iGB Affiliate 43 Feb/March 2014 | Page 32

TRAFFIC now be punished (at page level) where the anchor diversity had an abnormally low portion of ‘other’ types of links – i.e. not containing the brand, or keyword and nontransactional link text, often a call-to-action. We always have to keep in mind that in an increasingly variable environment where we are unable to isolate all the variables, much of the data, which ties together cause and effect, is merely circumstantial evidence, but over time, when the same cause and effect appears to result in the same symptom, we become more confident in our theories, and this confidence drives our strategy. So we have identified two ways for Google to find and punish spammers, and by applying these filters using different thresholds related to the norms within a SERP, Google now had a way to punish foul play based on entirely diffe rent, acceptable norms. These vary from SERP to SERP. Given that some SERPs are a cesspool, it was once again game-on for the spammers, simply a matter of spamming better than the competition (which in the world of SEO means: ‘business as usual’). 1. Those who believe rumours, propaganda, chatter and gossip 2. Those who analyse the data and make decision based on facts I had made it a personal policy to stay away from the forums, and make decisions based on facts. Penguin, in all its iterations, appears to be a simple incremental tightening of the thumb screw formerly known as May Day (with the addition of Disavow data), and since there have been a few iterations, we have to assume it will continue to tighten. We can see from Figure 1 that the latest iteration of this tightening of the threshold occurred on Jan 24 2014. Figure 1 Google leaves trace evidence for forensic seos to investigate The wonderful thing about an algorithm update is that a significant change provides those observing with a large volume of results which could be analysed and accurately interpreted. One way to do this is to group those sites which suffered and those that benefited and understand the common denominators of each group. I have no doubt that this was the reason Google had already begun making multiple changes simultaneously making it harder to isolate the variables and draw solid conclusions. By the time Penguin was rolled out, it seemed as if Google had also involved its press office (I was hearing about their shiny new algorithm on the mainstream media). Since there was now an increasing number of Big Data-aggregating-forensicSEOs, new methods of dispersion and misdirection were required. Propaganda was driving an increasing number of reactionary SEO decisions, which in turn was messing with the data and making it even harder still to isolate the variables. There are two types of SEO: 32 iGB Affiliate FEBRUARY/MARCH 2014 Figure 2 shows one of the many domains impacted on January 24 (casinoaction.com for the phrase: play casino games). As you can see, there is a increase in volatility which indicates some harsh punishment is being dished out, while the increased thickness of the cyan band indicates an abnormally large number of sites are being punished. Figures 3 and 4 show another site to be affected: oddschecker.com for the phrases ‘betting online’ and ‘online betting’. We have some compelling yet circumstantial evidence of an algorithm update on January 24, 2014, however, let’s look at the simplified anchor profile data, since we can’t rule this out as part of a wave Source: http://my.reach.ai Figure 2 Figure 3 Source: http://my.reach.ai - Casinoaction.com Source: http://my.reach.ai – Oddschecker.com – ‘betting online’