There is a tsunami of hate directed at GA4. If the truth be told, this is largely a reflection of the fact that nobody really likes change, especially when it is forced upon them, and GA4 is a fairly dramatic change.
I absolutely agree that GA4 does not really feel ready for mainstream use and lacks some really basic features that we have enjoyed in UA for years. I must confess, however, to a growing appreciation of some of the changes.
A good example is the abandonment of bounce rate. I have always argued that bounce rate is not an especially useful measure of quality as there are plenty of scenarios where a high bounce rate actually indicates a positive user experience. I much prefer the GA4 approach of active users and look forward to the day where we are not having to pacify panic / concerns about a high bounce rate on pages such as a ‘contact us’ page.
I wanted to share another example where I think GA4 offers a step in the right direction.
Anyone who has implemented a cookie consent banner in the frustrating correct manner will have probably experienced an alarming drop in reported traffic. Universal Analytics (UA) relies primarily on cookies, so anyone who rejects cookies will not have their session recorded in Google Analytics. If you are strict with your GDPR compliance, you should not set the cookie until you have the express consent to do so. The introduction of a GDPR compliant cookie consent banner can therefore create a bit of GA panic.
A good example of such potential panic is one that is close to home – here is a graph of traffic to the Browser Media website, as reported in UA:
It doesn’t need an eagle eye to spot an alarming drop in traffic in December 2022. A keener eye will spot an annotation at the time of the crash – this read “implement correct cookie consent”. Whilst it drove me slightly mad to kiss goodbye to the valuable information that was provided in Google Analytics, I felt that we should really toe the line and not poke the stick into the GDPR hornet nest.
On the basis that pretty much all our website traffic is driven by organic search, the graph above (in isolation) would lead to a very plausible assumption that we had done something horrendous with our own SEO. I didn’t panic as good old Google Search Console showed that there was no ‘off the cliff’ moment in December 2022:
I knew from other metrics, e.g. number of enquiry forms, that it was simple a reporting issue. Traffic levels had not evaporated and the UA reporting was simply a reflection of the fact that nobody was clicking on the cookie consent banner, as it was at the bottom of the blog post that they had (hopefully) enjoyed but had no need to then click on the cookie consent as they had (hopefully) found what they had been looking for. As an aside, there is an interesting debate about how intrusive to make the consent banner, but that is for another post.
GA4 was much more encouraging for the same time period, despite being subject to the exact same cookie headaches:
Yes, there is drop in late December, but that was the Christmas festive period where folks were quite rightly stuffing their faces rather than reading a digital marketing blog. So, how is it that GA isn’t as affected by strict adherence to cookie consent?
GA4 is essentially more capable when it comes to tracking user behaviour without relying on cookies. You do need to ensure that you have selected ‘blended’ in the reporting identity settings:
GA4 then uses a much wider range of signals to allow reporting of user behaviour, even when they have not given you permission to use cookies.
What is exciting, to a geek like me, is that this will only get better. As shown in the reporting identity settings screenshot above, we are still not enjoying the benefits of modelling on our GA4 account. I am hoping that this will be available to us soon as I am looking forward to analysing the impact that it has. Google will use data from similar users, who have allowed cookies, to model the behaviour of those users who do not consent to the use of cookies. You can read more about it here.
Whilst there is, of course, a risk that the modelling could get things horribly wrong and be very wide of the mark and it is still frustrating to have to rely on AI generate data, rather than hard user interaction data, this is definitely an area where GA4 has the ability to trump UA. It is an area that I will be watching with interest.
I hope this short post gives you a glimmer of hope that life in the post-UA era will not be all bad. There is some light at the end of the tunnel.