A Ghost in the Machine
Why your analytics may not be what it seems.
Google Analytics is a an excellent tool and provides many small business owners and technology professionals with data that pinpoints just what people are doing on your website, where they are coming from, how long they are spending there, and what they do when they get there.
Taking a step back, if you have a website and do not have Analytics, get it, it’s free and will provide invaluable insights into your users’ habits.
Like most small business owners the new tool can be a either all engrossing, or just too hard, if you are in the all engrossing sphere you will constantly check it but rarely delve deeper than the main Audience Overview.
A great tool, it gives you a snapshot of what is going on and just how fabulously your site is performing, or how it may not be. Like most great representations of data the devil is in the detail and of late there is most definitely a devil, or should I say ghost lurking in your analytics.
What you should be looking at
The Overview gives you the basic stats; how many sessions, how many users, page-views, session duration, bounce rate and new sessions. It is a nice quick overview of how your site has been performing over the last month.
There are other tools within this page for comparison to previous dates , previous years and the left hand column gives even further breakdowns on demographics, behaviour, technology and numerous others.
What most people focus on is how many actual users came to your site, it’s the first figure people mention and the only figure most business owners look at, and here lies the problem.
To actually find out where the users are coming from you need to move a bit further down the page and look at the Acquisition tab in the left hand column.
Cool stats and it all still looks fab lots of organic search, plenty of people going directly to the site, a nice number of sites referring traffic to you, and a few bits and bobs from social.
But where are those referrals coming from? Do you have links on other sites? Do your clients post links?. In this case referrals make up a large part of the total users 14.8%, in some cases it can be over 50%, and has an abnormally high bounce rate. Something is definitely up.
One more click and we can see where these awesome referrals are coming from, and now they do not appear so awesome. The first two on the list are from best-seo-software, and rankings-analytics.com, sites you have never heard of or had anything to do with.
These are what are known as ghost referrals and they can from a very large part of the traffic being reported as users in analytics. Put simply these are spurious links that generate traffic. It involves making repeated requests using a fake referer URL to the site the spammer wishes to advertise, but from your end it looks like a genuine user.
Why do they do it? Sites that then inadvertently link back to the spammer’s site add weight to their SEO, the business model is based on link building farms. They serve no good for the web, and are a massive headache for genuine businesses.
This is what Google has to say about it
“A referrer is a simple HTTP header that’s passed along when a browser goes from one page to another page, normally used to indicate where a user’s coming from. But users can change it, and some people will set referrer at pages they want to promote and visit tons of people around the web — people see it and say ‘Oh, I should check it out’. It’s not necessarily a link… there are some people who try to drive traffic by visiting a ton of websites with an automated script and setting the referrer to be the URL they want to promote… there’s no ‘authentication’… You can’t automatically assume that it was the owner of the URL if you see something showing up in your dashboard. Somebody is trying to do some hijinx.”
– Matt Cutts, Head of Google Webspam Team
What to do about them? Firstly go to your analytics, head to Acquisition, then Channels, then click on Referral. Spam sites generally have links listed like best-seo-software, buttons-for-website.com and a myriad of others, the bounce rate will at 100% or close to it.
So yes now you have discovered all those lovely users are not actually real. How to fix it?
If you are not comfortable working with Analytics it may be best to employ a web professional to fix the problem. They can add filters based on the referring site urls and eliminate most of the problem but you will have to keep an eye on it.
Most sites can be updated within an hour
There are some excellent articles on the web on how to fix the problem including this one by George Gergiev here
Originally published at – http://www.businessblogshub.com/2015/11/a-ghost-in-the-machine/