Panda 2.2 Update Part 2

[PPC_Chris]

The following was continued from http://www.webmasterworld.com/google/4326253.htm [webmasterworld.com]
According to Search Engine Land [searchengineland.com...]

Google has given us confirmation that they have ran an update to the Panda filter recently.

We have been expecting the Panda 2.2 update based on news coming out of the SMX Advanced conference. Matt Cutts told SMX attendees that Panda 2.2 has been approved, hasn’t been rolled out yet, but that should happen soon.

The update hit sometime late last week. I believe Google manually pushed out the Panda 2.2 update around June 16th.

Bad news for all of us still Pandalized waiting to have the penalty lifted.

[edited by: Brett_Tabke at 11:34 pm (utc) on Jun 21, 2011]
[edit reason] split from previous thread last week [/edit]

[Shaddows]

Not to get sidetracked (apologies, I started it and am continuing it), but G engineers are very specifically NOT hurting DMD. They are working on an algo, with no express intent to harm or promote any given web property.

They merely seek to return the results that give peak aggregate utility (trying to strip away the subjective connotations of "best") to their users.

As always, it’s the algo that did it.

-------------------

Added

Philosophically, it’s a bit like the death by automated drone argument.

Say you're an innocent civilian, attending a wedding, and you get killed by an automated drone (non-existent at this point), who had identified you as a hostile. Who killed you?
The person designing the drone?
The person designing the rules of engagement?
The person who launched the drone?
The person who wrote the "identifying" software?

If Google reduce their traffic to DMD properties, was it...
Mr Panda, with his machine-learning techniques?
The human testers, who created seed-sets?
The engineers who selected the criteria to apply to Mr Panda's heuristic?
The engineers designing the SERP user satisfaction system?

[edited by: Shaddows at 2:43 pm (utc) on Jun 23, 2011]

[dataguy]

Just FYI, in the past week I've seen a huge increase in scrapers spoofing the http_referer to make it look like it's coming from Google search. If I didn't have several layers of scraper detection, I'd be fooled into thinking my traffic from Google was improving.

I've read several reports from webmasters saying that they've been released from Panda, but their income hasn't returned. I suspect their traffic increase is actually due to scraper traffic.

[indyank]

I've read several reports from webmasters saying that they've been released from Panda

several reports! Would you be able to share them? anywhere you go on the internet to find for recoveries in panda 2.2 update, you will come back to webmasterworld. Here, we would have got reports from only 2 or 3. People in the SEO niche are making nice headlines and stories.

[danijelzi]

Does anybody know if its possible that Panda 2.2 is rolled out only for searches when a Google user is signed out? Not only that logged-in rankings are generally different than logged-out, but there's clearly an absence of scrapers outranking my content when I do a search when logged out. Also, my pandalized site is ranked better when logged out for about 20 queries I tried out. I have used two different G accounts for this test.

[AlyssaS]

danijelzi - if you've clicked on those scrapers while logged in, G probably thinks you like them, hence the reason they keep serving them up to you. Signed in search is personalised.

The only way to check for accurate rankings is to use the adwords preview tool, which is unpersonalised and allows you to select the data centre you want to check:

[adwords.google.com...]

[danijelzi]

Thanks AlyssaS, that explains the situation. Now its important to remember not to visit scrapers when signed-in stay there too long, giving Google an idea about how great and engaging they are.

[HuskyPup]

Signed in search is personalised.

I can assure you that it is also personalised even when not signed-in therefore to get their actual results you need to delete cookies on a very regular basis otherwise you will not get G's "normal" results whatsoever which begs the question "What does a normal user who is signed-in actually see?"

I bet it's nothing like my results.

[AlyssaS]

which begs the question "What does a normal user who is signed-in actually see?"

I bet it's nothing like my results.

Normal users will see what the adwords preview tool shows, for most keywords that they search just once or twice.

But you raise a good question. What is the difference between personalised signed in search and personalised signed out search? From what I can signed out search resets itself every algo change but signed in search continues to remember your history.

Anyone else got any thoughts on this?

[indyank]

signed out search resets itself every algo change

:) nope.It is until you delete the cookies (max. time)

[indyank]

Just learnt that there was a Philippine Spike for pandalized sites on June 6. Do any one see it on your graphs?

[AlyssaS]

indyrank - does that mean people who haven't cleared their cookies (i.e. most people) are still on pre-Panda searches, for those keywords they happened to search for before? Surely not?

[indyank]

signed out searches mean that your activity can be tracked only at the browser level and not on Google's database. Algos have nothing to do with personalization.

Google does expire cookies often and it isn't that the cookies remain always the same.

[edited by: indyank at 5:59 pm (utc) on Jun 23, 2011]

[indyank]

coming back to the philippine spike, I noticed it on 20 Jun (mine was a small peak for that month for one of the sites) but it looks like hubpages, wisegeek etc. experienced the Philippine Spike on June 6 (source - pandalized.com)

[supercyberbob]

I reported the .ph spike in one of these threads a while back.

But I haven't investigated further. My speculation is some kinda pre-Panda thing going on. Shrug.

Would be interesting to know what happened exactly.

Edit: I should probably mention my traffic now is comprised of a lot less US gooble referrals and a lot more international tld gooble referrals, relatively speaking.

My intra-day traffic graph is a lot flatter, if that makes any sense.

Not sure if that's a typical Panda 2.0+ thing.

[wiseapple]

Noticed that we also started getting more traffic from google.co.ph right around June 8th and actually it continues to build. Previous to June 8th, we had maybe one visitor from google.co.ph. They must be testing something. Also, some of our Panda issues seemed somewhat resolved. The results are different from say a google.co.id.

[KJBweb]

Just FYI, in the past week I've seen a huge increase in scrapers spoofing the http_referer to make it look like it's coming from Google search. If I didn't have several layers of scraper detection, I'd be fooled into thinking my traffic from Google was improving.

Scraper detection, care to share what forms of scraper detection you use?

One of my sites got smacked down the listings hard since the 1st of this month, and quite rightly so.

It was a local news aggregator with links to the source, mainly just to pull news (via RSS) for a specific area all into one place with small business listings on there and a small forum too.

As said earlier, a re-think for some is required where the primary goal needs to come away from the revenue and more towards the user experience... ie. Satisfy the user :)

Bounce rate is now more important than ever it seems.

[drall]

To this day not one site in my vert has recovered from the original panda. This is out of hundreds to thousands of sites that we watch.

Some public traded companies, some with tens of millions of backlinks. I have watched many take radical redesigns with no recovery and apparent further decay after redesign.

I dont think you the webmaster or SEO can do a damn thing to change what was put into play with Panda and you are wasting your time.

[rlange]

indyank wrote:
Just learnt that there was a Philippine Spike for pandalized sites on June 6. Do any one see it on your graphs?

Yep. I've mentioned before that my company's main site, which lost traffic over the course of about 2 weeks around the middle of March, saw a significant increase in traffic this month. That started on June 6. There was another increase on June 13, but that particular increase went away on June 17.

All traffic from the Philippines (to an old, old, old page about musical instruments in India).

--
Ryan

[derivative]

My major site that was heavily penalized with Panda 2.1 has returned to its first page rankings for several keywords. I decided to 'ride the storm' and not change a single thing. Lo and behold it returned as I couldn't find any major violations. I feel that if I keep changing things all the time, some variant of the Panda algo might keep it in the 'box for something, fluke or not.

[zeus]

I have now removed 1100 pages manually on google tools, be cause they are SO slow in removing pages from there index, but it takes a few minuts for a new site to be fully spidered. Now I have a cramp in the hand from all this manual cut and past.

[walkman]

Some public traded companies, some with tens of millions of backlinks. I have watched many take radical redesigns with no recovery and apparent further decay after redesign.

I dont think you the webmaster or SEO can do a damn thing to change what was put into play with Panda and you are wasting your time.

Smart words, I learned the hard way. Treat any messages from Google PR team as such, PR designed to delay, to protect their reputation and not tell the obvious. In the same category I would put the Google koolaid drinkers that refuse to admit that Google may have made a mistake or that the site is fixed.

At some point you will harm your site doing change after change because Google didn't respond to the first changes. So use your common sense, look at what others have done, results from it and choose the most reasonable explanation, even if it includes engaging in blasphemy and assuming that Google might be not as noble as they claim.
[en.wikipedia.org...]

[mhansen]

@Walkman +1

We did some hoop jumping after the Mayday changes last year and found that it was nothing but an exercise of blind faith. Change servers, reduce clutter, re-organize content, etc etc. When it was all said and done, our site recovered... but we really had no clue why.

We weren't really effected in either way from Panda... but we already decided that if we were, we would do nothing but find other streams for the lost traffic.

Incidentally, that was also part of our plan during the Mayday update last year, and we feel that those added streams of local marketing and demand generation have been the best investment of our time and money.

[flanok]

OK I have another theory,


but I need some input as I am not 100% sure of how cached pages work within Google.

My website is in 4 sections


1/ Home page General content (no Ranking Change) | 2/ Catergories of content (no Ranking Change) | 3/ Detailed pages Of Catergories ( 30% gains 70% losses in ranking) (mixed bag) | 4/ Very detailed information about the detailed pages (no ranking change these pages would probably rank just from the title tag and url as they are very specific and have none or very little external links)


As you can see it is only the 3/ "detailed information" pages (that also delivers the most traffic) that are seeing changes through the latest Panda update with 30% of pages gaining and 70% dropping.

In my opinion the content is all written by me, so I think it is consistant throughout and pretty good, although my grammer is not the best.


The important point is this. The "Very detailed" pages link to the "detailed pages" on an average of around 8 to 1. (only specific "very detailed pages" link to the related "detailed pages", gaining around 8 internal links each).

What I have noticed is that the cache for many of the "4/ Very detailed pages" shows a "page can't be found" message, even though the page is still in the index and ranks similar to before.


So my opinion is, I am losing internal link value whilst I wait for google to re-cache these pages.


This seems to be confirmed by comparing the "detailed pages" pages that have improved and seeing the "4/ Very detailed pages" still have a cached version of pages that links to them and the opposite.

The "detailed pages" pages that have losses, appear to have a large quantity of "4/ Very detailed pages" that have lost their cached version and show the "page can't be found" message waiting fot Google to revisit.



Unless I dont understand how cached pages work in Google and what I see is not what Google actually uses, I believe that when Google spiders revists these pages and reaches them I will regain the internal links and therefore rankings.


Is it possible that many of you out there have not checked a sample of cached pages through the levels of your sites, and it not a case of bad content but lost internal links during this interim period, waiting for pages to be re-cached.

My opinion only

[InsectOverlord]

Site was hit by Panda 1 to the tune of 20%. cleaned up the site a lot removing low quality pages. On the stroke of midnight last Friday/Saturday saw Google referrals increase 45%, which has held since.

Our SERPs are now better than before -probably the result of the extra work trying to fix Panda.

We are a UK based site with mulit-millions of Google referrals a month, ranking top 5 for our key search terms.

[stokelake]

Wow @insectoverload that is amazing. I know a load of people who have removed low quality pages, we removed 75% of our pages from google. But still our rankings have not recovered. Did you make any other changes?

[InsectOverlord]

Loads - we tried every piece of unfounded made up advice from the SEO forums! We blocked Google accessing our Affiliate store, although we left affiliate link on the pages, that might have been a biggie. Who knows what worked.

[miozio]

It's now a fact for one of our sites that Panda 2.2 was in favor of scraper sites. Malware sites blocked immediately by my computer security rank on Google with the full copy of our site front page (as seen on G cache) I can not enter these sites with spammy URLs but can view the contents using G cache. One of them prompts download of a computer scan program with a dialogue popping up!

Its an attack, really! It was not so bad a couple of weeks ago!

You know what? That does it Google, you are obviously becoming a malware and using you may harm my computer! Checked bing, these sites don't exist it the index.

[outland88]

Who knows what worked.

Or if anything at all worked!

Enough people are beginning to report that sites have recovered doing absolutely nothing that you can't help but wonder. Until this point I was on the line about Google letting people know whether a site was rammed with Panda or not. Now I’m not. Too many people have had to engage in immense amounts of work to recover earnings not knowing whether it would hurt them further. That’s not right based upon untold evidence they are doing wrong. On the other hand the big boys, material evidence or not, are clearly being given insider information to recover at a quicker pace. I can’t justify not giving equal treatment to all when Google can keep the site off the radar anyway even revealing the information. If they recover they are in compliance.

[brinked]

3 of my 5 clients which I took on after panda have recovered with the latest 2.2 update. I am not at liberty to disclose the names of these sites or the full details but I can tell you some of the changes they made. Some of the changes they made on their own as they realized their sites were lacking. The reason why they did not improve them sooner was because they were ranking so well and did not want to change something that is not broke.

client 1.

- deleted a bunch of outdated articles (from pre 2000 era)as well as pages that were useless to visitors such as sitemap, a change log etc.

- removed all nofollow tags at my advice that google may see that as an attempt to sculpt pagerank. This particular client had nofollowed every single menu navigation link (do not ask me why).

- Removed 2 300x300 square block ads that appeared above all content articles.


client 2.

They re did their entire website. They had a new website in the works anyway and I helped them structure the site navigation as well as instruct them on content placement. I also got them into video creation and lately they have really been taking off.

I am not putting too much stock into this recovery because the site was pretty new and was lacking big time with their content and usefulness. The only way to go was pretty much up.

client 3.

Ecommerce site with over 50k products, all have manufacturer descriptions. Lets just say they did not have to rewrite their product descriptions as I keep saying Panda is NOT about duplicate content.

They made 3 changes that took a total of about 30 minutes to do and they have made about a 90% recovery. They were hit back on the original panda, and then hit again when panda was released worldwide.


There are people out there that have recovered, probably a lot less than we have hoped for but they are out there. They are not going to come to us and tell us what they did and how they did it, why would they?

Just to add to this post. My own site that was hit by panda still has not recovered. It has made a partial recovery during panda 2.2 and the internal "widget" sections that I mentioned I changed the URL's for have recovered fully now after adding a trailing slash to the end of the URL's. The homepage is the only page effected as of right now. I am starting to think it is not related to panda at all.

[c41lum]

thanks brinked some good information there.