Some Penguin Insights

Thought I will share this with you guys and maybe we can see more tests done on Penguin.


The reason I shared them is this:

Site1 - Penguined

Site2 - No problem


You can clearly see the first one was killed on the 24th. What are the differences between the two sites?

The Penalized one

Has 15 pages.
Not all pages have links, there is one page that has 2k links, another has only 10.
I bought about 60 high PR private blog posts about 6 months ago.
It had a 10k wiki blast to one inner page.
Anchor text is not diverse, very concentrated.
Has one adsense unit at the bottom of the page.
Has SEO plugin installed (meta keywords, descriptions).

The Not Penalized One

Has 6 pages.
Every inner page has about the same number of links, homepage has a lot more (more on that later).
No High PR networks at all.
No wiki links.
Anchor text is incredibly diverse. There was a service on BHW called something like spam gone wild, it is closed now. This site got a review copy for that service. Basically they are xrumer links but with hundreds of different anchors, and the content is different on all of them. I don't mean spun, but actully different.
Has no adsense but 2 banner ads.
Doesnt have SEO plugin installed (meta keywords, descriptions).

My Conclusions

Link quality is bullshit. The xrumer links are doing well after Penguin, given that anchor diversity is there, not sure how important the unique content is.
Link velocity is bullshit. The xrumer links were built in a short timeframe.
Link velocity may be important for the high PR links. Lots of high PR links popping up at the same time -> high PR blog network flag.
Wiki links may not be good, but maybe its because the content is not 100% unique, Google sees lots of links with very similar content ->spammer flag. Google has seriously improved in detecting similar stuff lately. Maybe that is the real problem with all web 2.0 links?
Adsense may be bad, but I doubt it is really important.
Having no SEO meta tags may be beneficial?

Thoughts?

We need more data to conclude if there is a pattern here, it might just be random

Here's my situation - two sites, two emds, same linking shemes, same amount of content and everything.
One got penguined, one improved.
WTF?

the one obvious difference is anchor diversity, one has it and the other doesn't

Seems like google is getting really good at detecting organic, natural looking SEO work, and punishing anything that looks like more time has been spent on SEO than on actually making good content.

OP that was a good post.

But at the same time, I echo Ras' sentiments above. I have the same experience. WTF.

I want to add little bit of my theory too:

All my sites were EMDs with or without - in it.
Out of 15 only 2-3 sites are ranking, the surviving one are those which had lower number of comment links(with SB but Actual PR, I have stopped doing that now) or same links given to the sub pages.(these is one site to which I gave links to its page and those were not same as the EMD, the keyword was different)

The things which I can say are:
I will not give SB Actual PR links to my sites again directly.
I will use 3-6-9-12 spoke web 2.0 link wheel backed with 3000 LOW OBL Actual PR links.
Better use homepage links, I can get niche relevant homepage links ( not blogroll links)
Two sites don't have adsense on them, only one site has adsense code on it.

This is all what I can say on offpage, I have some thoughts on offpage too but those are different.

I just had an AHA moment.

Why are they running Penguin manually?

Checking link graphs and anchor text would be very easy that the regular algo could handle. The reason is they check where the link is coming from. They analyze the content like they do with Panda. That is the resource-heavy stuff. Sooo as I suspected the unique content on the linking page is really important, and where you place the link.

I guess they got some skills with Panda, identifying duplicate (spun) content, and they utilize it to check for link spam with Penguin.

the information are quite consistent with the latest update of google algorithm but is it somehow the same with the other niche site?