Welcome to WebmasterWorld Guest from

Forum Moderators: Robert Charlton & goodroi

Message Too Old, No Replies

Panda DATA ONLY - separating Panda from the bamboo

11:43 am on Mar 24, 2011 (gmt 0)

Junior Member

5+ Year Member

joined:Oct 27, 2010
posts: 61
votes: 0

There seems to be a ton of theories kicking round about Panda, and a lot of noise but little actual signal, so I'd like to kick off a thread just for data.

I'll start:

- Large site (1m+ indexed pages)
- Aggregator + UGC in entertainment
- Huge amount of content, but spread quite thinly (more content than the pages now outranking us however)
- Ton of backlinks (including 100k+ from one of Google's flagship sites... (not Blogger!))
- Strong level of user engagement (1m + registered users). Low (for our vertical) bounce rate (< 50%). High number of page views / visit
- Monetised only through affiliate links (no "ads" at all).
- Absolutely no black hat tactics (never bought any links, never created doorway pages etc)
- Only invest in SEO to a "hygiene" level - i.e. we are more interested in building a brand / best in class product, than an SEO empire. Or rather SEO is a means to an end rather than an end in itself.
- We have several small SEO issues that we've never cared to fix (see aforementioned desire to build a product, rather than a SEO site), e.g. a slightly f*ked up sitemap, semi-duplicate content due to problems with UGC / aggregation logic etc.
- Recognisable brand (especially amongst our demographic). Gets a ton of direct traffic. Best in class product (by quite a margin). Lots of word of mouth recommendations and constant praise on our feedback page.
- Huge amount of growth (5x) in the past year, and across the board (referrals, social, SEO, direct).

We've been hit:

- across the board to all page templates
- no recognisable pattern to the pages hit (some of the worst affected have the most content and strongest link profile)
- about 20% down on average
3:22 pm on Mar 24, 2011 (gmt 0)

Senior Member

WebmasterWorld Senior Member tedster is a WebmasterWorld Top Contributor of All Time 10+ Year Member

joined:May 26, 2000
votes: 0

I agree - plain information without theories and complaints is a good thing. Here's one site I work with.

Google traffic dropped 19% at Panda Update time (Feb 24)
Now down 12% (March 24)
Conversions remain stable - no corresponding drop
Lost traffic was almost completely for single words and a few 2-word phrases
3:30 pm on Mar 24, 2011 (gmt 0)

Full Member

10+ Year Member

joined:Oct 25, 2005
votes: 0

Are these sites replicated on other TLDs?
How wide spread is the duplicate content issues?
Is content scraped from other online sources?
What other "small" seo problems might you have with the site?

Any of the sites I've examined post Panda really did deserve the smack down after close inspection. (that's not to say some got hit that didn't deserve it of course, just mean the sites I personally reviewed).
3:36 pm on Mar 24, 2011 (gmt 0)

Senior Member

WebmasterWorld Senior Member tedster is a WebmasterWorld Top Contributor of All Time 10+ Year Member

joined:May 26, 2000
votes: 0

The site I mentioned has some duplication and scraping out there (many sites do), but it is minimal and scraped content does not outrank them - they are the original content source.
3:47 pm on Mar 24, 2011 (gmt 0)

Junior Member

5+ Year Member

joined:Oct 27, 2010
posts: 61
votes: 0

Are these sites replicated on other TLDs?
>> Nope

How wide spread is the duplicate content issues?
>> Minor (< 5%)

Is content scraped from other online sources?
>> Kind of, the model is similar to an IYP (although the site is not in local) - i.e. we have base records that are then fleshed out with UGC and editorial content

What other "small" seo problems might you have with the site?
>> Well.. not many that I know of. We constantly delete a ton of content, so we have a lot of pages that 404 (which I've had problems with in the past). We also have a lot of thin user profile pages etc. indexed, however this has never caused problems before. If we don't bounce back from Panda, noindexing this content will be top of my agenda (although tbh we've got more important product work to be doing atm than SEO).
4:21 pm on Mar 24, 2011 (gmt 0)

Preferred Member

5+ Year Member

joined:Mar 22, 2011
votes: 7

Any user content generation? The only commonality I could find in our sites that were hit Panda.
4:32 pm on Mar 24, 2011 (gmt 0)

Junior Member

5+ Year Member

joined:Oct 27, 2010
posts: 61
votes: 0

Yes, lots - close to 1m items of UGC (reviews, comments, images, videos etc).
4:40 pm on Mar 24, 2011 (gmt 0)

Senior Member

WebmasterWorld Senior Member 10+ Year Member

joined:Apr 23, 2003
votes: 0

My data:

- Small(ish) site - was about 200 pages, now about 100 pages (since I've deleted much of it)
- Informational and it gives users free stuff to download (all original stuff - printables, etc.)
- Some thin content that can't help but be thin (like a downloadable image for instance, with only instructions on how to download it).
- Some duplicate content that can't help but be duplicate (like a famous old poem or something like that).
- Some articles that I believe fits the footprint of content farm articles. Only had a couple dozen of these, and they were all original, written by me, quality articles, but they had that same how-to type format, and/or some of them targeted similar phrases (10 great widgets for teenage boys, The best widgets for preteen girls, etc.)
- A handful of duplicate content articles that was from using articles from article directories years ago.
- Backlink profile - nothing great, nothing really horrible. I'd call it "meh, so-so".
- Bounce rate of about 50%. Many come to specifically download something, do so, and leave. Average page views is 2 (usually going from category to specific download).
- Monetised through both adsense and affiliate links. I've removed all the affiliate links (since they really weren't doing well anyway), and reduced the adsense presence a little.
- No blackhat seo.
- Was one of the very first sites to focus on this niche 7 years ago.
- No UGC at all.
- Static site, no database, no autogen'd metas or anything like that.

I've been hit:

- across the board
- impossible to really gauge the impact since the site is seasonal, and this isn't the season for it.

What I believe might be causes for the smackdown:

- the couple dozen articles on the site smelled like a content farm - even though there were very few articles, they had that same type of content - how-tos, lists, etc. All written by me, all unique, all quality, but the format might be a farm type trigger. (What I've done about them: I've deleted them all and 410'd them).
- a handful of articles from years ago that I'd used legally and legitimately from article directories, that are all over the web. (What I've done about them: I've deleted them all and 410'd them).
- a couple dozen pages with dupe content that can't help but be dupe content, such as the words to an old famous poem. (What I've done about them: Some I've deleted, and some I've kept).
- Dupe content within my site: Some pages have the same "how to download this thing" instructions on each of the downloadable thing pages. (What I've done about them: Not much. Not sure how to reword that in so many different ways. I suppose I could link to one page of instructions, but that seems less than user-friendly).
- Too many affiliate links? Maybe, not sure, but since they weren't really doing much for my bottom line anyway, it seemed like a good time to dump them. So I did. That dumped quite a few pages since that was a whole section of the site.
- Thin content: Some pages just can't help but be thin. Here's a thing. Download it. Nothing much else to say about it. I've dumped a few thin pages, but some I just can't/won't dump because they are important to the site and to the user.
- Number of ads and ad placement: I've already mentioned that I got rid of all affiliate links and ads. I've also reduced the adsense presence somewhat. I've never felt the site was overwhelmed by ads, but that doesn't mean others feel the same way. So I've pared it down so that there's basically one above the fold and one at the bottom. If that's too much, then that's just crazy.
- I've seen some people say that misspellings might be an issue. I checked for that, and found only 5 minor misspellings across the site (unless false positives are in play, such as using names that aren't in a dictionary but are legit, or appreviations such as tsp for teaspoon).

Final thoughts:

I have other sites in similar niches that approach everything in exactly the same way as this site, yet they weren't hit. The differences include:

- the other sites didn't have an article section, so no "farm footprints", and old articles from article directories.
- the other sites had less affiliate links, but do have some
- the other sites also have thin pages that are just naturally thin (here's a thing, download it).
- the other sites have about the same number of ads and ad placement
- the other sites do have the same type of dupe content within the site, such as "how to download this thing".

So, if I had to compare them, I'd say the small article section on the smacked site was the biggest difference. I don't personally think a total of about 30 or so articles would constitute a "farm", but the footprint might be the same, and the fact that some of them were from old article directories might be the kicker.
5:17 pm on Mar 24, 2011 (gmt 0)

Senior Member

joined:Dec 29, 2003
votes: 0

Compared to my competitors, the only thing that could have done me in: tags, too many of them. However, I removed them a day after Panda hit on WebmasterCentral [google.com...] and since then traffic has only gone down. I also 'deleted' (at least with a noindex) about 35% of the existing pages, just in case. Google has cached them at least twice since Panda and now most of my site has a March 18 cache, with index and 2 other sections having a daily freshtag. No luck in SERPS though.

Looks, so far, that Google hasn't run the algo on everything after Panda, I can't point at a site that says "we did this and came back."

For a few minutes last night on AOL search, I think, I saw my site's ranking without any 'filters.' Loved it, but it was short lived :)
12:13 am on Mar 25, 2011 (gmt 0)

Preferred Member

5+ Year Member

joined:Mar 20, 2011
votes: 0

- Medium size site (100k pages indexed), long history.

- in entertainment/content

- Large amount of content, but all of it thick, detailed, original content. (thin content pages and people scraping or excerpting our content now outrank us)

- Ton of backlinks from a wide variety of quality sites

- Medium level of user engagement, low bounce rate, 2.5 pageviews per visitor

- Monetised through Adsense and other banner ads.

- Have never used black hat tactics at all, and have mostly focused on creating quality content and not so much on SEO at all beyond a very basic level.

- Had SEO issues which hadn't been fixed pre-panda. For instance Google for some reason indexed several thousand random search pages from our database that we never gave them and Google seemed to be generating itself. Google also somehow indexed some pages which didn't exist, still not sure how that happened. Had an old message board which Google had also indexed, even though again we didn't give it to them, however we hadn't "nofollowed" it because again, we were focused mainly on content and not SEO.

- User content generation? None except for the old, closed message board.

- Well known, well respected brand, especially in our area.

- Huge amount of growth in the past year, as much as 3x, in all areas from Google, to direct, to social, etc.

- Scrapers steal a LOT of our content, to the point there's no longer anything we can do about it. However most of our competitors who weren't hit have the same issue.

How We've been hit by Panda:

- Massive deranking of all keywords across the board. Lost nearly all top keyword rankings.

- Traffic from Google dropped 40% when Panda implemented, dropped further in following weeks down to 50% on average.

- Unable to determine any pattern in the pages hit the worst, some of the pages which were most heavily de-ranked had some of the most in-depth, extensive, widely shared linked to, and heavily trafficked from a variety of sources outside just Google. Panda penalty seems to have been applied globally, the keywords hit the least still sunk 7 - 10 spots. The worst ones sank 300 or more.

My Current Theories On The Cause:

- The only thing I've been able to come up with is all those random, search pages Google was erroneously indexing. Also we weren't using "canonical" in our headers so in some cases Google was also erroneously indexing bogus URLs to some of our pages, so that it looked like there were three copies of a page, when it was really just Google creating different URLs for the same page, which, again, we weren't giving them. It's sort of like Google created the problem then penalized us for it. But none of the duplicate pages google created were ranked highly or bringing in any traffic or all, they were just out there somewhere. But that's the only theory I can come up with to explain what happened. Have disallowed and used canonicals to theoretically eliminate this now, but there has been no change at all.
7:03 am on Mar 25, 2011 (gmt 0)

Junior Member

5+ Year Member

joined:Apr 1, 2010
votes: 0

- large site (upto 6 million indexed pages)
- UGC (but mostly dupe content like directory)
- conversion ratio same
- 20 % drop approx in traffic
- content replicate on other TLD's
- content was also replicate on the whole website
- dynamic website
- more than 4 million registered members
- industry (directory, business listings, products)
- hardcore OnPage SEO
- lot of backlinks - paid, unpaid
- monetized by adsense
- still ranking on some keywords (post & past update)
- normally all the websites in our industry has the same seo structure
- was ranking on branded keywords like toyota

possible factors:
1. the bigger the site, more the impact. Even of a little change
2. adsense
3. hardcore onpage seo
4. google is jealous of me :P