Google Signals - where's my data gone?

With the rush to implement GA4 before the final curtain fell for Universal Analytics in the summer, it's perhaps easy to miss or underestimate certain changes to your data which may have occurred following migration. Fear not however, CORE are here to shine a light!

Introducing Google Signals

Demographic data is often an invaluable tool in understanding who it is exactly who is browsing your website. Details around age, gender and interests can be vital in a number of respects, from remarketing lists to persona development, and for some is simply a fundamental feature of reporting on their website.

So we think it is therefore important to point out how opting in to the GA4 equivalent in demographic data, ‘Google Signals’, may lead to some unintended consequences for other aspects of your precious data.

First of all, a brief introduction to Google Signals. Google Signals is a feature first launched in 2018 which allows Google Analytics users to understand how site users interact with your website across multiple devices and sessions. However, it has recently gained increased prominence given the sunsetting of UA in favour of GA4 and now fundamentally if you require audiences in Google Ads or to see demographic data in your GA4 reports, you must have Google Signals activated.

No big deal right? Well yes and no. It’s certainly a simple enough task (just flicking a switch) but that simple task could have rather drastic implications on your data elsewhere.

In short, it can cause quite severe thresholding in your account. This means rows of data throughout your GA4 reports which contain relatively small figures will be hidden. Now the argument is if you’re only missing out on small portions of your data these can’t be that significant surely? Well if you’re anything like us from a data nerd perspective you’ll know that sometimes the most interesting insights can be gleaned from those grains of sand within your data. 

The biggest issue we’ve seen is the effect it’s had on event data and custom parameters. A client of ours with a relatively sophisticated set up of event tracking saw its figures decimated by the reduction to the point of uselessness.

Harumph indeed.

What’s to be done?

As mentioned, if you require Google Ads audiences or demographic data you’ll have no choice but to turn on Google Signals. However, there is something which you can do to prevent this from permanently altering your data and it involves something called ‘reporting identity’.

Reporting identities in GA4 have a distinct effect on the results in data you will see in your account as they each provide different combinations of methods of data gathering. There are a wide number of resources available which can provide further details around the methods themselves, however for our purposes there is only one identity that can really solve our issues - ‘Device-based’.

‘Device-based’ only utilises the ‘device-id’ method of data gathering and although is the least sophisticated in concept it provides the best alignment with Universal Analytics. To make the switch, access the admin dashboard of your GA4 account, click ‘Reporting Identity’, then ‘Show All’ and select ‘Device-based’. Your thresholding concerns should hopefully then be a thing of the past!

What’s also great about reporting identities is you can seamlessly switch between each option - you don’t have to make a choice and stick to it forever! In fact we recommend you do try out each reporting identity and see what effect it has on your data before perhaps looking to settle on a view in order to provide accurate benchmarks.

If you require any assistance with navigating this new GA4 landscape please don’t hesitate to get in touch and we’d be delighted to help get digital working for you.