The driving element behind this thread is another thread that you can read here: http://www.blackhatworld.com/blackhat-seo/white-hat-seo/289589-what-todays-search-engine-optimization-full-insight.html Long story short, it is too full of crap. Worse than that, it is written in a way that makes it look like it's some kind of tasteful dish made by a french chef. That last part is what kinda annoys me. I'm ok with freedom of speach and this is a forum where everybody's input is welcome, but making certain statements and tagging them as "correct" or "accurate" should happen only when you know what the fuck you're talking about. First of all, I will pick up on statements ExtraWinner made and explain them myself as well and point why I think otherwise. Very little of what he explained is correct or even close to reality. Some things are correct, but very few, and the overall message he transmits is not. Spiders Google spiders - 4 main groups: backlinks spiders, content spiders, picture spider, video spiders. NO! There is ONE single spider. ONE not many. The spider does one thing and one thing only. It discovers links and passes the page content to the indexing system. It started in the begining with a small number of sites that had many outbound links and from there discovered the entire web. The spider runs as a distributes system, obviously, but at a logical/software layer it is/behaves centralized. What happenes is the spider looks up the "database" of known pages (URLs) and visits those URLs. It builds a list with all URLs on the page and for every URL that wasn't in the "database" it saves it. Also, since it already consumed bandwidth requesting the page, it passes that page (the actual HTML and possibly other details) to the indexing system, so that there won't be need for a new retrieval. That's it! The spider's job is not to do any sort of analysys or "thinking" just to spider. The way it does this is not important, but there are a few things (useless) you might like to know: It does thousends of requests at the same time. Though it can do thousends of requests at the same time, it has to hit the web servers it crawls slowly so it doesn't produce a DDoS effect on them. That is why it has a subsystem which has to decide what URLs have to be crawled next but in such way that it doesn't crawl many URLs on the same server at the same time. It has to avoid loops. A single PHP script can generate an infinity of actual pages by using URL parameters. A subsystem of the spider has to make sure that the spider doesn't get trapped in such an "infinite" loop. Same system decides wether a URL should be crawled or not based on several factors. Won't go into details here but most importanly, it will not index too much duplicate content or URLs that are too deep within a subtree (which generally translates into "have low PageRank"). Just to make it clear, by PageRank, in this particular instance, I am not referring to the 0-10 value but to the real internal value and especially the page rank computed from the incoming graph edges. The indexing system does not really try to understand the text. It just saves it in something that is very likely similar to a reverse index. Because of the architecture of the reverse index they can figure out what pages are good candidates for a search query. Also because of the architecture of the reverse index it is somewhat easy to figure out poorly spun or copied content. There are some additional layers there, but the foundation itself is in the architecture of the reverse index. Google has been capable to detect duplicate content since a long time ago, it just didn't used that as a ranking or penalty factor, instead relying mostly on links. Even today with the Panda/Scraper update it still can't do this perfectly, though they say they got pretty good at it. For more information look on wikipedia the articles about web crawlers, search on Google videos for the presentations done by Google engineers about the architecture of their system and read the research papers on this topic. Google dance I have no idea why it happens exactly but I suspect it is a side effect of the computational process that has to do with figuring out what score that URL should get for the particular keyword. Very likely not all factors get computed at the same time. What I am pretty sure about is that it is NOT the result of many spiders that each give scores. Why do I think that? Because it would not be optimal design to have more than one spider or have separated components give scores (+ or -) by themselves. How I say it works is an educated guess only but I am pretty sure that it is not the result of multiple spiders each giving different scores. Sandbox The explanation though nice to understand and maybe seems to make sense, again is not correct. I have no idea how a penalty it figured out or computed internally but it is not a number resulted from adding + and - scores. For example there is no such thing as a content penalty. It's just that your content isn't a good enough candidate for the searched keyword. It is a matter of ranking content based on factors (e.g. relevance) and the content that doesn't rank simply didn't made it. Most penalties however are at link level. They can only be computed by analyzing the link graph (coupled with some other relevant info) and detecting certain "features" that the link graph for a site has. This is to a large extent based on probability/statistics and also takes into consideration patterns (e.g. 90% of links are 301 redirects). It is possible to use machine learning algorithms to detect patterns, but I wouldn't be sure that this is how it's done. I won't go into details here but I can tell you that when you have the entire link graph (like google has) it is very easy to spot link networks or link schemes. Add to that extra information that Google has (e.g. host, registry), and it becomes a very easy job. You can read more about this from research papers. If you do that you will also realize how link schemes many of you come up with (at least many I have seen around here) can easily be detected and are unatural. What you have to keep in mind is that google has a pretty permissive threshold so it doesn't accidentally hit clean sites. Obviously, that still happens but in Google's eyes it is acceptable collateral damage. Most important thing everybody should keep in mind is that though Google COULD do something it doesn't mean it actually does it. Google can do a lot of things and can use many signals to figure out stuff, however sometimes doing that would be too computationally intensive or would result in penalizing too many legit sites. Also, the penalizing process happens at various levels and a penalty could be eliminated if a subsystem computes that some other factors are more important. For example, if you blast very fast a ton of comments and forum posts to a new site it gets penalized. Do the same to an established site and nothing happens, or you may even help it. Also, statistically speaking if there are 9 out of 10 signals that scream "blackhat" then you can guess what happens. If it's 1 signal out of 10 nothing happens because the statistical confidence is low to make a decision. Authority Link graph data determines domain and page authority. That in turn coupled with the content score for that query determines position in the SERP. That is an overly simplistic view at the problem but is that part that matters. No matter how well you optimize anchor text and content for a keyword, a domain that is trusted will outrank you even though that keyword only appears once on it's page. I have a site that ranks for "keyword 2011" (or whatever year we are in) just because i have "Copyright (c) 2010-2011. All rights reserved." in the footer. If I have 20 ******** links with anchor "read more" on pages about winkels, dinkels, splocks, shmoks, etc. on top trusted domains/pages, then on my page I have "Copyright 2010-2011" and a user searches for "keyword 2011" I will rank for that #1, before your site that got blasted with 50,000 links with ScrapeBox and Xrumer, with several variations of "keyword". Why? Because if a stranger on the street will say he will take care of your 5 months old baby you will beat the crap out of him. But if your friend from childhood will say that, you know you can trust him. Then again, if me a SEO guy get dressed in a white coat and come in your reserve at the hospital and say "I have bad news. We did this test and you have colon cancer." you will believe me. Just because I have a white coat. Blackhat is about putting on a white coat not about yelling louder "I am a doctor". Again, you can read more about this from research papers. Is PageRank important? Both are correct: There is no PageRank PageRank is the fundamental measure to rank a site What do I mean? There is a "rank" in the algorithm. That is what matters. It is derrived from many "ranks" combined using a complex algorythm (not simple addition but complex rules). It is not just a matter of adding 2 average links + 3 average links + 4 super weak links and getting the result 1. Also, that "2+3+4=1" that ExtraWinner presented is really misleading. I can look at it from multiple perspectives and it can mean for me many things. Same with the people that read it. Some only read it in one way. But Joe's way is not Mike's way. Also, how Google computes real PR is not even close to how ExtraWinner explained it. Not even close. It isn't even the same as described in the original PageRank paper, but I would guess that was the foundation for today's PageRank. The toolbar PageRank is not very important though. While it gives a good estimate in most cases, you can easily get to rank #1 for a solid keyword with a PR0 page. N0F0llow? Worse than useless. Worse? Yes. How is that? Well, Imagine that every time you have an orgasm somebody cuts a finger from your hand, then from your feet, then your ears, then your nose... now nobody wants to have sex with you anymore because you look like a monster from a bad zombie movie. WTF am I talking about? N0f0llow is worse than useless because it is dangerous. It is easy to build n0f0llow links (e.g. ScrapeBox), you build many links (=spam), you shit on somebody's site to do that, you get the sensation you're doing something because of the large numbers of links that you build, everybody tells you it's good, everybody does it. Result? #1 ranking in google? Maybe for some weak keyword. Just your site being flagged in Akismet as a spam source and a bunch of other nasty "penalties". If 10000 homeless dudes smiling, out of the blue vouch for this dude I don't know that he is a good babysitter, do I trust him with my baby? Not really. So you see, n0f0llow is not evil or good. No. By itself n0f0llow is a tool. Noobs make it worse than useful because how they use it. I wouldn't mind a n0f0llow link from Wikipedia but I would hate to have my site blasted to 50000 blogs with a generic comment and same keyword. Why? Because my site is too good for that. My site gets "real" links. My site is too good to look at your site. My site won't play with your site. Oh, btw, if you want a technical answer or results from tests of why n0f0llow doesn't help, I could give you that as well, but I won't. I just want to say that nothing is absolute. Not even n0f0llow. It is not good or bad. It CAN be good OR bad, good AND bad or neither. What is hot these days? Same things that was yesterday and will be tomorrow: Offering value! Presenting your shit in a nice package so that the value can easily be seen! What i recommend: Multi-site link exchange from paragraphs of text in the post. Your own PROPERLY BUILT link and content networks. Links from high authority sites. Guest posts. Contests / prizes. Game changers - radical stuff that will give you an edge. I have some. Do you, or are you just following the pack? What is hot as in popular: Some sort of crap that doesn't work too great if you have real competition, but is noob friendly. ExtraWinner recommends to submit to just one (or maybe top) article directories. I recommend to blast content everywhere. If you can post some content with a link on even the shittiest sites on the internet, do it. Why? Because domain, IP and C-class diversity is one of the main factors that improve the trust rank. Then again, content is expensive. Or is it? Spin properly and in a year, with 4h of work a day for spinning, you could produce 150,000 articles. I'm not saying you will actually be able to do that. I am not even saying you know how to spin properly. I'm saying it is doable. Now, the really big problem is actualy finding 150,000 different hosts to post that content on in order to maximize the effect of the links. Is Google smart? Yeah, both the people and the software. After all, the software manages to beat some of you, doesn't it? ExtraWinner, google guys may be on BHW to check latest noob trends. They won't get pissed though, you can be sure of that. I am pretty sure figuring out how to cope with the really blackhat and intrusive techniques that the organized crime cartels use, is higher on their priority list than what script kiddies do. By the way I am not implying all BHW members are noobs. By my measure most are, but there are also some that could easily make me look like a noob. Too few and very silent, as they should be. Perfect website? ExtraWinner said "having around 10 links on each page. External links to authority sites that are anyhow under same category.". Ever heard of the term "SEO black hole"? YouTube is a "SEO black hole". All big sites are "SEO black holes". Linking out to authority sites makes those sites even bigger authorities. Instead do your best to get some links from those sites, so that YOU become the authority. Also, did you pull those suggestions from your ass man? You did. You may think you didn't but you did. How do I know. There is no such thing as "perfect website". Having certain proportions, certain percentages for types of links, etc. WTF? You might have some sites that have those characteristics and are doing great, but that doesn't mean it is a generic "rule" or even an "in most cases" thing. A while ago some folks like you came up with a magic number for keyword weight on page at 4% or so. Then every other "expert" out there made it 3%, 5%, 10%, whatever, just so his number is "unique" lol. Not even somebody from Google would be able to say what the "perfect site" is. There is no such thing. Perfect site is a site with 1 million pages of unique content, 5 million links including about 50,000 links from top 10000 sites on the internet. I mean, wtf man, are you that much of an expert that you know how perfection looks? Do you even know how to measure content uniqueness? Do you know how Google measures it? There are many ways to do it. Which one should I use to make sure I have 80% unique content? Some big citations? Why not instead have other big ass sites quote my site and put a link in there towards my site? Accuracy of info? SEO now is not what it was 5 years ago. I know a guy who does SEO since 10 years ago and he sucks at it. I mean don't get me wrong, he is succesfull, he makes a lot of money, etc. but not because he's a great SEO but because he's a great business man. Also, my personal oppinion is that the information you presented is not only inaccurate but also misleading at a subtle level. Very misleading. Also incorrect, which is a different thing than accurate. There's one thing to miss a rabbit by 1 inch and there's another thing to not know there was no rabbit there in the first place. ExtraWinner, I see that you replied to some that didn't "liked" your post that you wanted to keep it simple and easy to understand. That is not an excuse for thin, misleading, incorrect information. The "noobs" you wanted to help apparently, don't get smarter from incomplete and incorrect information. In the case of SEO, incomplete can easily translate into "very wrong/incorrect" and on the long term affect your approach to SEO in a very bad way. While what you said might seem plausible, possible, might seem it makes sense, etc. the complete message you transmitted is wrong. The entire substrate message and in many cases factual data is incorrect. Also, in SEO you can't do accurate tests. As a matter of fact you probably get more accurate guestimates by just thinking like a SE engineer, throw at yourself the problems he would have and try to find solutions. If your a good coder with a solid foundation in critical thinking you might just pull it off. I recommend you to find the research papers on the topic you are interested in before you think you know how things work. After you will recover from the amazement, remember that Google implements it more effectivelly than it was outlined in that research paper. Anyway, this is my view on the "full insight" you offered. My thread doesn't have much insight. As a matter of fact it is quite "thin" and incomplete. I would say in the "SEO school" it is at kindergarden level. In the end, you choose in what you believe, regardless of the truth. It is best to research yourself, but most people prefer the eazyness of believing instead of the struggle of researching. ExtraWinner, I'm not trying to attack you as I don't know you and have nothing against you. However, I wanted to make some things clear.