| 11:19 pm on Dec 8, 2011 (gmt 0)|
I've been watching WebMaster Tools for the website I moved the pages to, some important terms from the pages I moved are already in the top 10 there. Keep in mind it's not a "new" domain I moved the pages to, it's my three year old hobby site which Panda eventually came to love.
But I did start another new experiment yesterday, 301'ing a 30 page subject off my main site, which is also Pandalized, though not as badly as the previously mentioned one. In this case, I moved the pages to a domain that I set up as a blog for friends who never used it. WordPress was installed and there was a single page up, I nuked everything, created a new HTML home page and sent the 30 pages over to it. I just hope that one page blog wasn't Pandalized, I have no way of knowing:-)
| 3:41 am on Dec 9, 2011 (gmt 0)|
OK UPDATE] Dec/08
SKYROCKETING in positions , completely regained all old positions as of today DEC / 08 and somehow started ranking for extremely competitive keywords !
And I will repeat again ADS Do Not Influence nor their position as long as they are moderated !
| 8:33 am on Dec 9, 2011 (gmt 0)|
When i have recovered few months ago, i had the strange feeling that my site was human reviewed after changes that i've made. Is possible to be a (learning) machine that simulates human quality raters.
| 9:40 am on Dec 9, 2011 (gmt 0)|
So what did you do Donna? Spill the beans please.
| 9:42 am on Dec 9, 2011 (gmt 0)|
So what did you do Donna? Spill the beans please.
I was deep crawled on the 8th, 5x normal crawl amount, then smashed again today, getting tired of this crap, getting around 1k uniques, that was 25k a few months back. Everything I try just seems to make it worse.
| 2:23 pm on Dec 9, 2011 (gmt 0)|
My original experiment is still going strong with one caveat. The "problem" page I talked about in the thread about trusted and untrusted sections on a single page hasn't been reindexed since I completely rewrote it Wednesday night. So every other page in the collection is getting over 200 visitors a day from Google and this page is getting just TWO! For those who didn't read the other thread, this is a page that was ripped off and syndicated both in whole and in parts, with Google reporting tens of thousands of copies of various parts. I'll feel a lot better when they take another look at it.
I've been keeping an eye on SiteLinks, and while Google has dropped all the pages I moved from the SiteLinks for the old site, they still show them immediately after the six-pack as returns for the old domain, even though the URL shown is the new domain. No idea how long that takes to resolve.
The experiment I started Tuesday with 301'ing an entire subject off my main site to another domain has already kicked in, with partial day results from Thursday indicating that those pages will draw at their pre-Panda levels. Interestingly, I had Google Alerts in place for URLs from the main site, of the form, www.mainsite.com, and Google sent me a notification for the most important page that got moved this morning. But Google Alerts was always flaky, so who knows.
If nothing else, I'm having fun with this stuff for the first time in many years.
| 4:34 pm on Dec 9, 2011 (gmt 0)|
Good luck, content_ed. I've done a lot of experimenting with adding subdomains and changing domains, and my experience has lead me to conclude that Panda is simply recalculated much less frequently than the other updates. Changing domains properly can have a positive effect within hours, but my sites have always eventually gone back to being Pandalized.
None of the other site improvements I've made have had any effect, so having a few weeks of high traffic here and there have been nice, even though it didn't last.
Hopefully, your mileage WILL vary.
| 6:28 pm on Dec 9, 2011 (gmt 0)|
Two questions, if you would.
1) Did you ever try moving a collection of pages to a new domain rather than the whole domain?
2) Did you move to a brand new domain, or to a domain that was unaffected, or beloved by Panda?
| 8:43 pm on Dec 9, 2011 (gmt 0)|
Yes to all four scenarios. Sometimes when I only moved a collection of pages, I didn't get the initial jump in traffic. Other than this, they all followed the same pattern and within a month traffic was reduced.
I should point out that on several sites, after I split content into subdomains by user accounts, traffic was eventually reduced but it was still better than it was before moving the content to subdomains.
| 4:05 am on Dec 10, 2011 (gmt 0)|
I might be compelled to retract the "smoke and mirrors" comment above IF there is long term results. In general, however, what has been reported here (WW) and other sites is a gradual decrease in all site indexing/serp results. Could it be that G has met the web and it is bigger than G? Don't know... but I have no doubt that all search engines are looking very hard at what is sometimes called "duplicate" (within a site) and "redundant" (found on other sites and/or scraped). And that the web continues to explode with "sites" as every Tom, Dick and Harry get on board. Tom, Dick and Harry, these days, means COUNTRIES providing content.
| 6:07 am on Dec 10, 2011 (gmt 0)|
@tangor Its seems like it, but the google way of letting us know is extreme .
| 2:16 am on Dec 11, 2011 (gmt 0)|
Two more questions if you don't mind.
First, did you ever try moving pages or domains back where they started after the traffic dropped again.
Second, when the traffic was reduced for pages on a new domain within a month, was it reduced to roughly the same level as before the move, or higher/lower. I don't ask about subdomains because you mentioned they stayed a little higher and I'm not using them in any case.
Oh, and another second question. Since the timing of the drop was variable, did you notice if it ever coincided with Panda updates? They were reported to be happening pretty frequently over the summer.
I did a little more page moving tonight just to complete a subject grouping. I'm not unhappy with how the sites are organized for humans at this point, so even if they get pushed back down where they were before the move, I'll keep them where they are.
Oh, and final second question. Did you ever see a recovery at the next Panda update for a site you moved the pages away from? That's the part of this experiment that attracts me the most since my sites aren't large to start with, and I'm moving whole subjects. It would be funny if Panda looked at the Left Behind content on a site and let it out of purgatory because it had an issue with the pages that moved on.
| 2:50 pm on Dec 13, 2011 (gmt 0)|
One week update:
Google visits continuing at pre-Panda levels for the pages, now up to a dozen, moved from the 80% down Pandalized site to the 300% up Pandaloved site. No signs of the move in either WebMaster Tools account yet, and Googling the original domain name still shows results for pages that have been moved, though it correctly identifies them as being located on second domain.
The second experiment, moving several dozen pages from my less-Pandalized site to a new domain with no Google status has also resulted in a traffic boost, but not to pre-Panda levels, more like halfway there. While all of the feature pages are close to double week-on-week, the visitor shares aren't distributed the same way as a year ago. This I would attribute to non-Panda changes in the algo and changes in competition.
I suspected from the start that the transfer to the previously unused domain wouldn't be as effective as the transfer to the Panda beloved domain, but I wanted to try it both for the sake of the experiment and to keep the subjects segregated. It will be interested to see if the two moves are treated differently if and when the mother-in-law crashes the honeymoon.
Side note that maybe should get a thread of its own, but the page I rewrote (@2,500 words) from top to bottom because Google was completely ignoring heavily pirated and syndicated sections, recovered appreciably as soon as it was reindexed. Last Tuesday it got 1 visitor from Google on the new site, yesterday it got over 60.
I really reworded it as opposed to rewriting it. The meaning of the text is the same, just expressed differently, though with the same basic vocabulary. I'm fooling around with doing the same for some other heavily infringed pages, even though it's a lot of work and probably hurts the quality a little since the language is so forced. The other pages in the group were from the same original date, so it didn't have anything to do with the age of the original content.
| 6:03 pm on Dec 17, 2011 (gmt 0)|
Update Dec 17 :
Rankings have grown 15-20% of pre-panda across the board. Approximately 2500 pages noindexed. Indexed pages under 500 and stable now. No spikes or dancing. Long story short, since the change only positive results in less then 2 months. Crawling has been around 20-25%% of total pages every day for the last 2-3 weeks. Cache is finally frequent.
| 2:10 pm on Dec 20, 2011 (gmt 0)|
Passed the two week point this weekend with no changes in previously reported ranking. It's the start of our lowest traffic period of the year, this week is poor and Christmas week is horrible, so I won't be able to identify minor changes. If the wheels come off it should be obvious, otherwise I won't post another update until after New Year's.
The pages moved from the worst Pandalized site to the site that Panda loved continue to draw at pre-Panda levels with just two exceptions. The two most popular pages in the group, those that had drawn well over 500 visitors a day from Google each before Panda on the old site, are only drawing half that today, but it's still several times what they were getting three weeks ago.
Both of those pages have been altered since Panda hit by removing links to our other Pandalized site which ranked just as high in the same area (I broke all links between those sites). So it could be that Google really used to like the links on those pages, or it could be that ten months without the steady addition of organic links they would normally have seen has dropped their value such that even if Panda was eliminated, they would still rank lower. That's the sucker punch waiting for small sites that do recover from Panda if they had some super high ranking pages in competitive areas that needed to draw organic links on a daily basis just to maintain their place.
The second experiment won't reach two weeks until midweek, but things there haven't changed either. This was the move from a Pandalized site down about 66% to an unused domain with unknown Panda status. Traffic bounced around 100% but it needed to bounce 200% to get back to pre-Panda.
| 5:55 pm on Dec 20, 2011 (gmt 0)|
Interesting post Content. Thanks for reporting back here! I'm definitely keenly interested in how it turns out.
I found a competitor that has 301'd his entire ecom domain to a new shorter "brandable" domain name.... and his traffic in SEM rush when compared to his old domain is exactly the same as it was back in February... and his rankings have visibly skyrocketed in the SERPS. This is definitely an interested "loophole" in the Panda affect.
| 5:23 am on Dec 21, 2011 (gmt 0)|
According to Tedster's log [webmasterworld.com...] the last Panda update was in November. Until the next Panda update all these developments are theoretical.
| 7:16 am on Dec 21, 2011 (gmt 0)|
Content refinement is necessery to recover the website from Panda penality
| 7:17 am on Dec 21, 2011 (gmt 0)|
|Rankings have grown 15-20% of pre-panda across the board. |
@Donna - not sure these recoveries are due to the regular iterations. Must be something else you are doing that allows the rankings to climb in between.
I am seeing recovering pandalized sites that have newly loaded pages with unique quality content, newly indexed and ranking at No1 within 24-48 hours for only those pages. Pandalized pages have generally moved up.
It seems like Pandalized rankings can adjust between iterations. Anyone else out there seeing this ?
| 6:43 pm on Dec 21, 2011 (gmt 0)|
@Whitey - I'm also noticing that NEW FRESH pages on my pandalized site have an easier time ranking (albeit for long tail key terms) than my existing pandalized pages (which seem to be in a sort of sand box)... I've also seen on two occasions someone from Google HQ checking out the newly created page(s).
| 7:48 pm on Dec 21, 2011 (gmt 0)|
@Whitey , newly pages rank no probs, but I was mentioning my old penalized pages that are climbing based on the same rankings as before. Only difference I see for the 1st time is that G is trying to in my opinion classify which pages are informational or directorial. As an array of keywords bounce up or down for example : "widgets info" array kws goes up and "widgets examples" array goes down and vice verse .
| 10:42 pm on Dec 21, 2011 (gmt 0)|
OK, this is too funny. I was just looking at Analytics data for the moved pages, and I realized that my search traffic has been rapidly falling - from Bing and Yahoo! (same thing).
For the highest traffic of the pages moved, Bing and Yahoo! used to account for around a third of the visitors. Last couple days they've been down around, ZERO.
So whatever happens with Google going forward, apparently Microsoft figures that moving pages is evil. Maybe I should sign up for their version of WebMaster Tools and see if it makes a difference.
| 3:05 am on Dec 28, 2011 (gmt 0)|
Update Dec 27 :
Still going up in rankings approximately 15% since Dec 17 (last update I posted ).
| 5:31 am on Dec 28, 2011 (gmt 0)|
"Oct 15-17 noindex all review pages and deindexed my forum a week later , removed H1 tags and xml maps. That's all . "
Are you referring sitemap pages to be removed or sitemap.xml? Also I am using <h1> tags for all my title articles, so are you saying you use H2 or other for your article titles?
| 9:37 am on Dec 28, 2011 (gmt 0)|
I tend to believe all these penalties most of your are facing are over optimization and some new or updated link algo. Panda is just the smoke screen. I rest my case !
| 1:00 pm on Dec 28, 2011 (gmt 0)|
Donna, I noticed somewhere that you said content zoning plays a role too.
By content zoning do you mean the proportion of space above the fold that is content, with respect to ads and pictures?
This may be a factor on my sites where I have used large pictures and large call to action text. It's not exactly thin content in my opinion it's just that the good stuff is further down.
I will look at removing buy buttons and making text/images smaller if that is what it takes.
| 3:37 pm on Jan 4, 2012 (gmt 0)|
It's now been over four weeks since I started the 301 experiment which has two parts. I'll review a little since these posts are so infrequent, but full details can be read in my previous posts in this thread.
In experiment #1 I 301'd a dozen pages (in two phases) from a site that lost 80% of it's Google traffic to Panda 1.0 and 1.1, to my hobby site that was hurt in 1.1 but then recovered on its own and then saw it's Google traffic jump to 3X pre-Panda on October 14th.
Nine out of 12 pages recovered to pre-Panda levels within a day or two of the move, and are still there. Three pages improved without recovering fully. One recovered fully after I reworded it from top to bottom, the other two are doing much better but are still down from pre-Panda, around 30% for one and 50% for the other. They were previously the strongest pages on the Pandalized site, drawing in the 700 - 800 visitor a day range, so a year out of the spotlight might have cost them rank since they lost organic link building momentum. The twelve pages added around 2,500 visitors a day to the hobby site.
Bing and Yahoo stopped sending any visitors to the pages as soon as they figured out they'd been moved, so this would be a bad experiment for anybody who counts on Microsoft for a big chunk of their traffic. In terms of Google, moving the pages clearly got them out from under Panda, it's now just a question of whether Google will slam the whole site when they get around to the next Panda update.
In experiment #2, which I started exactly four weeks ago, I moved a collection of around 30 pages from my main site, which was hit in Panda 1.0 and 1.1 (though not quite as badly as the site above) to a site I've owned for three years but never used beyond installing WordPress. The collection of pages are currently drawing around 800 visitors a day, half of that to the most popular page.
All of the pages moved to that site jumped appreciably after the move, but not back to pre-Panda levels. I would say they were down about two thirds when they original site got Pandalized, and now they're down around one third. Hard to say with the seasonal traffic. I attribute the lack of a full recovery for any of the pages to the fact that this site is Panda neutral, neither loved nor hated.
I'm less sanguine about the pages that didn't immediately recover on being moved pointing to specific problems after looking back at a full month of Analytics. In one case, the rewritten page went from 16 Google visitors the first week to over 300/week after the rewrite was indexed, and those 16 the first week are within Analytics margin of error for zero. Those differences, by the way, were about what it did pre-Panda and post-Panda on the Pandalized site. It was actually the least popular page in the collection for some time before Panda as well.
But on the other two pages, looking at a full month's data I can't really see the affect of the rewrites, and I hate to hang a theory on a single page. For what it's worth, I am highly confident that rewording pages on Pandalized sites has no immediate effect, but I decided to do it anyway for the pages that were the most attractive to content thieves in case it helps when Panda does update.
I'll report in again after the next Panda update, it will be interesting to see whether the destination sites or the source sites have their status changed.
| 4:17 pm on Jan 4, 2012 (gmt 0)|
Thanks for keeping up with the detailed posts.
| 8:41 pm on Jan 4, 2012 (gmt 0)|
As you mentioned... the real test is going to be after the next Panda run!
| 6:03 pm on Jan 7, 2012 (gmt 0)|
Much appreciated @content_ed and @donna, great info thanks
| 11:51 pm on Jan 7, 2012 (gmt 0)|
Amazing information! Thanks content_ed and donna!
| This 78 message thread spans 3 pages: < < 78 ( 1  3 ) > > |