Recovered From Panda: Well I Guess So!

[zehrila]

Today, either they reverted their panda update or i am just being lucky to see old rankings, some of which are even stronger. What i did so far.

1: Changed hosting
2: Changed nameservers (just for the heck of it)
3: Used WP recommended robots.txt file for blog
4: Blocked some internal dupes and weird urls through robots.txt
5: Fixed some 500+ broken links/images.
6: Noindex Nofollowed thousands of pages.
7: Removed adsense altogether from main pages. Atleast 1 unit still exists on some very deeper pages, which i have noindex, nofollowed.

Further things i have lined up to do.

1: Thousands of canonical tags (they seem tricky to implement)
2: Thousands of 301 redirects.
3: When i create a Widgets page, it creates 6 other pages e.g green widgets, blue widgets etc. It at times get very hard to fill those pages, so iam in a process of enabling only those pages, for which i have actually added content. It shall significantly remove thin pages.

I just hope its not a temporary thing and ranks stay. There is still a lot more to fix in my site, the errors, which i was not bothered about and took rankings for granted.

Hope it helps someone.

[irock]

zehrila,

May I ask when you implemented the fixes mentioned in the original post? I just like to know how long it took for Google to re-rank your pages.

[zehrila]

May I ask when you implemented the fixes mentioned in the original post? I just like to know how long it took for Google to re-rank your pages.

I did not do anything the first 9 days and then i started making changes slowly, first thing was to take ads off, and switch hosting and nameservers, don't know if it has anything to do with nameserver or hosting, few of my other sites hosted on old hosting are still affected.

Started observing spike in ranks 3 days ago.

[irock]

So, you implemented the fixes 10 days ago? Also, did you completely remove ads (I assume it's adsense) from your site?

[zehrila]

I have been implementing changes from past 10 days, there are still too many changes to implement. I removed ads from Homepage, Main Product Pages, reduced ad units from 3 to 1 on deeper less important pages.

[walkman]

Looks like Google gave someone else some experiment time on the serps. I lost whatever pages I was ranking for and definitely G couldn't gauge anything from the few users it sent. Site:domain.com is still a mess, reporting long ago deleted pages.

[crobb305]

Walkman, when did you see this happen? Did it happen this morning?

[zehrila]

Just wondering if any one else has applied these steps and saw any recovery?

[nippi]

I've recovered as of 36 hours ago.

Day1 of recovery, traffic doubled, adsense tripled
Day2 of recovered, 4 hours in, traffic has doubled again, back to my best levels of 18 months ago.

Seems my fixes have not only got me back from Panda, but also from something else.

I rewrote poor content. By poor, it was either too short, badly phrased, spelling errors, partial or exact copies from other sites. Articles are now varied in length, and high quality. Not all articles fixed, but a large proportion, and the worst ones done first.

I stopped google indexing thin pages. Detail pages with little detail are now gone.

i did not speed my pages up and they are not real fast.
i did not remove any adsense and i have 2 blocks above the fold, one at the bottom of the page.

[indyank]

nippi, what is your bounce rate and avg. time on site like now?

[nippi]

bounce rate 3 days ago was average 73% with 1.10 average time on site

now its 65% and 1.37

becuase higher quality pages are ranking i guess

[brinked]

nippi,

thanks for sharing your experience.

I rewrote poor content. By poor, it was either too short, badly phrased, spelling errors, partial or exact copies from other sites. Articles are now varied in length, and high quality. Not all articles fixed, but a large proportion, and the worst ones done first.

That is a very important observation when you point out articles varying in length. I have a site that was hit in Panda 2 and it is a very high quality site and being that its more of a media site, I had content written for each category page. This content was actually created as an attempt to avoid any type of "thin page" penalties but all pages have pretty much the same article length (about 200-250 word paragraph).

Any natural content site will have articles of varying size and formats (ie some will have more pictures than others, some will be 4 paragraphs long, others will be 10 pages etc). This makes a ton of sense and I can not believe I did not think of it earlier. Either way, I have removed all of this filler content about 2 weeks ago so I am still waiting to see if it has any kind of effect.

Thanks for sharing and congrats on your recovery!

[dickbaker]

Nippi, I think yours is the first story of a real recovery that I've read. Congratulations.

A question, if you don't mind: was any of the thin content you had manufacturer descriptions or other text that appeared everywhere on the web? You say you had duplicate content, and I'm just trying to figure out how duplicated it was.

Another question, if you don't mind. Did you regain many or most of the spots you'd occupied for various phrases pre-Panda, or have rankings simply improved, but you're not necessarily re-ranked for the same phrases? I'm really curious about statements from some people that sites can regain their rankings. For example, if you ranked #4 pre-Panda for the phrase "XYZ blue widget" and the page Google showed was www.example.com/acme_xyz_blue.html, is that page now ranking close to the same spot for the phrase "XYZ blue widget"?

[nippi]

dickbacker

thin content was

(1) Manufacturer Descriptions.
(2) Scraped Meta Descriptions for business descriptions
(3) Same repeated descriptions from mnay other directories.
(4) Short pages under 150 words. Some under 30 words.
(5) Wikipedia knockoff articles.
(6) Some 15% modified articles.
(7) Some member contributed total duplicate articles.


My rankings had been solid on many phrases for years prior to panda

I have regained rankings pretty much to the exact positions i held for all content prior to panda 1.

Its a complete, total, absolute, undeniable, Panda is not affecting my site at all recovery to exact rankings prior to panda on 98% of my keywords that I monitor

[indyank]

Some 15% modified articles.

do you mean articles where just 15% of the content are unique and the rest are either from the product manufacturer or other sites?

So what % of content in a page has to be unique? The answer to this will be crucial as Google seem to have made a big change here.Obviously, there will be articles where many of us will like to retain manufacturer's description, but add our own content (text and/or media) and functionality, if any, over and above it.It doesn't make sense if google is expecting people to completely remove the original manufacturer's description.

Did you remove the pages that were thin (I think you considered pages to be thin if they had 150 words or less) or you added noindex to those pages? How exactly did you remove them as I am hearing google isn't ignoring pages with noindex tag in computing the site level panda score.

Congragulations on your recovery. :) Sorry for forgetting to do that before.

[nippi]

What i meant was i was trying to pass off as unique articles copied from elsewhere, with about 15% of the article's content rewritten to make it look like a totally new article.

i dont think making sense is the issue for google, its about not having the same content appear in masses of its results.

i have programmed my site so thin pages get a link to them which is hidden from google, in addition to adding no index no follow meta tags to them, so they are neither crawled again, and also removed.

thanks for the congrats everyone.

please beleive me when i say ive done nothing but fix content.

adsense is the same
page speed is untouched.
links are no differernt.

it a content quality filter.

[indyank]

What i meant was i was trying to pass off as unique articles copied from elsewhere, with about 15% of the article's content rewritten to make it look like a totally new article.

So you added 15% unique content to articles scraped from other places and google considers them as unique now. that is interesting.None of the sites I deal with and affected by panda have that kind of duplicate content.While you could pass of articles with just 15% (or 15/115%) unique content, i am struggling with sites that are far more unique!

i have programmed my site so thin pages get a link to them which is hidden from google, in addition to adding no index no follow meta tags to them, so they are neither crawled again, and also removed.

So you are hiding those links to googlebot and still google is fine!

Now I understand how google should be treated going forward.

[tedster]

So you added 15% unique content to articles scraped from other places and google considers them as unique now.

I thought he was saying that the pages Panda originally nailed were built that way - scraped content modified by 15%. He radically improved or removed content that was created that way and his rankings returned. No more gaming, in other words, but real original content.

[c41lum]

Hi nippi

in addition to adding no index no follow meta tags to them, so they are neither crawled again, and also removed.

When you say you "removed them", what do you mean?

[zehrila]

c41lum: I assume he means, he blocked them using robots.txt and and added No Index attribute in his pages so that search engines don't index them.

Nippi: Good luck, my traffic is growing ever since i got out of panda.

[c41lum]

what im trying to understand is, did nippi 'remove' them from his site linking structure. Or was it a case of just using the noindex tag but linking to them as he did before.

[nippi]

yes, the old articles with only 15% difference are gone.

i have not removed the thin content pages.

a. links to these pages are set up so google will not follow them. its entirely legitimate to have pages for people and not for search engines. For example, i might have a business in my directory with little or no details. Such a business will still have a detail page accessible by users, so if and when the business owner finds it... they can take control of it and update it. But no point in having this detail page indexed, its got no more content than the stub in the business category pages section.

b. such pages which are linked but a link google can not follow, are also set to no index no follow, so they removed from search engine indexes.

They remain accessible on my site, and i want them to stay that way.

for the reord, i remain completely recovered from Panda.
In my mind its just a content quality threshold filter like all of those that come before it.

[indyank]

Ok. So you are saying that you had rewritten content in those articles to make them unique. I was saying the same thing in my original question. But my actual question was did you rewrite them completely or added more content to those pages to make them unique. I am trying to understand this as it will help in deciding whether to retain the manufacturer's description for google.

I feel that retaining the manufacturer's descriptions is useful for users, though it shouldn't be the only content on the page.If google is expecting people to add their own 100% content for products, without retaining any manufacturer's description, it isn't good for users.It is a case where the page is fine for users and all other search engines except Google.

such pages which are linked but a link google can not follow, are also set to no index no follow, so they removed from search engine indexes.

i guess you are saying that you added the nofollow tag to all links to them from within your site.If you are saying that you added "noindex, nofollow" to those thin pages, then it is something different.You are only preventing the googlebot from following links on those thin pages, but not the pages themselves.

[Whitey]

I feel that retaining the manufacturer's descriptions is useful for users, though it shouldn't be the only content on the page.If google is expecting people to add their own 100% content for products, without retaining any manufacturer's description, it isn't good for users.

I'm 100% behind you on this. Removing aggregated product descriptions is neither practical from the users point of view and the site owners, where the expense could not be met for large sites. And writing replacement unique content could be a waste.

Siteowners would have to be on the look out for other improvements which can be more economically applied that can reverse them out of this mess.

[tedster]

There's a big difference between scraping and then spinning/enhancing a bunch of articles and reproducing manufacturer's descriptions. I would definitely not drop a manufacturer's description completely. Including them just CAN'T be a Panda trigger.

[Whitey]

But would you go along with blocking non performing product pages during onsite tests in preparation for some Panda analysis (if&when).

[nippi]

I'm sorry, I'm simply not in favor of manufacturer deescriptions.

(1) Usually they are too short.
(2) if google does not want 50 pages of results with the same wording, why would they favor your page over another if both ahve the same content?
(3) It's often in US style rather than Australian English which for me is not good.
(4) It pays no attention to keywords.
(5) its often got links to the manufacturers site in it, which I don't want.
(6) Its usually ..... crap and boring.

eg Blue cotton party dress with white frills and a white print. Prewashed/ color fast. Adjustable waist

This beatufiul little party number is sure to delight your little girl. In a divine summery blue embellished with white frills and a modern Bali themed print, it's suitable as both a party, formal or even an "I just feel like wearing a pretty dress down to the park" dress.

It can be worn above or below the knee and the waist is adjustable to fit differernt sizes. The material is 100% lightweight cotton, is very soft, prewashed and shrunk so it will keep its colours, shape and size after washing.

etcetc

I write product descriptions that give more information, allow the shopper to bond with them. I get unqiue content, and descriptions that convert more customers. It works. No competitors in my marketplace other than me do it, and none rank as well for shop products(now that I'm unpandaed)

Ok, if you are selling tvs and computer parts, not sure what you can do to redescribe such items... but for me its rewrite all the way

[indyank]

Hmm...This is what I suspect...

I would say that having the manufacturer's description is really useful even in the example you quoted.

egBlue cotton party dress with white frills and a white print. Prewashed/ color fast. Adjustable waist

This beautiful little party number is sure to delight your little girl. In a divine summery blue embellished with white frills and a modern Bali themed print, it's suitable as both a party, formal or even an "I just feel like wearing a pretty dress down to the park" dress.

I would still like to retain the description (italicized) in my page as it gives a brief description (features) of the product.

I would then go on to add more useful content in the form of images, videos and laying down the benefits and usage.

but if the google engineers were thinking like you do and enforce uniqueness like you say...it surely is not good.

[suggy]

Nippi, we write all unique content for our website (and pretty much in the same style as you). We still got pandalized big time!

My concern is that quite a lot of this content has been scraped over the years. It's mainly done in parts (so a stolen paragraph from our page is blended with others to make nonsense!).

When I do a search for those stolen phrases, the theif ranks above us in google's results. For our unpandalized competitors, they rank ahead of their scrapers/ copycats. Why?

I checked the headers of the top ten results and noticed we're the only ones returning last modified date. This is refreshed every time the database is updated (for prices,etc). So even old pages, look modified recently.

Then I noticed my nemsis scraper's pages have older last modified dates. Did we inadvertently make them the original in Google's eyes?

Anyone with a similar problem found the same scenario?

[c41lum]

The only site of mine which has recovered is the one i deleted all the content on. It was 2000 page now its just 2 pages. It was a niche price comparison site, we didnt use any generic manufacturer content or images. We had 3/4 sentences of unique content on every page all written in house. No Adwords all advertising links were nofollow. The site is on a 1998 domain and has been comparing prices on this certain niche since 2002.

Now its only 2 pages the site ranks number two for a hyper competitive term and number 8 for a term i never dreamed of getting on the first page for.

Im mad that i had to delete all my pages but I could see no other way. Now I have a 2 page site that ranks well but no way of monetizing it Im scared to put any kind of affiliate links on the site just in case we get smashed again.

[walkman]

Even a PHD dissertation has lot and lots of direct quotes. The truth, however is that we don't know what triggers Panda. And then we have this: if you want to write a brief Bill Gates bio to enhance whatever you have on a page, you have to write the year he was born, what he did, what he founded, with whom etc. Rehashed? Duh! I am not going to turn Bill Gates into a US Civil War General to be unique.

Month four is a few days away but I'm no longer buying the 'content' thing for a lot of sites.

c41lum, probably all the juice left is going to those 2 pages.