In Q3 of 2020 Google made a huge shift with its beta release of Tag Manager. They enabled the Tag Manager in a server-side environment by leveraging Google Cloud App Engine infrastructure and a JavaScript focused framework. Now, It feels like the many pieces of a puzzle coming together as they are meant to be. Possibly, a well-planned strategy in advance of this measure.
You can definitely tell this is the beta version, as there are many features missing in server-side GTM. Notably, Google does face challenges that most other companies will not experience.
Google should put this technology into perspective and shine the correct light on it. Yet, this approach feels like a counter-attack to ITP. It does require more than a simple approach to circumvent a User's consent or to track anything that GDPR or ePrivacy would prevent. GDPR/ePrivacy does not explicitly forbid anything, but rather it only requires you to get the consent of the User. However, it is still your job to align the strategy with any applicable legal requirements. As with any solution that holds great possibility, it also comes with great responsibility. The excellent news is that with server-side tracking approaches, there are security features that will actually improve the user experience and data privacy!
As of now, the presence of the server-side GTM is not groundbreaking. When following the first tutorials to implement Google Analytics server-side with GTM, you will notice the lack of many features. It is a quick setup with an absolute mainstream focus and will work with any website that already uses Google Tag Manager in the front-end. A good starting point is Simo Ahavas Blog.
The enhancements of the basic integration come next and explain when you can move to first-party cookies set via HTTP headers: Building a custom server-side client in server-side GTM.
It is a good thing to have a server-side solution that comes with low entrance cost. However, the economic price does lack some important features:
1. Third-party integration:
In its basic setup the server-side container will sit on a GCP managed domain and URL, and your website visitors will still have to call a third party. This can be changed quickly with a CNAME integration though. This isn’t mandatory and requires your own initiative.
2. Third-party libraries:
As of now the server-side GTM has an API “getGoogleScript” which is capable of serving analytics.js or gtag.js. You can make those libraries a first party integration if you map the App Engine via CNAME to your domain. However, the gtm.js from your front-end Tag manager will still be served from a third party. Google is officially working on a solution for this, but it could take some time.
3. Mostly front-end focused solution:
When working with GTM server-side containers, the most important part is still your front-end GTM container. Most sites that utilize GTM will still rely heavily upon dataLayer, GTM and its front-end tracking approach. Custom implementations of protocols and scripts won’t be something that I would expect to become mainstream. That would be a high-maintenance, development and resource eating monster of a project to undertake.
Now, let's discuss the GTM Server-side option and give it a rating. When it comes to technical feasibility the GTM solution gets a high score. As described, it's a quick and easy setup and absolutely feasible even for small sites. It will fit perfectly into your existing ecosystem if you already have GTM installed in your front-end. In regards to data sovereignty, this solution does lose some points, as you don’t have the option to choose your infrastructure. You will be bound to the Google Cloud. As for data protection, the “easy” integration of GTM Server-side will not make a huge shift, as this approach heavily relies on front-end work. The “classic” GTM will still require lots of code to make the approach work. It is prudent to consider the weak points of this approach carefully.
In my opinion, GTM server-side seems like a good entry point for most websites. It will come with some features and benefits. But, if you want to go full server-side and control level data (100% data experience), then it would be better to build a custom solution of your own. Not satisfied with this solution? — We have listed different solutions in our free ebook!
