those criteria at the point it is run. If I am correct, this explains why some sites can recover if they address why they do not meet the current criteria between runs of Panda, but sites (even if they have not changed) can escape when the criteria change with each run.

Im sorry, but you cant have mass winners/losers on the same date as every panda run if demotions are individual and timed.

Panda penalizes websites as a whole, not individual pages.

Sort of true, but the whole picture is a bit more nuanced. First, some pages and keywords get hit harder than others. And it is still possible for an individual page to keep a good ranking - I've seen it in a couple cases. So IMO there is a page-specific component involved in Panda.

Just looking at my Analytics Data.

Bounce rate soared on August 12th and September 27th to almost double the previous average. So much for Google giving users the results they want.

Incredibly these peaks in bounce rates occurred for the terms that bring the most traffic to my site as well as for long tail terms. It is almost as though Google has decided not just to send me less traffic but also to make sure they send me more folks who are less likely to want to buy from me.

Talking about the dates of Panda and reports of recoveries...

Why, then, do we seem to have the largest number of recovery reports at the time of each update?

I feel, without any proof of my own, that Google likely tests the Panda changes on one or several different data centers first, which is why we hear a small number of reports of recovery first.

Then after they run the data through its test phase on the limited exposure it gets from that single data center, it spreads the Panda sauce throughout all data centers, much like the other updates Google rolls out.

Much like others have seen, I noticed a large spike in GBot activity in late August (Approx Aug30 - Sept2nd) which led to a ranking change around Sept 16-18th. IMO, this was the roll out to the testing data center. Now, two weeks later, it is rolled out to others.

I would have to look through past update threads to see if the pattern was the same... but I seem to recall seeing something similar.

My sites are showing improvements. Some history of what I've done - I have about 100 sites all geo-specific. Layout is the same, content is geo-specific. Some sites have had link building, most sites had content re-writing, 2 sites i didn't do. Some of my observations:

Let's start with the 2 sites i didn't do anything with - my "control" sites - starting Thursday they started showing modest improvements - 4% and 9% increase in traffic.

Sites with new links but no new content - larger improvments - 10-15% (this is the next smallest group of sites)

The largest group of sites all had most of their thin content rewritten they are showing on average 3-5% increase in traffic

Sites with links and Content showing 8-10% increase.

The increases i'm measuring are small - This past Thursday - Saturday compared to previous Thursday to Saturday.

I know it's a short duration but it does show increases across the board for various tests. I will post more as the days go on.

The 2 sites that had no changes had no changes because they sailed through all panda updates but the last one - thats why i didn't change them however they both suffered a bit in the August update.

I also noticed a lot of my big brand searches that got hit back on the second iteration of Panda are starting to recover. We lost 90-95% of our brand searches on panda2 but i'm seeing huge (100-300%) increases in these brand searches these past few days.

Bounce rate soared on August 12th

That seems odd. Daniweb experienced the same thing on Aug 12th where she also mentioned that her bounce rate increased on that day too.

Her quote:

Then, bounce rate increased by 10% on August 12th for no apparent reason -- I did not make any changes immediately before and could find no evidence of an algorithm updating happening that day -- and has been stuck at the higher value ever since. Additionally, average time on site decreased by nearly 50% exactly on August 12th. Again, I cannot figure out what was special about that day. The fact that we were hit by Panda now is most likely a reaction to what Google discovered on that fateful day in August.

Was your bounce rate increase for that one day or did it stay there?

It is almost as though Google has decided not just to send me less traffic but also to make sure they send me more folks who are less likely to want to buy from me.

I have also experienced this at times where traffic is higher than normal but conversions are just lousy. So I agree, almost as if they are testing different markets to see how they respond to your site.

I think that all they do now is to gather "user experience" data and test what they have at the moment of a Panda "update". It's not really an update but a test with what they have so far. For every "test" they collect more data and they will keep doing that until they find a balance between their stuff and to show some sites in the serps.
I think we can try to figure it out, but if we don't know what data they collect and how it is used, SEO is dead in the water. I think we can try to change whatever we think is right, but we'll never "get it right" according to Google. It wouldn't surprise me if they read this board (Hello Google!) just to figure out what we are up to and how we can "fool" their algo. I don't think Google want us to do SEO. They want to decide for themselves who is supposed to be on the first page.

Unlike what many commentators have suggested, the end of each penalty period bears no relationship to the date of the next Panda update.

The data doesn't support this notion at all. Personally I can tell you that my panda penalty was lifted on the exact day of a Panda run. And here on this forum, every time there is a run, there are numerous reports from people having their penalty lifted on that day.

Shatner - per my above post, in response to yours , could you clarify if you did anything major after you got out of Panda.

I'm testing my belief that sites that break the threshold need to keep making improvements, otherwise they become vulnerable again and slip below the threshold. Amongst important factors I believe is the URL /Content/Freshness ratio - provided it is quality content.

First, some pages and keywords get hit harder than others.

Tedster - Any theories on why Google is hitting so selectively?

[edited by: Whitey at 12:03 am (utc) on Oct 4, 2011]

Bounce rate soared on August 12th

Google Analytics updated how visits were being calculated on August 12th
[analytics-ninja.com ]

this is really a significant change and it seems that people aren’t understanding what is going on. The main things that people seem to be complaining about are:

-Increase in visits
-Increase in bounce rate
-Decreased average time on site
-Decreased pages per visit

Many people thought that panda update was nice to them

First, some pages and keywords get hit harder than others.

Tedster - Any theories on why Google is hitting so selectively?

Here's my take on it. Panda is neither one - not just a site focused algorithm and not just a page focused algorithm either. It looks at BOTH!

A set of low quality pages can poison the rest of the site - there's a kind of decaying negative factor that spreads through the sites internal linking. Google doesn't want to send visitors to poor quality pages (no matter their relevance score) and they also don't want to send people to pages that are neighbors to those poor pages, either.

At the same time, it is possible for a page here or there to stand out as being really top quality, and they can get lifted over and above the general demotion that most of the site suffers.

there's a kind of decaying negative factor that spreads through the sites internal linking

Do you believe that added to that Google is being harder on the more competitive terms and verticals, or that it doesn't matter?

I'm thinking the relativity of the content with other sites may also play into it, since I'm seeing sites with great link profiles and content that stands OK, that have been busted. Maybe it's isolated within the site's personal poisoning as you say - the parent URL's i looked at did have about 30% poor quality pages linked to them and remained Panda'd. But those that surfaced in the last update were on child URL's where there wasn't a problem with the other child URL's and never produced much traffic anyway. Maybe the scoring also includes some consideration to folders - but I'd be careful playing with that, get it wrong and the whole site may be effected as you suggest. Just get the entire sites fixed and humming IMO

There has also been some suggestions by others that duplicate content overlapping from parent to child may cause a problem - but again i see some sites holding strong with that. It makes me think this algo has a long way to run before it's "stable".

@Rasputin well could be the contents came from other sources but these 4 blogs been running for more than 2 years now. Though only two got hit.
I think i will revise it by adding my opinion on each content. :P

BTW my flagship supersite is gaining back the traffic after being pandalized which I lost -60% traffic.

Since last Saturday it increased by 110% every day.

Is it possible that Panda does not like sites which link out to many other sites as an editorial service to the user?

honestman - FYI, just the opposite was being assumed on another thread at the end of August...

From around the web
http://www.webmasterworld.com/google/4356511.htm [webmasterworld.com]

I noticed that some large websites have begun linking to other sites in a "from around the web" section of each page.... Has this become an optimization technique webmasters should embrace? Are search engines rewarding sites who interlink to other sites?

For the most part, I fall on the side of thinking that well-chosen outbound links are more likely to help than hurt.

You also mention that some of the links are organically reciprocal, and then go on to say that you do a lot of internal cross-linking. I'm just wondering whether you might have gone overboard on your linking in general, particularly on the cross-linking.

There's no way to generalize about what a good quantity is, except perhaps to remind you that you probably can't do anything like the amount of outbound linking or contextual linking that, say, Wikipedia does. Also, there are places where one-way navigation linking makes a lot more sense than cross-linking.

From Google's viewpoint (as best as I can get a handle on it) the idea behind Panda is to give an improved experience to the user's of their SERPs. So there sure could be a kind of scaling according to how competitive the niche is - we know pretty well that Panda only gradually got around to the long tail in later iterations.

Now does that translate into something like "comparative quality" within a niche? I've got no data to judge from, but it sure might be the case. And I've got to agree with your other observation, link profiles alone sure don't give any immunity from the wrath of the Panda.

Now does that translate into something like "comparative quality" within a niche?

This would logically have to be true. But it appears Google is acting illogically, which makes it so difficult to second guess.

The websites in my niche are download sites (galleries). 99% of them don't bother to add any textual content and still manage to receive masses of traffic. The largest receiving 400,000 visitors per day and has no textual content on most of their pages.

I've spent tens of hours, writing textual content for the pages and to absolutely no effect. Zero.

writing textual content for the pages and to absolutely no effect. Zero

Did you/they have image tags? What's the difference between you ?

there's a kind of decaying negative factor that spreads through the sites internal linking

I'm speculating that another way to look at this is that low-quality pages anywhere on a site statistically reduce the likelihood of user-satisfaction within the site. Similarly, inappropriate navigation choices statistically reduce the likelihood of user satisfaction.

Chances are that you get more positive points for user satisfaction within your site, but you at least avoid negative points if you don't send a user back to Google to run the search again.