How to ensure your Google Analytics report is accurate
How often do you gaze inquisitively at your Analytics report? My guess is fairly regularly, but do you actually understand what you’re looking at? If the answer is no then this article is for you so keep on reading.
I speak to many business owners and experienced marketers working at agencies and companies of all sizes about the traffic to their website and more often than not they give me the facts and figures from the default dashboard of Analytics. Depending on a business’s visibility it can make for impressive reading when they say we have had several thousand visits this month.
The question to ask yourself or your marketing/SEO team is… “How accurate is that data really?”
The trouble with the default dashboard in Google Analytics is that is gives you all the data from the past 30 days. That may sound great at first but the Internet is a big place and your target market is likely a very small proportion of that. Therefore the types of questions you should have at the very outset are. Where is the traffic coming from? How good is the quality of the traffic? And what is the traffic doing when it arrives on our website? If you’re not currently getting the answers to these questions on a regular basis then you should be asking why not.
Let’s look at an example from one of our clients.
This shows the default dashboard in Google Analytics and some of the data collected over the past 30 days. It says we have had 2,660 visitors, 2,032 of which were unique. The pages viewed, bounce rate and time on site all look healthy so we could surmise that our client is doing well and traffic is increasing slowly.
Let’s dig a little deeper into this data. If we visit “Acquisition -> Overview” this provides a breakdown of the total traffic. Using the same example we see the 2,660 total visits come from the following sources.
- Direct: 957
- Organic Search: 813
- Referral: 764
- Social: 103
- (Other): 15
- Email: 8
The best place to start evaluating your traffic is from where your referrals are coming. Clicking into “Referrals” will bring up a list of the top 10 referrers of traffic to your website over the past 30 days. Do you recognise those sites? Do any jump out at you as being spam? If you’re unsure look at the key indicators such as time on site, bounce rate and pages viewed as this will give you a good idea of whether that traffic is genuine or not.
Looking at the client example we can see the vast amount of referral traffic is nothing but spam. In fact from the top ten referrers of traffic, only one is genuine, which I have blurred out for client confidentiality. The other nine and more besides are all spam. How do I know? Let’s take the first one as an example.
It has supposedly sent 393 visits to the website over the past 30 days and yet:
- The bounce yet is 99.75%
- The pages viewed is 1.00
- The average time on site is 00:00:01
Not exactly the kind of traffic we want to be including in our reports or analysis of how much traffic we’re getting to our website.
There are many ways to exclude the referral spam traffic to your website. You can block the traffic from ever reaching your website using your .htaccess file or have your hosting company block specific referrers at the server level. You can also set up filters in Analytics to exclude traffic using any number of pre-defined or custom filters.
One of the easiest and least-technical ways that you can do yourself is by using Google segments. Segments are just what they say, a way of segmenting up your traffic. We always share practical advice on this blog so have created a step by step guide for you to follow below.
How to exclude referral spam from your Analytics reports
Exercise: Exclude referral spam traffic
Visit “Acquisition -> Overview -> Referral”
Show all rows so you can see all referrers (bottom right hand corner)
Export the list of referrers to a CSV file and open in Excel. Remove those you recognise and others that appear genuine i.e. looking at time on site, bounce rate and pages viewed. If in doubt view the website of the referrer yourself to see if it is a genuine website and related to your business.
Next, add a “New Segment”, click “Add new” and give the Segment a name, for example “Exclude Referral Spam”
Under “Advanced” you will see a “Conditions”
For “Filter Sessions” choose “Source / Medium” and ensure the next box says “Contains”. Add each of your referring spam websites, choosing “OR” to add a new box beneath. Effectively you’re asking Analytics to filter website A or website B or website C etc.
Once finished and depending on how many referring spam websites you had in your report you will most likely have something like this list.
Save that Segment and now we want to view the data with the referral spam excluded. Click back to “Audience -> Overview” and then “Add Segment” once again only this time view “Custom” on the left hand side and select your new Segment. You will be presented with a report showing all visits in blue compared with the spam referrers you have segmented (removed) in orange.
Going back to my client example, I can see that referral spam was making up a sizeable 22.28% of the total traffic during the last 30 days. By excluding the referral spam we get a much clearer idea of the quality of the remaining traffic, again looking at my example I can see the bounce rate average has decreased, the session duration and pages visited have both increased. These are all positives to take from the actual traffic the website is actually receiving.
Depending on your website and the amount of referral traffic you’re receiving, this one segment alone could alter your Analytics report quite significantly. We can however take things further and get an even more accurate account of Analytics data.
How to remove non UK traffic from your Analytics report
Depending on your business and your target market, segmenting data by location may be another great way of gaining an accurate insight into the actual traffic numbers that matter for your business. The first thing to do is establish if you’re receiving traffic from countries outside of your target market and if that traffic is genuine or spam.
You can very quickly see where your traffic is coming from by visiting “Audience -> Geo -> Location”. The map will give you an idea of where the most traffic is coming from, scrolling down the page will give you exact numbers. Once again here you’re looking for high bounce rates, low time on site and low pages per visit.
Using my same client example I can very quickly see that during the past 30 days, the client has had 423 visits from Russia. Now my client has no interest in the Russian market and couple that with the data showing as being very low quality we should discount this traffic from our reports as 423 visits from 1,985 (having excluded the referral spam) is a large proportion.
The United States is similar in that quite a lot of spam traffic is being generated from within the country to the client’s website. That said, the bounce rate and average pages visited suggests that not all of the traffic from the United States is spam.
*A further examination of this data for the United States showed that most of the traffic was “direct” and it was spam, however a small proportion was “organic” and was genuine.
As you look through the various countries listed you should quickly be able to see those sending spam by looking at the real-user indicators as we have discussed already. The easiest way to exclude traffic from non-relevant countries is to create a second segment that “includes” only the countries you’re targeting.
For this particularly client, their service is solely targeted towards a UK audience, therefore I can add a new segment to this affect. Much like you did above, “Add Segment” and then give the segment a name, choose “Demographics” and put “Location” as whichever country you wish to include.
How to get a truly accurate assessment of your traffic
Combining the above segments will give you the most accurate idea of the traffic you’re getting that matters to your business. I am sure you will agree it is far more beneficial to know the actual quality traffic you are receiving each month, rather than the inflated default figures that provide no ROI or user engagement.
The above report shows the three segments, the blue line represents all traffic over the past 30 days. The orange line represents the traffic excluding the referral spam. The green however represents the one that provides the most accurate data as this is UK only traffic, which shows no referral spam as being generated from the UK.
Quite significantly, the traffic that makes a difference to my clients business only makes up 47.78% (1,271 visits from the 2,660) of the total traffic to the website over the 30 days. The quality of that traffic is far superior than it appeared at first glance as well.
- Average pages visited increased from 2.36 to 3.45
- Bounce rate reduced from 60.53% to 36.35%
- Visitor duration increased from 00:02:35 to 00:04:28
If I wanted to provide my client with a report from Google Analytics that outlined this most accurate of data that can be relied upon, I could remove the other two segments leaving just the UK segment in place and the report would show as follows.
I hope you can see that these basic changes can highlight big differences in the data you’re looking at or reporting on so be sure to do some investigation and apply the above settings to ensure the data you’re looking at is as accurate as possible.
Finally there are a couple things to note about this approach, the referral traffic segment will need regularly updating. You will know this to be true if you activate the segment then view the referrers still listed. If any show up and are unrelated or spam then add them to the existing segment by clicking the downward error and selecting “edit”.
There is so many more things that can be done with Google Analytics to gain an even greater insight into the quality of your data and crucially how you can utilise the reports to help you make better business decisions but that is for another time.
Everything I have outlined in this post can be undertaken by someone with little or no Analytics experience. If you found this post useful or have a further question just leave a comment below and I will do my best to get back to you.