Note from Darren Slatten (SEO Mofo and World’s Best SEO):
What’s up, Mozzers?! At one point or another, most of you have probably asked yourselves something like: “How did Darren Slatten become the World’s best SEO? Where did it all start?” Well, ladies and gentlemen, the author of this post, Fernando Chavez, was my mentor and direct supervisor over at…um…some big SEO company. When he started training me, I had never even heard of SEO…but now look at me: simply incredible. You might say he is to Obi-Wan what I am to Darth Vader. Did I eventually surpass his skills and knowledge of SEO? Most people would say “no, Darren, no you did NOT,” but in MY mind…I totally did.
In any case, I’m very excited to see Fernando finally break out of the “big SEO company” setting and start his own company. He has already started putting up some great content on his website, and this article in particular is brilliant, if you ask me. The ranking model he describes is one that I strongly believe in and have always relied on, but I was never willing to share it with anyone (mostly because it’s so difficult to explain). Since Fernando was determined to publish it, I figured I might as well contribute to it and try to steal some of his glory. My contribution is the sweet diagram accompanying this article. I recommend you print it out or have it visible on your screen when you read this. The superscript numbers in the article reference the various elements in the drawing. This should help you better understand the logic behind this ranking model. So without further ado, I give you…The World’s 2nd-Greatest SEO.
It’s pretty obvious that no one will ever be able to crack Google’s overall ranking algorithm. In fact, Google engineers probably couldn’t tell you exactly why one site is ranking where it is for a particular query without researching that site like any SEO would. And I don’t think they know exactly how algorithm tweaks will affect rankings without testing them first either. The point is that there is no one person out there who knows exactly how Google’s rankings are calculated, so it’s useless to try to understand the algorithm at that level.
Despite the fact that we can’t crack Google’s ranking algorithm completely, it is possible to break it down into the parts that really matter for search engine optimization. Over the past few years I’ve developed a very simple way to think about how Google might calculate rankings. Whenever I am asked a question about SEO, I go back to my simplified algorithm and determine how the issue at hand might affect the relevance calculation for the targeted keyword. If it doesn’t affect the landing page’s overall relevance, then I say confidently that it’s probably not worth worrying about.
The first point that I need to make before explaining this ranking calculation is that PageRank is not a direct factor in Google’s calculation of a documents’ relevance score. In other words, it’s possible for a page with higher PageRank to be less relevant to a given keyword than a page with lower PageRank. This is why you’ll often find sites with lower PR outranking sites with higher PR. It is not uncommon to see this because PageRank is based solely on how pages link to each other and is independent of keywords. On-page factors and link relevance are far more influential on rankings than PR.
I do not make this point about PageRank to say that you shouldn’t try to increase your site’s overall link importance. You absolutely should. The more PR your site has, the more off-page SEO value your pages will pass to each other, which is obviously a good thing. I just want to be sure you understand that PR is based on links and not keyword usage.
Okay, so now that we have PR defined more clearly, we can move on to how rankings might be calculated by Google. At a fundamental level, the quality of a search engine’s results can be measured by their relevance to the given search query. Sure, we’ve all heard of metrics like trust and authority, but a very trustworthy, authoritative document doesn’t do you any good unless it’s relevant to the keywords you searched for. Therefore, the ranking model in this article is focused primarily on keyword relevance. First we’re going to divide the overall relevance (200) of a page (100) to a particular keyword into two subsets: on-page relevance (210), and off-page relevance (220).
On-page relevance is pretty easy to understand on a basic level. Although there are potentially dozens of factors that could come into play when optimizing a landing page, the truth is that using the targeted keyword in a page’s Title tag and body content will get you most of the on-page relevance that you could possibly obtain for that keyword. Don’t overthink it; just make sure your keyword phrase (and related phrases) is used prominently and naturally throughout the page (215).
Off-page relevance is more complicated, though it is still relatively simple to calculate in basic terms. The trick is to realize that every one of a page’s inbound links (300) has a relevance associated with it. The potential landing page’s total off-page relevance for a particular keyword is the summation of all these relevance values. Essentially, keyword relevance flows through links like PageRank does. So the more relevance a page has for a keyword, the more it can pass on to your landing page through a link.
The relevance value for each individual inbound link (300) to a landing page (100) is dependent on two things:
- How relevant the linking page is to your keyword (260).
- The PageRank of the linking page (270).
I believe these two metrics are multiplied somehow, giving us a combined value, which I will call the Link Relevance Product (250). Based on this assumption, you can see how if the overall relevance (260) of the linking page is zero (i.e., the linking page has nothing to do with your keyword), then the Link Relevance Product (250) of the link will be zero as well. This would be a good way for Google to negate the value of paid links between sites that have no keyword association.
Because PageRank is a factor in the link relevance calculation, you have to consider it when developing inbound links to your landing page (100). However, you should not be fixated on the Google Toolbar value. Obviously links from higher PageRank pages are better, but you should mostly be concerned about whether or not that page is in Google’s main index. If the linking page (300) is supplemental, then it’s probably not worth your time to pursue. Instead, spend your time pursuing links from pages in Google’s main index. Most SEOs do this already, but I wanted to stress this as a best practice, because if we assume that supplemental pages have little-to-no PageRank, then their Link Relevance Product (250) would be virtually nothing as well.
The off-page relevance (220) of a landing page can be further broken down into two subsets – relevance from internal links (230) and relevance from external links (240). Since it is so easy to manipulate internal link relevance, it is fair to assume that Google would want external links to have more influence on rankings. Therefore, we can reasonably assume that the off-page relevance values (240) coming from external links are given more weight (235) than those from internal links.
You could see how Google increasing the value of external off-page relevance (240) for the keyword would decrease the effectiveness of more sophisticated types of spam. For example, many people used to buy old sites with a moderate amount of PageRank for the sole purpose of putting up keyword-relevant pages that would link to the site they were optimizing. If you did this nowadays, then you might not see the rankings boost that you used to. This is because the pages you put up on the old domain will have a minimal amount of external off-page relevance (240) for the keywords you’re targeting. Why is that? Because that old site’s external links are going to be relevant to the original topic of the site, not the topic you changed it to after you bought it.
Please note that I am not talking about PageRank when I suggest that external links have more SEO value. Links are links when it comes to PR calculations. PR only factors into each inbound link’s relevance as part of the Link Relevance Product described above.
So to summarize, Google’s relevance calculation can be broken up into the following three subsets:
- On-page keyword relevance of intended landing page (210).
- Internal off-page keyword relevance (230), defined as the summation of the Link Relevance Product (PR multiplied by keyword relevance of linking page) for all internal links to the landing page.
- External off-page keyword relevance (240), defined as the summation of the Link Relevance Product (250) for all external links to the landing page. This subset is multiplied by a factor (235) so that it has more influence than the internal link relevance summation.
In theory, when a user queries Google, the pages that appear at the top of the search results are the pages that have the highest relevance score (200) for that keyword, based on the summation of these three subsets.
Google’s actual ranking calculation is obviously much more complex than what I have described here. However, I feel that focusing on these core ranking factors will always be a winning strategy. The amount of influence that any SEO changes have on your Google rankings will constantly vary based on algorithm tweaks. But changes that you make to influence the ranking factors I’ve described will theoretically increase relevance, which will always be the primary interest of search engines.
Once you understand this theoretical calculation, it’s much easier to see the reasoning behind the step-by-step SEO guide on SEOCipher.com. The three phases described in that guide correspond directly to the three relevance subsets described in this article.
