Google is a notoriously secretive company. Whether its new products, changes to the service or potential buyouts, Google has always kept its cards close to its chest. This is never truer than when considering the algorithm which drives the search engine rankings.
For a company that employs around 50,000 people across the world, it has been said that fewer than ten of these know exactly how websites are ranked and indexed. Of course, Google is unlikely to make public exactly what its algorithm contains, as trying to find it out is big business. Figures from Econsultancy suggest that, back in 2011, the SEO industry was worth £591 million in the UK alone. This represented a year-on-year rise of 18 per cent, so if that upward trajectory has continued, it can today be estimated to have a value of nearer £971 million.
Given these figures, it should come as no surprise that Google is unwilling to reveal just what goes on behind the scenes of its algorithm. After all, it can then utilise paid search to provide companies the chance to reach higher up the SERPs through its Adwords platform.
Despite all this, however, still a little information manages to trickle down to the world of SEO. Of course the accuracy of what’s proffered cannot be proven, but it nevertheless offers SEO professionals a starting point from which to tackle their campaigns. It doesn’t just indicate what to do but (often most importantly) what not to do, as the ramifications for this can often be much more pronounced.
So with this in mind, let’s take stock of some information Google has drip fed down to see exactly what it is we know about the algorithm and what remains a complete mystery.
Google has to regularly update and tweak its algorithm to ensure it remains up to date. Whether this is to mirror changing search trends, keep abreast of new tech or to try and eradicate villainous back-door ‘black hat’ SEO techniques, it needs to ensure the algorithm is agile and ahead of the market.
Exactly how often this happens, though, is anyone’s guess, with estimates ranging from two to 2,000 times a year. Strangely, both of these could be correct, albeit in very different ways – all depending on the definition of ‘update’.
Google is universally viewed to be tweaking its algorithm, which means it could well undertake 2,000 (or maybe even more) every single year. These tiny tweaks are unlikely to have a noticeable impact on many rankings, though, but instead combine in order to make Google the best of its kind on the market.
On the other hand, large-scale updates such as Panda, Penguin and Hummingbird have been implemented in the past few years which had a much more pronounced impact. They generated a great deal of interest ahead of the roll outs, then again when sites either scaled the peak of Google’s results pages or crashed out altogether.
This isn’t all, as Google is also known to “refresh” certain parts of its algorithm, which doesn’t usually introduce any brand new aspects but instead tweaks existing ones to ensure they remain right up to date. Similarly, the larger updates also get a refresh, with these following the typical setup of decimals indicating just how major the change was (such as Penguin 1.1 being a small upgrade, compared to Penguin 2,0, which was decidedly more revolutionary).
The official word from Google tells us a little, but certainly not everything. It claimed that, in 2010 alone, 13,311 precision evaluations were carried out to test the impact of algorithm updates and see whether results were positive or negative. In addition, a further 8,157 ‘side-by-side’ experiments A/B tested results pages to see which was better in the eyes of Google users. Plus, 2,800 ‘click evaluations’ were also undertaken, to get an idea of how users would react to certain amendments.
As a result of all this, Google said it launched 516 “improvements to search” during 2010. Of course, these improvements are not on the route to an end destination at which point Google can sit back and know that everything is done. It will instead continue rolling out updates to keep ahead of the curve, so these improvements are roundly expected to increase year-on-year, not steadily decline. Proof of this comes from figures revealed about updates in 2012. Back then, some 18,812 precision evaluations were carried out, as well as 10,391 side-by-side experiments and 7,018 live traffic experiments, all of which led to the launch of 665 improvements.
So the number of updates that Google rolls out may now be covered, but how many ranking factors are there to monitor and test in the first place? Google’s official line is that there are “over 200”, but again an exact number still eludes us. Of this “200”, some have been explicitly revealed by Google, whereas others remain estimates based solely on activities which garnered success in the real world.
What makes it doubly-difficult to second guess Google’s algorithm is the fact that each ranking factor has a different weighting in terms of importance – and therefore, impact. To put it another way, Google’s head of webspam Matt Cutts has specifically acknowledged that domain age does indeed come into play where ranking factors are concerned. The sickener came, though, when he admitted that domain age may not be particularly important, especially where the youngest ones are concerned. “The difference between a domain that’s six months old versus one year old is really not that big at all,” he teased.
In short, this means that even if SEO professionals stick rigidly with only those factors that have been recognised by Google, it may not work out particularly beneficial. This is because these officially acknowledged ranking factors may have less impact on overall SERP placing than some others which have not been admitted by Google and are instead born more out of speculation.
Ingenious SEO professionals have used their initiative to find out ranking factors that Google is not keen to disclose. One example is the use of patents to see what’s hidden up its sleeve.
Google – just like any other business – endeavours to keep a competitive advantage, so is regularly filing patents to ensure the likes of Bing or Yahoo! cannot replicate its search functions. One such example was domain registration length, which detailed those sites which had “squatted” on a URL, compared to those which get more regular use. This means Google could use domain expiry date as a ranking factor.
It’s worth remembering, however, that this is only a patent submission, so whether it’s something Google has actively implemented remains to be seen. After all, it’s not an alien concept for companies to submit patent applications solely to prevent their competitors from doing the same.
For those wanting an exhaustive list, there are numerous sources dotted around the web which purport to provide 200 (or more) ranking factors. These should be approached with caution, though, as around 90 per cent is speculation and even those which have been acknowledged by Google may actually have minimal impact in the grand scheme of things.
It’s commonly accepted that knowing what not to do can be just as (if not more) important than understanding what to do. Rising up the Google rankings can be a long and intensive process in which a great deal of time and effort is rewarded by a jump of just one or two places. Conversely, undertaking activities which contravenes Google’s rules and results in a penalty can result in being dropped like a stone down into the depths of Google’s pages – or even not being indexed by them at all.
One of the more recent cases which hit the headlines was that of Interflora, which Google penalised by effectively eradicating it from all search queries, even those where users actually searched for the brand name, ‘Interflora’. This, it was said, was down to the retailer manipulating links in order to artificially improve its search ranking through paid advertorials.
Interflora was alleged to have paid newspapers to publish advertorials on their sites, complete with links back to interflora.com. Some years ago, of course, this would have been viewed as positive link building, as connections were established that users could happily follow between two reputable websites. More recently, though, Google has said this is not beneficial but instead rather detrimental, as the by the very nature of being paid for, these links eradicate any credibility that such sites may possess.
In the resulting aftermath, Google reiterated its stance that such links should be ‘nofollowed’, so that Google’s crawlers don’t come across it and, therefore, don’t end up penalising the site as a result.
Of course, there is a grey area around penalties in a similar way to that of ranking factors. Few people are acutely aware of every single issue which could result in a penalty, not to mention the importance weighting that each issue is given and specifically how much rope a site is given before it hangs.
On top of that, there are questions around whether the likes of Interflora were made an example of, simply because they are more well known. Other sites which had done the same, some have argued, may not have experienced Google’s full force.
This poses quite the teaser, with two clear paths. Option one is to stick rigidly with Google’s rules and be safe in the knowledge that no penalties are on their way. This is the safest option, but also that which is likely to require the most work, ingenuity and more than a little luck to beat the competition to Google’s pinnacle. This also requires webmasters to keep on their toes for the foreseeable future, as Google can (and has) changed its algorithm so an activity which was once recommended is now worthy of a penalty.
Option two, meanwhile, involves living a little more on the edge and straying toward what Google deems as being ‘negative SEO’. This is likely to be a quicker way of achieving success, but also a much riskier strategy. If it fails – and bear in mind Google is very adept at finding out bad practice – it can take even more effort to get out of the penalty quagmire than it might have done to stay on the right side of the law. Not only that, simply getting back to the old position a site occupied before the penalty can take a very long time, so building upon this further will surely take more time in the long run.
With such uncertainty around the issue of penalties, though, option one is unquestionably the safest. Google is not only continually moving the goalposts, but also endeavours to only hint that it’s done so and drip feed information on where they’ve been shifted to. It’s like a tech version of Battleships, which means that staying on the right side of the law is difficult enough already, without trying to crawl ever-nearer the edge.
This article may well have posed more questions than it answered, but such is the nature of the beast when dealing with a company as secretive as Google. It can’t be blamed, of course, as there’s a very good reason for feeding a little – but certainly not all – information down through to its users.
Either way, one certainty is that if educated guesses were currency, Google’s algorithm would almost be laid entirely bare. Professionals who have worked in the SEO industry for many years have experience with countless campaigns across all manner of different industries and verticals. These people are also some of the most ready to buy into the world of open-source technology, where information is shared freely between interested parties for the betterment of all involved.
As such, the information may only be estimates, guesses and straw poll results, but they come from valuable, quality sources. With this in mind, we may actually know more about Google’s algorithm than first thought.