Google's PageRank algorithm has been around since, well, as long as Google itself. Yet there still seems to be a lot of confusion about what it is, how it works, how important it is and how we can use data based on it.
While all of those things have changed to a degree over the years, as Google as refined and developed the way it does things, the basic facts have remained constant.
Before we begin diving into PageRank, however, we need to make an important distinction between two things:
- PageRank: This refers both to the algorithm itself and the score given to each page as a result of it.
- Toolbar PageRank: This unofficial term is the name usually given to the publicly displayed PageRank score. Officially, it's only available through the Google Toolbar (hence the name), but in reality many SEO tools provide it.
We'll look at these two things separately.
PageRank: What is it, Exactly?
PageRank is what made Google what it is today: a heuristic analysis of the web's link graph, or, in simpler terms, a concrete, mathematical way to ascribe importance to a web page based on the links that point to it from other web pages.
This is what initially made Google's results so successful compared to the nine other search engines that had come before it. Although they weren't the first to look at links, they were the first to do it in such a meaningful way.
I won't go into great detail on exactly how it works, but if you're interested, Matt Cutts has a decent bit of information about it as the intro to his post about why PageRank sculpting no longer works.
The algorithm itself is described in Brin and Page's seminal paper, "Anatomy of a Large-Scale Hypertextual Search Engine." In other words, it's public knowledge.
However, it is also clear that PageRank has evolved a lot over the last 10 years. The basic structure is probably pretty much the same as it has ever been, but it has become much more sophisticated in many respects (for example, introducing an additional factor for detecting and valuing links based on where they are on a page, as well as on the strength of the page).
For this reason, PageRank as presented in the original paper is usually referred to as pure or classic PageRank, as opposed to its modern counterpart.
It's also clear that Google now uses many other factors to produce the search results that you see every day. Some of them, as with PageRank, will be based solely on links. Others will be based on all sorts of other factors.
Google doesn't reveal how important PageRank is as a part of this total mix, but while its importance has clearly diminished over time, it is still an important part of the total ranking algorithm.
But now that we've discussed the PageRank algorithm itself, what about the number that is derived from it?
Google is also suitably vague about this (and, remember, they don't reveal it either: they only reveal the Toolbar PageRank), but, in essence, it is a logarithmic scale (like the Richter scale for measuring earthquakes) with a base of about 16.
So, that means that a PageRank of 2 is 16 times bigger than a PageRank of 1; a PageRank of 3 is 16 times better than a PageRank of 2, and so on. Just to give you an idea of some scales here:
- PageRank 4 is 4,096 times better than PageRank 1
- PageRank 5 is 65,536 times better than PageRank 1
- PageRank 5.1 86,475 times better than PageRank 1
- PageRank 7 is 4,294,967,296 (over 4 billion) times better than PageRank 1
So, hidden within that deceptively simple scale from zero to 10 is actually a massive range of importance. And this brings us nicely onto Toolbar PageRank.
Is Toolbar PageRank Just the Number From the PageRank Algorithm?
No. Google would like you to believe that this is the case, but most SEO professionals believe that there are other ranking factors from different algorithms mixed in with this number.
However, the main problem isn't whether it is or isn't just PageRank, but the fact that there is precious little information about whether it is or isn't. There are also a number of other problems with it:
- It's only reported to one significant figure. It should be obvious from the above numerical examples that even the difference between PageRanks 5 and 5.01 is far bigger than the difference between PageRanks 4 and 5. So, the single figure that we're given bears remarkably little information.
- Even ignoring that is it/isn't it PageRank debate, the number given by Google is probably not an accurate reflection of a page's PageRank anyway.
- It's only updated for each page three or fours times per year. In contrast, Google's internal representation of a page's PageRank is probably updated almost constantly.
It should be clear by now that although the PageRank algorithm itself is important, Toolbar PageRank is next to useless as an SEO metric. I would recommend not using it at all as a measure of an SEO campaign's performance (what you should be measuring we can discuss another time!).
It does have one worthwhile use, however, which is looking for manual penalties (although even in this context it should be used with caution). If you are monitoring the rank of a page over time and it suddenly drops, that is a good indication that a penalty has been applied.
Likewise, if you're comparing a page's PageRank with its equivalent mozRank (an indepedent figure that guesses at what a page's PageRank really is) and the PageRank is markedly lower than the mozRank, that may also indicate a manual penalty. This can be useful when selecting link building partners.
It can also have some use for analyzing how Google perceives the importance of pages within a single site by comparing PageRanks across the domain and for looking for missed opportunities (i.e., if you have a page with a high PageRank but no traffic, it probably means that you need to work on incoming anchor text and on-page optimization for your chosen keywords). However, I wouldn't recommend using it to compare page across different domains.
Busting Some Myths
And lastly, lets finish off by dispelling some common myths and misconceptions around PageRank:
- PageRank acts like a vote for one page by another. It's a great description for lay people, but it really doesn't work that way -- and is actually quite misleading. If you are, or want to be, an SEO professional, it's really worth understanding PageRank properly: read Matt Cutt's post that I linked above (or read the original paper). A good analogy would be that is calculates the probability that a web surfer will end up on any given page if he or she follows links around the web at random. One immediate impact from this is that linking out actually reduces a page's PageRank. It also shows us that a link from a page with many incoming links is better from a page that has just a few.
- Sites have PageRank. No, they don't! PageRank applies just to individual pages, not to sites as a whole. People often refer to the PageRank of their site, but what they are actually referring to (unwittingly or otherwise) is the PageRank of their site's home page. It is thought that Google does have another link based algorithm, often referred to as DomainRank, that works in exactly the same way as PageRank, but at the domain level, not the page level. But that's a different story. It's also worth noting that home pages tend to have higher PageRanks than pages within a site, as they tend to have more links pointing to them.
- PageRank relates directly to traffic. It doesn't. PageRank is an indicator of a page's latent ability to rank, but says nothing about how it is actually ranking or for what terms it is ranking. Think of it as a potential. For actual rankings to occur, Google also has to associate a given page with one or more keywords, however strongly, and it uses entirely different algorithms to do that (both link based and otherwise). So, it is not uncommon to see, for example, a PageRank 3 page getting orders of magnitude more traffic than a PageRank 5 page, because the former is strongly associated with some popular keywords, and the latter is potentially not strongly associated with anything. However, it's reasonable to assume that between two pages equally associated with the same keywords, the one with the stronger PageRank would rank higher in the results (although all the usual caveats in the information given above still apply).
- PageRank shows how important Google thinks your site is. No. PageRank is just one of many factors that Google employs for indexing sites and ranking them.
- PageRank is Google's only link-related algorithm. Again, false -- although when Google first started this was probably true. They now use calculations based on links for all sorts of other things to do with indexation, ranking, and keyword association.
Having at least a decent working knowledge of PageRank is vital to the working SEO, both in terms of helping with link building and understanding the distribution of PageRank scores around the web. However, as an SEO metric, outside of some very narrow circumstances, Toolbar PageRank is almost completely worthless and should be approached with extreme caution.
The Original Search Marketing Event is Back!
SES Denver (Oct 16) offers an intense day of learning all the critical aspects of search engine optimization (SEO) and paid search advertising (PPC). The mission of SES remains the same as it did from the start - to help you master being found on search engines. Early Bird rates available through Sept 12. Register today!