For the last decade or more the SEO community has been obsessed with backlinks. Whether it be filling out profiles to get a quick link such as the famed Angela and Pauls packages or automated tools such as Xrumer and SENuke. Blog commenting, shout boxes, wordpress templates, sql injection, website widgets are just some of the methods which have successfully shot websites to the top of the search results.
In August 2009 Google announced it was working on new architecture to run its web crawl and search processes. This was coined The Caffine Update and was rolled out in the summer of 2010. There is an excellent post on the google blog that goes into basic detail about how caffine works.
Since the Caffine update Google has been implementing tweeks to its crawl, index and search programmes on a daily basis. By January 2011 they had rolled out over 200 update/tweeks to the algorithm.
Caffine has made it simple for Google to discredit backlinks from commonly targeted platforms. As soon as they see a site ranking somewhere it shouldn’t be they can look at why and manipulate their algorithm accordinly. It seems they can react in this way at quite a pace too from what we have seen this year.
So worthwhile backlinks have become harder to get. In our opinion they still hold weight and affect the serps but not to the same degree they once did. In our experience and that of others social networking has come of age as a quality indicator to the major search engines.
Google is openly counting urls in tweets, retweets, facebook likes, facebook shares/mentions and these seem to be one of, if not the, most influencing factor in the SERP’s.
Time to google some competitive terms and see what is going on.
From this you can see that the site isn’t using backlinks to get on to the first page for the highly competitive term “cialis”, not directly at least. There is a chance they are using redirects to flow link juice from other sites. If you google the site there is some evidence of automated backlinking (possibly xrumer/drip feed blasts). We double checked with Yahoo site explorer and the entire site, excluding this domain reports 216 backlinks. This is far from the hundreds of thousands that we used to see a year ago for these types of results. Additionally it has a dmoz listing and a pagerank of zero!
Googling the following dork returned 480 low quality results:-
“http://www.yewtree.bham.sch.uk cialis -site:www.yewtree.bham.sch.”
We believe sites like this are ranking due to the high trust rating Google is giving to the domain as a whole. It is a uk school tld and has built up a history of trust with Google. Whoever has accessed this and used it to redirect to a pharmacy has taken advantage of this trust and manipulated the onsite and offsite content to rank for cialis.
We thought they were probably using a cloaking script to show a high quality article to the spiders filled with lsi (related) terms. We knocked up a quick script to imitate a googlebot spider and see what the site returned. Interestingly only the description seemed modified although this cannot be the case as the title we were sent back was “Welcome to Yew Tree Community School” and the title in the results page was different. So there has to be some cloaking going on.
So we suspected that they may be using social networking to spam the search engines…
http://search.twitter.com/search?q=www.yewtree.bham.sch.uk – Returns no results…
A facebook search turned up nothing either except for some bing search results. Click through from here and it shows the original school website with the meta description.
<meta name=”description” content=”Joomla! – the dynamic portal engine and content management system” />
Can you guess what cms has a zero day exploit in the wild???
So there is no evidence of manipulation of the social networks but definitely some cloaking going on particularly targeting Googlebot.
So has domain trust become one of the biggest ranking factors in Google? So much so that nothing external is required?
Does the site use redirects to flow juice from external sites?
Does anyone have a better googlebot imitator that can get the the page that google is receiving? (forge Browser::BROWSER_GOOGLEBOT perhaps)
Does anyone have any other ideas we haven’t gone through here as to how sites like this are ranking?
The big takeaway from this for white hat seos is the influence of domain trust. High quality aged domains and websites may well hold a higher value than ever for organic search. It certainly has a great power or the black hats wouldn’t go to the trouble of hacking kids schools websites. Google does appear to be onto this, note the “this site may be comprimised” note in the serps, probably for the cloaking. I wonder if Google is using a programme like selenium on a large scale to test for cloaking.