Following on from
this thread , in which even Matt Cunts himself even says that you may well have to abandon your site if it gets slammed by Penguin, I tried to recover 5-10 of my sites hit by this retarded update.
Many people are preaching X or Y conclusions about Penguin, but they are mostly guesses. All I have is experiments per site in which I changed 1 variable and considered the results. In other sites I changed multiple variables at the same time to see if the penalty 'trigger' was because of a combination of things (as it can't always be just one thing that is triggering the penalty). As with anything SEO related, my results are provided "as is", and are never 100% certain. No one's analysis can be. If, during any future update/Penguin refresh, any of the sites do change, I will add it on this thread. I began testing almost immediately after the first update around 24 April and so I made it before around 24 May for the next refresh. It appears that Penguin does not slap sites constantly, rather on each refresh, however once a site is slapped, if it is ever to recover, it must wait until the next refresh.
The results are not very promising. I PMed briefly with Expertpeon regarding the random nature of the update and how it seemed to be very much to do with links (the main constituent of web spam these days) and its overwhelmingly punitive nature. His findings seem to be somewhat consistent with the results I got. Here goes:
Variables changed:
1. relevance of backlinks (a. link context, b. page meta tags, h1 etc.)
------> result, those sites which have already been slapped simply moved further down after addition of more 'relevant', high quality links. This actually surprised me. However, other sites which were not slapped by Penguin, have maintained their rankings even after the refresh, after I added relevant, high quality links to them as a sort of 'prophylactic' against any future slaps. Nonetheless, this latter result is inconclusive because these other sites may simply not have been slapped regardless of what links I added after the initial update. More curiously, there is no difference in the link profile between slapped sites and non slapped sites. I backlink all my sites in a similar way.
2. anchor text dilution using SB (99% nofollow links)
-------> result, this technique definitely does not work. I had heard from many places that too targeted anchor text would cause a penalty, so I made a list of around 50 generic anchors as well as naked urls in different format e.g. http://mysite.com, http://www.mysite.com etc., and slammed them to a 1 million AA list. After the next refresh there was no difference.
3. building more high quality DF links from different sources (targeted anchors)
--------> result, this technique did not work. Again, a rumor going round was that link diversity, not just in terms of anchor text, but also the source of the links, was a key requirement for avoiding a penalty. One of my sites had around 700 DF links already. I slammed it with another 1200 from a variety of different high quality sources. There was no difference in rankings for penalized sites. I actually thought this variable would bring my site up to the top again, even if it was penalized, because if I use these kind of diamond links on a fresh site I will shoot to the top for most of my keywords.
4. decrease on-page optimization (aka on-page pessimization)
--------> result, after reducing densities of any keywords above 3% down to 1% (absolute max), one of my sites that had been slapped from 500 visitors per day to 0 visitors per day, recovered 10% of its traffic for about a week, before going down to the abyss again. A 'dead cat bounce', hence, this doesn't seem to be the main target of Penguin either. Another site, I reduced keyword density, de-optimized meta tags, h1 tags. Again, no difference even after the refresh.
5. anchor text dilution using high quality, DF links
-------> result, after slamming another site with similar types of links as in variable 3., but instead with generic anchors and naked urls. Again, surprisingly, no improvements even after the refresh. If anything, the site went down.
6. dropping links with no context (high quality, DF links)
-------> result, although Matt Cunts is the most notorious bullshit man, I could not ignore the 'examples' of webspam that they were targeting in the original webspam update on 24 April on the G blogspot. Namely the examples of links dropped in completely irrelevant, completely nonsensical, poorly auto-spun context, common amongst second-rate spammers. I changed the link context to be relevant (using not always precise keywords in the content, but rather LSI) in variable 1. above, but I also wanted to test simply removing all the context. Previously this was not very effective at all, but since Penguin was retarded, my thinking was that it might now be effective once again. Unfortunately, this shit doesn't work, just as before!
7. 301 redirect from penalized domain to new, non-penalized domain
-------->result, site rose to the top as if not penalized, but 1 week later penalty was passed on and fell to the abyss.
Conclusions from the above results:
My Guesses / Misc comments regarding Penguin
Disclaimer
The above is only the result of my own experiments, and my guesses/comments are just that, guesses and comments. The algo again may change at any time in the future. I would also welcome any of your input and findings regarding this update.
Happy spamming,
JB
this thread , in which even Matt Cunts himself even says that you may well have to abandon your site if it gets slammed by Penguin, I tried to recover 5-10 of my sites hit by this retarded update.
Many people are preaching X or Y conclusions about Penguin, but they are mostly guesses. All I have is experiments per site in which I changed 1 variable and considered the results. In other sites I changed multiple variables at the same time to see if the penalty 'trigger' was because of a combination of things (as it can't always be just one thing that is triggering the penalty). As with anything SEO related, my results are provided "as is", and are never 100% certain. No one's analysis can be. If, during any future update/Penguin refresh, any of the sites do change, I will add it on this thread. I began testing almost immediately after the first update around 24 April and so I made it before around 24 May for the next refresh. It appears that Penguin does not slap sites constantly, rather on each refresh, however once a site is slapped, if it is ever to recover, it must wait until the next refresh.
The results are not very promising. I PMed briefly with Expertpeon regarding the random nature of the update and how it seemed to be very much to do with links (the main constituent of web spam these days) and its overwhelmingly punitive nature. His findings seem to be somewhat consistent with the results I got. Here goes:
Variables changed:
1. relevance of backlinks (a. link context, b. page meta tags, h1 etc.)
------> result, those sites which have already been slapped simply moved further down after addition of more 'relevant', high quality links. This actually surprised me. However, other sites which were not slapped by Penguin, have maintained their rankings even after the refresh, after I added relevant, high quality links to them as a sort of 'prophylactic' against any future slaps. Nonetheless, this latter result is inconclusive because these other sites may simply not have been slapped regardless of what links I added after the initial update. More curiously, there is no difference in the link profile between slapped sites and non slapped sites. I backlink all my sites in a similar way.
2. anchor text dilution using SB (99% nofollow links)
-------> result, this technique definitely does not work. I had heard from many places that too targeted anchor text would cause a penalty, so I made a list of around 50 generic anchors as well as naked urls in different format e.g. http://mysite.com, http://www.mysite.com etc., and slammed them to a 1 million AA list. After the next refresh there was no difference.
3. building more high quality DF links from different sources (targeted anchors)
--------> result, this technique did not work. Again, a rumor going round was that link diversity, not just in terms of anchor text, but also the source of the links, was a key requirement for avoiding a penalty. One of my sites had around 700 DF links already. I slammed it with another 1200 from a variety of different high quality sources. There was no difference in rankings for penalized sites. I actually thought this variable would bring my site up to the top again, even if it was penalized, because if I use these kind of diamond links on a fresh site I will shoot to the top for most of my keywords.
4. decrease on-page optimization (aka on-page pessimization)
--------> result, after reducing densities of any keywords above 3% down to 1% (absolute max), one of my sites that had been slapped from 500 visitors per day to 0 visitors per day, recovered 10% of its traffic for about a week, before going down to the abyss again. A 'dead cat bounce', hence, this doesn't seem to be the main target of Penguin either. Another site, I reduced keyword density, de-optimized meta tags, h1 tags. Again, no difference even after the refresh.
5. anchor text dilution using high quality, DF links
-------> result, after slamming another site with similar types of links as in variable 3., but instead with generic anchors and naked urls. Again, surprisingly, no improvements even after the refresh. If anything, the site went down.
6. dropping links with no context (high quality, DF links)
-------> result, although Matt Cunts is the most notorious bullshit man, I could not ignore the 'examples' of webspam that they were targeting in the original webspam update on 24 April on the G blogspot. Namely the examples of links dropped in completely irrelevant, completely nonsensical, poorly auto-spun context, common amongst second-rate spammers. I changed the link context to be relevant (using not always precise keywords in the content, but rather LSI) in variable 1. above, but I also wanted to test simply removing all the context. Previously this was not very effective at all, but since Penguin was retarded, my thinking was that it might now be effective once again. Unfortunately, this shit doesn't work, just as before!
7. 301 redirect from penalized domain to new, non-penalized domain
-------->result, site rose to the top as if not penalized, but 1 week later penalty was passed on and fell to the abyss.
Conclusions from the above results:
- Penguin is a very punitive update. It errs on the side of over punishment, rather than leniency, in order to aggressively combat web spam and reduce the profitability of SEO, possibly in order to compel people to use Adwords to generate traffic instead.
- Penguin seems to be based on link spam, the collective most dreaded fear of most SEOs at BHW. No matter what you do on page, in order to reduce on page optimization, the penalty is not removed.
- Despite popular rumor, anchor text diversity doesn't seem to be a large contributing factor. Many of my other sites with higher anchor text %ages and less raw urls, generic anchors etc., remain unpunished. SB dilution does not work, just as higher quality DF link dilution does not. I was surprised by the latter, as I did think SB would probably not work due to NF links not being counted on the link graph.
- Once you have a Penguin penalty, you cannot move 'up' the search results as you might once have done. It's like a ball and chain being attached to your site. Building over 1000 high quality links to my site which would easily catapult a fresh site to the top positions, had no effect, or even a negative effect, on rankings.
- Links with no context, e.g. footer links etc., still have low value compared to relevant in context links above the fold.
- Penalty is passed on via 301 redirect, as with other penalties.
My Guesses / Misc comments regarding Penguin
- It is very difficult to get out of the penalty, once you are hit. Even if you CAN get out of the penalty, it will cost you, thereby reducing your ROI. In such cases it may make better business sense to start a new site, as Matt Cunts quite smugly suggested. However, it may not make business sense at all to even try to recover an old site, or even make a new one, especially if it would cost a lot to rebuild it. In such cases, SEOs are screwed.
- There is a *seemingly* random element to the update. It may or may not be random, but if it is, this greatly diminished SEOs ability to consistently rank for their desired keywords. I put 'seemingly' in asterisks because I don't think it actually is random, but there may be a fine line between penalized and not penalized - because all my sites are done the same way, yet some have been slapped and some haven't. Some really great quality "white hat" authority sites have been slapped, while other crappy thin sites are rising to the top, and vice versa.
- A huge amount of 'little guys', those who are more likely to link spam due to limited resources, are being removed by this update showing even MORE preference to big brands, in particular amazon. Others include walmart, bestbuy, target, facebook (in particular bizarre cookie cutter facebook notes pages directly linking to amazon via affiliate links), manufacturer pages, even when the search term indicates that the searcher would like more information, reviews, opinions etc. on the product/query, rather than a simple, very thin product page riding on the back of an authoritative domain. Again, this discourages regular guys from even trying SEO due to the reduced ROI (the top 5 positions are really the only places to be, but if they are occupied by amazon, bestbuy and walmart, it's no longer worth it).
- A lot of completely bizarre results are being observed. New, Thin sites with simple english and no real value in the content, with less than 100 links, are ranking for pretty decent, rather competitive terms, 10k+ exact local, while higher quality sites with 1000s of links all appear to be punished. This leads to confusion as to what exactly DOES make a site rank #1 because so much shit is at the top. But if you yourself make a shitty, thin site with less than 100 links, it is very unlikely you will rank. Again, there seems to be a lot of randomness/ luck in this.
- Now, instead of 100% of the time more links = higher rankings, on average more links = higher rankings 50% of the time. Either you get links and you avoid punishment for building too many / too crappy links and you win, OR you get slapped into the oblivion. Therefore, links are still very necessary but the ROI of link building is decreased. I am still making OK money every month but ROI has been cut in half for me. Even if you are spamming well, or you are spamming poorly, the average ROI will decrease as a result of this update. That's not to say SEO can't get you a very good ROI even today, but the fat square jawed loser Matt Cunts is tightening his square jawed grip on our balls.
Disclaimer
The above is only the result of my own experiments, and my guesses/comments are just that, guesses and comments. The algo again may change at any time in the future. I would also welcome any of your input and findings regarding this update.
Happy spamming,
JB
