When I saw Tom from Distinctly Digital had taken a look at the [web design Hertfordshire] SERP in an attempt to see why the sites were ranking as they did I couldn’t resist having a dig myself.
After a quick glance at the backlink profiles of the top ranking sites three main points jumped out at me:
- The lack of variation in links pointing to the ranking sites
- The strong performance of the #1 site given its anchor text profile suggests it is massively over-optimised, and
- As a commenter raised in Tom’s thread, the apparent poor performance of AWDltd.co.uk relative to its link metrics
The tools I have used for this analysis are Majestic SEO, CognitiveSEO and Search Metrics.
But first –
Social is a Red Herring
The first “ranking factor” Tom considers is the number of social shares received by the top-ranking web design companies in Hertfordshire.
He produced a bar chart of social shares per site highlighting the trend line which decreases from #1 – #10.
“We can therefore assume that for the keyword ‘web design Hertfordshire’ the number of social shares is an important ranking factor.”
No!
Even ignoring the fact that Google have come out and explicitly said that they do not currently treat shares on Facebook and Twitter any differently from other links on the web, this should not be an assumption.
Just because a trend exists doesn’t mean it has any significance – especially when looking at a sample size of 1!
Princeton University recently published a study in which they stated that Facebook would lose “80 percent of its peak user base between 2015 and 2017.”
Their source data for this experiment – Google trends. Facebook’s Mike Develin debunked this following the same correlation must imply causation principle to imply that by 2021 Princeton would have no students and by 2060 the world would run out of air.
Correlation <> Causation!
Yes good content is far more likely to be shared than dross, but social shares do not directly result in increased rankings – yet! And, even if they did it would be just one factor of many.
Link-building in 2014
If I came to you with a link-building proposition for your clients – building footer links using exact match anchor text you’d rightly turn me down. That’d be before I got to the bit about many of these links being site-wide and on sites of unrelated theme.
Yet these are the same links web design companies are building to their own sites – for many of whom this makes up a large proportion of their total links.
It doesn’t take a genius to know these aren’t the type of links Google wants to view as “good” in its algorithm.
I wouldn’t want them in my link profile…
Large Scale Site-wide Exact Match Footer Links are against Google’s Guidelines
Shocker!
Google’s Link Schemes page clearly outlines this type of link as unnatural.
“Here are a few common examples of unnatural links that may violate our guidelines:
<snip>
- Widely distributed links in the footers or templates of various sites”
Although in further discussions both Matt Cutts and John Mueller (CTRL+F on “footer”) have been typically vague there is evidence to suggest that this can cause problems, particularly if commercial anchor text terms are used without variation.
But… Penguin Free Zone!
Incredibly there doesn’t seem to have been a big shake-up in the [web design hertfordshire] rankings when Penguin 1.0 rolled out (April 2012). Despite the fact that exact match links targeting this very phrase is the most common anchor text for some of the sites ranking today.
In-fact (though I am not pleased to see it!) check out this correlation*:
The same trend* occurs when looking at the proportion of total linking C-classes that use exact match [web design hertfordshire] anchor text.
So much for over-optimisation penalties! Perhaps Google has learnt this is the sector-norm?
In most other industries link profiles like this would surely have resulted in link warnings and penalisation.
*all the normal caveats apply, correlation <> causation!
Where things are heading
Google is only heading one way on spam – getting better at detecting it and tougher at dishing out penalties to combat it.
Of course there is always LOADS that they miss and often grey-hat techniques get hit when other sites appear to blatantly breach the webmaster guidelines and yet continue to rank.
Low quality links
In addition to links from client sites there is some crap in some of the top ranking backlink profiles but it seems that the only real impact Penguin had on the space was to hit a couple of sites with extremely large number of exact match links from lower quality, lower authority, free-for-all sites.
I mean come-on… 149 linking domains with Hertfordshire misspelled!
…from quality content content like this…
So whilst stuff like the above got wiped out in April 2012, the fact is today the space has sites ranking even with an unnaturally high proportion of exact match high-PR links. The top ranking site has 80% exact match links!
So long as the link sources aren’t linking out to too many other sites or bad neighbourhoods it seems that these techniques continue to be relatively safe in this space – for now.
Brand
Hopefully you will have noticed in the two anchor text charts above that the sites ranked 2nd, 3rd and 4th, did not have any exact match [web design hertfordshire] anchor text reported. This is not to say they aren’t using commercial anchor text links in their client footers, just different variations.
Classifying key generic terms the top 10 sites might be targeting as well as individual brand terms in Cognitive SEO produces this:
Popcorn Web Design, Fluid Studios LTD and Touch Creative all have a relatively high proportion (67%+) of branded links and – see below – all have steadily been improving their rank for the investigated term over the last couple of years.
Two stand-out sites
Two sites seem to stick out from the others as appearing particularly anomalous.
Advanced Web Design
The headline metrics for AWDltd.co.uk in Tom’s post appear to suggest that with its current rank of #7 the site is performing worse than it should compared to its peers – something called out in the comments by Michael Harrington.
The Search Metrics graph above tracking its ranking shows that for the key term being investigated from summer 2012 to spring 2013 Advanced Web Design were generally ranking 2nd, with a few fluctuations to third. Since spring 2013 however there has been a slow gradual decrease in rankings.
There is no sign of a manual spam action being applied to the site or big algorithmic hit so I suspect what has happened here is that over-time some of their inbound links have been removed or devalued by Google.
Also – Although the site is linked to using the target anchor text phrase 20 times, these sites are only hosted on 7 different C-class IP addresses. This is actually fewer than the sites ranking in the positions around it…
Shy Studios
Hertford-based Shy Studios have held the top spot for [web design hertfordshire] for the past two years.
Almost certainly over-optimised – buy hey, it ranks.
No-one knows why the sites rank as they do…
…and that’s what makes SEO so interesting.
and of course the rankings have changed since Tom looked at the top-10 ranking sites… Google doesn’t sleep.
—
Image credits: Red Herring (Tim Parkinson) CC BY 2.0, Extrapolating XKCD, Majestic SEO, Search Metrics.
Great post Phillip. Completely agree with you re sample size and correlation/causation – my bad 🙂
What tool did you use to chart the historic rank positions?
With pretty much everyone in this space having a high proportion of exact match anchors, it will be interesting to see what Google promotes when it starts cracking down.
Don’t worry it’s just a pet hate of mine – especially when it comes to people equating social shares to better rankings!
Historic ranking data is from Search Metrics’ Essentials package http://suite.searchmetrics.com/en/research
It’s as if normal SEO rules don’t apply to this keyword!
Number 3 is currently: http://www.digital-results.com/web-design-hertfordshire/
Wrong in so many ways!
This reminds me a bit of a ‘payday loans’ case study that showed everyone ranking in the top 10 had some kind of penalty. Assuming there is a relatively small amount of sites trying to rank for this keyword and that the vast majority of them are over optimised, its possible Google is treating this as the norm and is ranking the site(s) it still believes to be most relevant to the search.