| 7:22 am on Jul 28, 2011 (gmt 0)|
@freejung - Yes, that was one of the most important changes, the bounce rate reduced dramatically. They're saying now that Panda 2.3 was pushed around July 23, but we started recovering mid July, so maybe Panda 2.3 was released earlier.
Yet with all the updates the scraper problem still persists, in my post I've shown the example where Matt Cutts' blog was outranked by a scraper, and this has been happening for months.
| 9:44 am on Jul 28, 2011 (gmt 0)|
Softland, around the 11th there was another update. Maybe different data is pushed at different times? IMO Panda is a black or white thing, it assumes that all sites checked must have certain things or they are bad, no wonder so many innocent sites have been caught in this. So just by linking to Cnet you may have lost your house or job...even though the content was 100% identical and the experience even better since cNet might be a faster download than from your site. But go tell that to Panda....
Heartless morons, imagine the strife they caused for the past 5 months and counting. Unless you find that technicality and fix, Google doesn't care about content. That's how good of a search engine it is, content neutral.
My keywords improved slightly, but nowhere near enough.
| 1:01 pm on Jul 28, 2011 (gmt 0)|
@Rasputin - we were hit in February, but I described it as slow decline because we had lots of other sources of traffic. So unlike some that saw a huge drop, ours was a slow slide down the precipice.
Traffic came back up around July 22-23, and has been very very strong ever since. Fingers crossed that we don't get hit when G decides to update Panda again
| 1:50 pm on Jul 28, 2011 (gmt 0)|
Walkman, I'm 100% with you on the external linking issue. CNET was indeed faster and also spared us bandwidth. Judging from all this might be that the fact we were linking externally and had a high bounce-rate (in the eyes of analytics, because that wasn't actually a bounce) faulted us. Plus that users were spending very little time on the site as they were eager to download.
Anyway, I know what a nightmare it is to deal with this and wish you all good luck. If only Google was more open with webmasters that are registered via WMT and are legit (I'd even sign-up for some manual verification or something similar).
| 6:27 pm on Jul 28, 2011 (gmt 0)|
So my site has recovered 90% each time from Panda 1, 2 and 2.2 with a slight loss on 2.3. My site has millions of pages with an index rate of about 50% (slowly climbing).
Here's what I did:
- Moved any pages that were thin (less than 3 sentences) to a subdomain. 301'd them.
- Any content I know was duplicated elsewhere (manufacturer descriptions and such) I loaded in AJAX so that it was tougher for search engines to read it.
- Massively expanded interlinking between pages.
- Removed banner ads from above the nav, and kept square ads (320) just peeking above the fold depending on resolution.
Hope that helps.
| 8:02 pm on Jul 28, 2011 (gmt 0)|
|Massively expanded interlinking between pages. |
The word "massively" here sets off some alarm bells. Was this designed with users in mind?
| 5:26 pm on Aug 1, 2011 (gmt 0)|
Trafic is still same , bouce rate is slowly getting steady at about 44% now on average.
Will keep you updated
| 1:29 pm on Aug 2, 2011 (gmt 0)|
I see a competitor who was on the original Sistrix list of Loosers in Panda 1 come back now - after Panda 2.3. They are back on page 1, right below me. Large site owned by major company with lots of link juice. From what I can see the site has not taken away any of their ads away - they have one leader banner in header, one on left column, one in right column and one in the footer. However, seems they have changed their architecture of their pages and focused more on theme-ing each page.
Regarding ads above the fold, on my own site, I have also slowly added back some ads (Adsense) to my site as well and did also not change my ranking. I was originally pandalized in Panda 1 too but recovered in about Panda 2.1. In Panda 2.3 full recovery eventhough now my site is -80% less pages in index (de-indexed 80% of site).
Right after Panda 1 this is basically what I did:
* Take away many above the fold ads
* Non-index 80% site for pages that are blank or not impt
* Remove excessive linking - internal, external, affiliate links
* Dynamic Theming of each page more tightly with "related posts" themes in side post and bottom of page
| 11:08 pm on Aug 2, 2011 (gmt 0)|
Suprised to find I have a small success story.
History: A number of my sites were hit by Panda 2.1 - despite extensive reading of numerous threads I still can't quite fathom why. Since a block of my sites were affected I guessed that there was something that Panda didn't like and that some sites were ok in their own right but were guilty by association. On the whole the sites that had more of: shared analytics account, shared adsense account, shared IP address were more likely to be affected.
One of my small sites has recovered. It was position 1 for its keyword and dropped at the time of Panda 2.1to position 10. It has consistently stayed in position 10 till I looked today when it's in position 2. Traffic that was down to 30% of previous has been back to old levels yesterday and today (so i guess the recovery was yesterday).
What I did: This site had adsense, was on the same ip address as most of my affected sites but did NOT have analytics.
In mid may I moved the site to a new host. For a couple of weeks I removed adsense but then at beginning of June I thought blow it lets at least get some income from adsense and put it back on! I added an FAQ page. I added a Facebook Like Button (which has achieved a grand total of two likes) and tried out a third party social bookmark toolbar which I meant to remove again but didn't get round to! I may have tweaked some text but nothing major.
And that's it
Oh but I should mention I did also get rid of about 10 of my other sites that I thought were of low value.
My conclusions: It could be a fluke but this was the only site that I moved to different hosting and this is the only one that has shown significant signs of change. It's early days and i don't want to get carried away but there might be something in this that helps someone else. It is only a very small niche site so small fry in the big scheme of things. My next step in a few days will be to move another site to different hosting and see what happens.
I will report back on any other developments.
| 6:18 am on Aug 3, 2011 (gmt 0)|
Seems like an admission that some filter(s) were loosened up and sites have recovered as a result.
It's likely they would downplay the impact of this, as it can be seen as a "rollback" if people escaped Panda after 6 months without making any changes.
Of course they don't want to say they are letting thousands of
|webmasters out of Panda jail that they now realize maybe shouldn't have been there in the first place. |
Congrats to those who recovered! It would help a lot if dates were mentioned, as any who recovered due to panda 2.3 might have been due to loosening of filters, rather than any action they took.
To my disappointment, very few told about any sort of recovery.
| 3:37 pm on Aug 3, 2011 (gmt 0)|
I thought I'd post an update of my progress. I posted that my analysis showed that the pages with thin content were my problems - I did rewrite most of them on many of my sites with mixed results.
Keep in mind this is before I realize Panda would be a fairly regular occurence.
It seems to have had mixed results - While most sites appear to be recovering there are a few still stuck in pergatory. I can't explain why most recovered and some didn't - structurally they are very similar.
I can say that as Panda was being re-run every month I noticed sites which I had reworked since the previous update and up to 7-10 days before the update seemed to be "picked up" during the update - anything 0-6 days before didn't get any boost from that iteration of panda, however they did get the boost on the next panda run.
I just hope now we get at least one more panda update (I'm sure there will be many many more) so that the remaining sites finally do get reassessed.
One interesting trend I started to see after the most recent panda update was something I've noticed the past 3 years until after the march panda update. It used to be I could tell you when an algo shift was being applied because my traffic would spike for a few days then settle down.
For example, I could tell you when the Feb panda update started and stopped because of the traffic spike i got in late feb.
However I couldn't see the spike in March, or April and only slightly increased in May, however in June and July I could see the spike again - however not as pronounced as it historically was.
On the plus side of things, my bounce rate hasn't changed since the June update. I mentioned that I had seen it jump up in the May update then drop in June. It didn't change during the July update. It's now the lowest it's been in a year.
So where's my traffic it? Well it's not fully recovered but I am probably 90% back to pre-panda levels in the US and still climbing. I've been holding steady at about 75% of pre-panda levels in Canada sites but I recently just started rewriting content in Canada so i'm expecting that number to rise as i get through another panda update or 2.
| 5:05 pm on Aug 3, 2011 (gmt 0)|
Here is my update:
I did get back the same traffic levels to a low traffic site I manage. This is in the entertainment niche and it would get around 2.2K visitors per day before panda hit this site on march 22.The traffic declined to around 1.3K per day by Mar 29 and stayed around those levels until July 21.Thereafter, it started to go up. Traffic reached 2.2K on July 30 but went down slightly to around 2K on the next two days.
It is a low traffic site with not anything great from a quality perspective.Being in the entertainment niche, it had a lot of images.
Added "noindex, noffow" to all pages that would have only an image with very little text on them.For example, photos from a gallery were displayed as thumbnails on the main page and clicking each of those thumbnails would lead to a page that contains its full image (with alt text) and a small caption.I added the above meta tag to those pages.
I know that Matt Cutts had said that pages with just images (like those on flickr) weren't affected by panda. However, I decided to make this change within one week from thereon as I didn't see anything else to do on that site. Moreover, flickr had a lot of UGC in the form of comments on its image only pages.
But there are far better sites in terms of quality which had good traffic before panda and they continue to drop.
I am still clueless on what panda is looking at. I hope the above summary would be of help to a few in that niche.
[edited by: indyank at 5:40 pm (utc) on Aug 3, 2011]
| 5:23 pm on Aug 3, 2011 (gmt 0)|
I do like to add that "thin pages" are just one form of "shallow content".
But I do know of a site that has several thin pages ranking highly for some popular keywords.
This site would have a lot of links on that page to other internal pages (which are are just search result pages) but the anchor text used would be variations of the keyword that it ranks for and some of these links would be within header tags.
Example: If a thin page on this site ranks for "gadget" all it would have on that page would be some 10 words on that gadget. There would also be links (around 20) to other pages within that site , with anchor text like "gadget free", "gadget free for <item>", "free gadget for <another_item>" and so on. out of the 20 links, three would be within h2 tags, 2 would be within h3 tags and so on. These links would either be on the sidebar or towards the bottom of page i.e. outside the main content area.
The page would also link to other related categories i.e. categories related to this gadget category. All these would also be present on the sidebar or as breadcrumbs.
But one another thing to note is this is a mega international site with domains geo targeting various American and European countries.
My guess is that panda algo parses a page and extracts keywords.When it finds diversity in the use of such keywords, it probably considers even thin pages to be of good quality. i.e. when the extracted keywords are relevant to the topic of the page and are found within header tags or as anchor text on that page, it probably considers that page to be of good quality! But this a guess based on the above example.
| 10:54 pm on Aug 3, 2011 (gmt 0)|
bouce rate - i really dont hope they put to much weight on that, be cause I have A LOT of pages with 100% bounce be cause they enter my site at my final image page which they have searched, means the user gets what they need in less then 5 sec. which I think is very good for my users, but maybe I have to do so that they have to read a lot of text, which would not have anything to do with there search but on topic and when its read they get the image, means pist users, but happy google. I remember a time when we had to make sites for users not google, EVERYTHING now is doing sites for google not users.
| 3:46 am on Aug 4, 2011 (gmt 0)|
what you see as bounce in your analytics tool never counts.They have methods to determine all those metrics from the content.
Moreover, most of these metrics are used to correlate their other findings on "Quality" (Google quality) and not used as a direct ranking factor.
| 3:14 pm on Aug 4, 2011 (gmt 0)|
I agree that bounce rate isn't a factor - and it is explainable because the page quality wasn't the best but it's better now which is why my bounce rate has dropped.
In my case it was an observation - that bounce rate was/is an indicator of when a panda update occurs.
| 5:41 am on Aug 6, 2011 (gmt 0)|
about 2 weeks back on one of our Pandalized clients , we did a lot of changes.
What we did :
Website had a lot of affiliates , and many pages such as index.html and index.html?uniqueid#*$! were in google index and receiving trafic.
First (using the parameters in webmaster tools) we "no index" all pages that had "?" in URL string
Then using anaytics we have identified all highbouce pages (most were product.html?uniqueid#*$! and made sure they are out of the google index . While we did use"?" exclusion tool , google took about 10 days ++ to eliminate most of them. We used the URL exclusion (webmaster tools) to speed up the process . We also found some pages (very few) with high bounce rates and almost nill time on page. We eliminated those as well
Third ,we discovered 2 days back that client had duplicate pages for each product (targeting diffrent words with same "unique" txt) . So today we have URL excluded (and "no index) all those suplicate pages...no comment .
we are looking now to rewrite all unique content that was highly SEO optimized and make it more "human". This step will be implemented in a week.
Nothing was achived in terms of recovery after the past 2 weeks in terms of increased traffic. 0%
However in the last 5 days (about) we have changed the bounce rate from 55% to about 43% and we feel confident that we achive -40% with minimal page exclusion (very weak pages that have consistently gave high bouce rate for the past 10 days.). (take that algo for "important/relevant/quality")
We still feel confident on our ability to achive partial or total recovery of this particular client , but we wanted to share our methods with the comunity regardless of results. After all , my team and I , we learned a lot from all of you.
| 7:12 am on Aug 6, 2011 (gmt 0)|
|Nothing was achived in terms of recovery after the past 2 weeks in terms of increased traffic. 0% |
If you achieve improvements inside the Panda 6-7 week manual update cycles, then these would be for other reasons.
Approx. dates so far have been :
Panda 2.3 > 22 July
Panda 2.2 > 18 June
Panda 2.1 > 09 May
Panda 2.0 > 11 April
Panda 1.0 > 24 March
It looks like Google is gathering it's data at the beginning of the cycle, so improvements that fall outside of this may take longer to update. Given that observation, research and research testing may take added phases and rework, climbing out of this algorithmn is likely to take a while.
One thing that concerns me is that some sites may have to be taken down by blocking pages and then rebuilding them. The Google standard has always been to build quality sites from day one, rather than tweak them mid term. That's a tough call on any business which went the wrong way, which is why i suspect Google restricted itself from 100% elimination of Panda "prone" sites, to give owners a serious warning shot. It has probably eliminated a lot of business' in the interim.
So in claims of success and the various tactics engaged, it would be helpful to compare them against the cycles. I wonder how many got it right in one go.
| 12:23 pm on Aug 6, 2011 (gmt 0)|
Did you mean 24 February?
| 12:14 am on Aug 7, 2011 (gmt 0)|
@Suggy - Sure ... [webmasterworld.com...]
| 11:40 pm on Aug 23, 2011 (gmt 0)|
Panda 2.4 ...has passed and website has lost 10% in trafic. Most days , our bouce rate is under 40%. We are now changing the TXT on each page (old txt was unique but heavily optimized)
Those are the last changes to the website. Will keep you guys informed
| 5:03 am on Aug 24, 2011 (gmt 0)|
Panda 1 took 30% of my traffic, panda 2 took another 20% I have removed a lot of thin content and added some more of good quality pages, but for now the traffic is just steady after Panda 2. About 6000 visitors a day. It really feels like Google put some time penalty, as nothing I do changes much for now.
| 5:50 am on Aug 24, 2011 (gmt 0)|
It can take a while for your removed pages to be excluded from the index. I wouldn't give up yet. I feel I am on the verge of being released on the next Panda update now.
| 5:55 am on Aug 24, 2011 (gmt 0)|
|I am still clueless on what panda is looking at. |
I think we all are. Panda is the Great Contraction of the Web (ie. getting webmasters to kill all those light paginations, thin content, image only pages, etc.) with no explanation of what Panda was trying to achieve and they win, we lose... and still try to play the "only" game in town.
Panda benefited Google in reducing the number of "live" urls (removed, merged, 301'd) which makes their infrastructure work faster. Unfortunately, there were no other benefits from Panda because the USERS never noticed the difference... only webmasters did.
What I take away from this experience is that the days of 2-4 page 500 word articles with ads has expired... and that is because G has to index ALL THOSE "thin" PAGES and it is costing THEM money (not to mention our bandwidth, too). So what better way to roll back than put the scare of "thin" and (not said yet but this is a THIN page) paginations of a single page article over a number of urls.
This is my opinion only, but it makes perfect sense to me.
| 10:27 pm on Jan 30, 2012 (gmt 0)|
Just realised that one of my small sites seems to have made a recovery. Wish it gave me a simple at a glance way of getting some of my other sites out of the Panda doldrums. Anyay here's what happened:
The site covers a whole range of different subjects, all pages written by me and unique. Probably no more than 100 pages in all. Hit by Panda in May2011 - Panda2.1 I think it was.
Traffic that had been very steady for several years went down to less than 30% of what it was.
Then 3 weeks ago I moved it to a different host and in the process got rid of about 20% of the pages which had very few visitors.
My stats show that the very next day my traffic started coming back.
My traffic for the last 3 weeks stands at 85% compared to the same 3 weeks last year.
So no quick fixes here as others have done these simple things and not had the results - but maybe an encouragement that if you keep trying things, somewhere you may tip the balance!
| 12:33 pm on Jan 31, 2012 (gmt 0)|
@nickreynolds, that is good new, congrats, and gives hope to those who are still Pandalyzed.
Anyone else here see any Panda related improvements by moving the same site to a new host and in the process getting rid of lots of sub par content?
Has anyone here seen Panda related traffic improvements by doing a full redesign of their Panda site (and keeping it on the same domain)?
| 2:01 pm on Jan 31, 2012 (gmt 0)|
I'll confess I didn't read this entire thread. However my blog got the penalty I assume was Panda due to an extra ad on the top of the page that ironically enough was suggested by Adsense. I removed it and notified search through webmaster tools. They responded that they would remove the penalty.
| 6:59 pm on Jan 31, 2012 (gmt 0)|
@jerryodom - I don't think that was Panda
| 7:02 pm on Jan 31, 2012 (gmt 0)|
|They responded that they would remove the penalty. |
They responded to you that the extra ad on top was indeed a penalty?
So those affected by the above-the-fold algorithm can file for a reconsideration request?
| 8:30 pm on Jan 31, 2012 (gmt 0)|
Ive kepy a personal log of changes to my wordpress based site, its small but maybes it will help other small sites.Ive checked and took out any site or company references..
Prior To Being Hit – 1/10/2011
So MYSITE has had this current layout now for over a year, its has worked well but I still wanted to improve on it, however as I was wary that Panda was doing the rounds I have held back on changing to my new layout Template which is by BLAHBLAH and will eventually add alot more functionality for users. So I decided to make changes to this template and spruce it up. Below is a list of changes made on October the first, bare in mind MYSITE has sailed through all previous iterations of panda.
Menu bar moved to left hand column ( People read left to right) so I thought putting all my latest articles on the left navigation area would help people spot new articles, take a read and help lower bounce rate and keep people on the site.
Added a Questions and Answer Section – I’m trying to make the site more interactive to users.
Changed the background colour of the site – Again just aesthetics but less contrasting so easier on the eye.
As said, all this was done on the 1st of October and I was hit on the 13th - 16th
The Recovery Starts Here 18/10/2011
As I had been fine up until those changes I’m hoping they contributed to the issue, however Ive made a few more changes.
The Navigation bar has returned to the right hand column.
Adsense has been trimmed to now only one unit above the fold.
I have removed non essential plug ins to aid page speed. The site has seen a 2 second page speed increase from this.Hopefully users will see this.
Google+ has gone, yes it was getting clicked and had been on for a few months however I think the speed of it loading was more detrimental and I had yet to see any referrals from people who had plussed a page.
I am increasing internal links between pages.
Ive ran WC3 Validator and got any errors down to just 5, of which i cannot fix as they are beyond my coding skills, never less I fixed over 25 issues.
Finally Page Titles, I have removed the domain name from all page titles and now just run with the standard title of the article.
That’s all for now, this wont be a quick recovery but I will see what future Panda iterations do to the site and will try to improve accordingly.
Ive modified my robots .txt to the below, I don’t want google spidering and indexing things it does not need to or may class as thin content.
Ive have now got the W3c Validation errors down to 0. Fingers crossed that helps. I have also started increasing the detail of the problems on each article. This may take a week or so to complete.
OK, so MOST of the articles have had the problems re-wrote now, google has acknowledged in webmaster tools that the website is now loading almost twice as quick but no sign of a recovery. On top of this google is sending RUBBISH traffic, I am getting traffic for keywords that are almost unrelated to the site so bounce rate is taking a beating. I do now have 11 new articles to put up on the site but struggling for motivation at the minute however they will be up next week.
November 19th Panda Update
Google announced via twitter that there was a small panda update on the 19th of novemeber, on the same day I removed all amazon affiliate shops. Since this date im seeing a 20% increase in traffic, I shall continue to monitor this over the next week and post any “panda flux” which Matt Cutts says can continue after a week and panda updates.
Still no recovery but the fight goes on, ive changed how archiving is indexed by spiders setting them to no follow. Im wondering if this was being classed as duplicate content. Starting to run out of ideas…..toying with a full re-write in the new year and finally putting the new layout on…EEEEEK.
Well, I did it. Ive redid the whole sites layout with a new theme. Its a lot cleaner. This is my last change on the site, I will give it 3 months now to settle. Happy Christmas.
Success!, I can confirm that today I have seen a boost of 20% extra traffic the site appears to be recovering from google panda! Over the next day or day or so ill monitor google organic traffic and keep you updated. I’m also noticing referral traffic from other sites up also, Im guessing they seem to be recovering also. Maybe this could be seasonal, we shall see.
| 8:34 pm on Jan 31, 2012 (gmt 0)|
Ontop of the above I re-wrote all of my Meta Descriptions.
| This 173 message thread spans 6 pages: < < 173 ( 1 2 3  5 6 ) > > |