Co-occurrence ? - or something else - A Google Analytics Case Study
These days, you need to have your eyes glued open with the amount of SEO content swishing around the internet. It is seems highly likely that by blinking, you might actually miss critical 'need to know' information. Each week, there are a plethora of 'must reads' - and sometime back in November (in case you missed it), 'the must reads' included these three posts.
- Rand's post regarding anchor text is weaking post
- Bill Slawski's rebutal on Co-Citation Observations
- Joshua Giardino - It’s Not Co-Citation but it’s still awesome!
I have to cast my mind quite far back to remember another time when one topic created so much discussion within the SEO community. It had most of SEO's pundits coming out of the wood work to give their thoughts and opinions. However, it also had me reflecting on something I've been seeing for a while, but have never bothered to figure out why it was happening.
The question that literally had me up all night was:
Why does Google analytics rank at position #2 (at time of writing) for the term 'ga' in the UK?
There was no logical reason why it should be doing so well, especially for such a generic term. Or was there?
To figure out why this might be happening, I started first with the basics.
If visit http://www.google.co.uk/analytics/ the title tag is displayed as
<title>Google Analytics Official Website – Web Analytics & Reporting – Google Analytics.</title> Clearly, 'ga' does not appear in the title tags
Nowhere on the landing page is Google Analytics referred to as 'ga'.
Perhaps there are a number of incoming links pointing to http://www.google.co.uk/analytics/ that have 'ga' as the anchor text. To check this, I fired up all three link analysis tools below.
|Tool||Anchor||How many sites found Linking (at time of writing)|
From the three tools I used, only Ahrefs.com found a website that linked to google.co.uk/analytics/ with the term 'ga'. However, one link shouldn't be enough to make Google Analytics rank for a generic term like 'ga'.
Could Google be simply using search data to make the connection?
So I decided to look at places where related terms can be found on Google
- Google related search
- Google insights
- Using the tilde operator
- Google AdWords tool
- Google suggest
1. Related Search
2. Google Insights
Google Insights also do not show a relation between 'ga' and Google analytics. This holds true even if you enter 'Google Analytics' in the search bar.
3. The Tilde Operator
Interestingly, Google does show that GA and Google analytics are used interchangeably.
4. Google Keyword Tool
Searching for keywords or phrases related to 'ga' on https://adwords.google.com/o/KeywordTool revealed that both terms do not show as related to each other. I also could not find any association after examining every result individually.
|Term||Top 10 Related terms|
time in ga
big fish ga
google web analytics
wordpress google analytics
what is google analytics
google analytics login
google analytics wordpress
google analytics for wordpress
joomla google analytics
google analytics tracking code
google analytics code
5. Looking at data from Ubbersuggest's Google Suggestion tool
The term 'ga' and google anayaltics does not seem to appear as part of the same search term.
Could there be a co-occurence of GA and Google Analytics within documents ?
To find out, I looked for pages where Google Analytics is mentioned and the word 'ga' also occurs. Some simple operators that let me test this are listed below
|intitle:google analytics intitle:ga|
|url:ga intitle:google analytics|
|intext:ga -intitle:google analytics|
What I've found is that if you plug them into Google, you can see that there are many cases where ga or GA and Google Analytics are used interchangeably or mentioned together.
Could this be the cause of the 'anomaly' ?
Takeaways and summary
If you cannot get a link within a document, is it ok not too worry that much about it as Google can connect two words if they occur together enough times across the web? For example, if your brand receives a lot of online press, just incase, you should also try to ensure your target terms are always mentioned.
Secondly if this is the case, we might need to redefine how a competitor analysis is conducted. If Google is looking at co-occurrences, it is possible that this is already factored into your competitors ranking. By only analysing links for ranking insight, we could be overlooking how terms co-occur on non-linking documents.