Six Google Analytics faults that destroy your data
Google Analytics is the world's most widely used analysis tool for websites.
Over 95% of our customers use Google Analytics to analyse traffic to their website, which of course has many benefits:
- It's free up to a certain level (over 500,000 sessions are randomly tested).
- It is relatively easy to set up using a simple script on the site.
- It connects to other Google tools, such as Adwords and Google Search Console.
- And most importantly: You can extract some very valuable information from Google Analytics.
Google Analytics is, however, not the perfect tool. It has its limitations when it comes to blocking scripts, incognito mode and ad-blockers, and one of the biggest complaints in these cross-device times, is that it's not that great at measuring users as they move across devices. That's why you can't expect your figures to be 100% accurate.
In some cases the fault will be non-critical, but in others it can give you deficient or incorrect data, which means you make the wrong decisions, lose oversight or miss important trends.
Over the years, we have worked with hundreds of Analytics accounts across different industries, companies etc., and we can highlight 6 typical faults that we see time and time again.
1. Deficient/inconsistent UTM tracking
UTM tracking is a standard way of measuring the effect of your campaigns in Google Analytics (an exception is Google Adwords, which is automatically connected to Google Analytics).
By placing a piece of code after the links to your website, you can see exactly which campaigns generate traffic and turnover, which media they are on and so forth.
The fault tends to occur if you have forgotten to enter the UTM tracking in:
- Facebook campaigns and/or posts
- Banner ads on different media
Another fault is inconsistencies in how you name things. This is important. Note that capital letters/small letters make a difference; if you write both CPC and cpc, they will appear as separate channels.
A horror story
Putting UTM tracking on internal links on your own site can prove to be a deadly sin.
I still remember a horror story from a few years ago, when a webshop had put UTM tracking on all the banners on their home page. The fault made it impossible for the webshop to trust their figures for sessions, rejection percentages etc. Measuring which channels, landing pages, etc., were performing well or not became a very laborious task.
They had a completely lopsided view of how much traffic they were getting or what was working. In fact, they generated 15% less traffic than they had anticipated, and they had no idea how their turnover was distributed across channels.
You’re probably thinking this was a one-man band webshop. But it wasn't. The site belonged to a major Danish retail chain, with a group turnover of +1 billion a year.
2. More than one Google Analytics script on the site
There is generally nothing wrong with having more than one Google Analytics script on your site. The problem arises when there is more than one Google Analytics script that refers to the same property.
If you put the same Google Analytics script on the site twice, at least 2 page views will be issued, and your rejection percentage will be unusually low. It will also look like you have more page views than is actually the case.
If your rejection percentage is below 10%, it is highly likely you have more than one script on the site. Check it out and remove it.
3. Failure to remove bots and spiders
Not all bots and spiders mean big problems. There are also some 'good' ones, which are simple bots from, e.g. Microsoft that crawl your site. They can lead to some odd spikes in 'direct' traffic with a high rejection percentage.
Go into the report in Target group > Network and see whether any of the service providers look odd. Look for rejection percentages of +95%.
On this screenshot, we can e.g. see an account where +5% of the sessions come from bot traffic from Microsoft!
To exclude these bots, go into Administrator > Settings for views and tick "Exclude all hits from known bots and spiders".
4. Faults in cross-domain tracking
Cross-domain tracking is one of the most common faults we see, and it's really tricky to remove. Cross-domain tracking is most common for sites that have the following characteristics:
- The site exists on more than one domain/sub-domains.
- You have to go through a payment gateway during the checkout flow.
What to look for
Check whether your own domain performs under source/medium in the report and whether there are any sources with titles such as:
- application.laanlet.dk f
If a large amount of traffic and/or turnover comes from here, you have a problem that you need to deal with.
In many cases it will be sufficient to add the sources (and your own domain) under Property >Tracking information > List of referral exclusions.
But not in all cases. So check if you suddenly start to see large rises in traffic from Direct / None. If this is the case, you will usually need to make some changes in your Google Analytics implementation. A subject you can read more about from Google.
5. Internal traffic
This problem occurs mainly on websites where a significant amount of traffic comes from people inside the company, e.g. on intranet-sub-domains. And on low traffic sites, it can lead to a MAJOR part of your traffic being from you - which can really distort your data.
That's why we tend to recommend a sort of filtered viewing, where internal traffic is removed. Go to Administrator > Filters > Add filter
NB:We generally recommend that you also have a 'clean' profile without any filters. If there is a fault somewhere, then you always have clean data to go back to.
6. Significantly higher or lower turnover in Google Analytics than in other systems
As mentioned at the start of the article, you can't always expect e.g. turnover figures to match 100% with actual figures. There are also too many fault sources:
- For example, you haven't uploaded returned goods in Google Analytics
If the difference (removed returned goods) is more than 10%, there's usually something amiss. Here are some of the most common fault sources:
- Incorrect currency in ecommerce tracking script. Such as EURO instead of EUR (common on bigger websites where payment in multiple currencies is possible)
- Some transactions are measured more than once - usually known as duplicate transactions. This often happens if the receipt page loads more than once, and the set-up hasn't been taken into account
Lunametrics has written a guide to how to remove this error. But only spend time on this if it's a really critical issue.
Do you find yourself questioning other data in Analytics?
Your organisation's set-up is of course unique, and your key figures are probably different to those of your competitors. You're always welcome to contact Novicell for a no-obligation review of your online tracking, or to set up bespoke reports.