Understanding Server Side Tagging with GA4
Companies are switching more and more from third-party to first-party data. What are the reasons for this? And what can a privacy-first data strategy look like with Google Analytics 4 (GA4) and server-side tagging (SST)?
What is server-side tagging?
Server-side tagging ("SST" for short), or server-side tracking, means that data is sent from a server to the recipient (vendor) instead of from a user's browser. The principle behind this is:
Classic Tracking:
Classic tracking, also called "client-side" tracking, works via Javascript and tracking pixels in the browser. When visiting a website, the browser downloads and executes a tracking script (e.g. analytics.js for Google's Universal Analytics, gtag.js for GA4). The tracking script in turn collects various data such as visited page, clicks or device information and sends it to the vendor, in our example Google:
This means that there is a direct connection between the user and the vendor. This makes it relatively easy to prevent tracking. A request to "google-analytics.com" can easily be identified or blocked, and it is precisely this mechanism that so-called ITPs ("intelligent tracking prevention") take advantage of. Examples of well-known ITPs are, the privacy settings in the Safari or Firefox browser or browser extensions such as Ghostery or the Privacy Badger.
In addition, an IP address is always transmitted when data is sent, which makes it difficult from a European perspective to track without consent.
Server-Side-Tagging:
With server-side tagging, a separate tagging server is placed in between. Since this server is operated by the provider itself, it is referred to as "first-party" in comparison to the vendor's "third-party" server. All data now flows from the user's browser initially only to the tagging server, from where the data is then distributed to one or more vendors:
This brings with it a whole range of advantages, which we will explain in detail below. The most important fact is that the data now no longer flows directly to "google-analytics.com", but to "tagging.mywebsite.com", for example. ITPs no longer take effect and sensitive data, such as the IP address, can be filtered out on the tagging server, so that some data may also be transmitted to the vendors without consent.
Google Analytics 4 (GA4) & The Google - Server Tag Manager
Even though we talk about server-side tagging, the user data must first get from the browser to the tagging server. For this transport, Javascript must be used in the browser, which captures the required data and passes it to the server tag manager.
GA4 is the latest analytics software provided by the market leader Google. It brings a whole range of advantages compared to its predecessor, "Universal Analytics" (also called "GA3"). These include, for example, an improved data model that enables more detailed tracking of complex processes, but also more features with regard to data protection. Machine Learning can be used for predictions into the future and with the export to BigQuery, every company can now build a free data warehouse.
Universal Analytics will no longer be able to collect data after July 1, 2023. Companies that have thus far relied on Universal Analytics should therefore think about migrating to GA4 in good time. We would be happy to support you in this process and create a free & non-binding migration plan for you.
Google Analytics 4 is wonderfully suited for use with a server-side tag manager. The tool is used to send the data from the user's browser to the tagging server, from where it can be distributed to the vendors. The flexible data model, which is based on events, makes it possible to provide data for all web analytics and online marketing tools with just one request. The GA4 data can be sent from the server tag manager to e.g. Google Ads, Facebook/Meta, AWIN, AT Internet, Econda, Microsoft/Bing, LinkedIn or Piwik in addition to GA4.
Google Server Tag Manager:
Google's Server Tag Manager is our recommendation for implementing server-side tagging. The reasons are:
The software is free for everyone
Installation can be done with a few clicks in the Google Cloud or on any other server running Docker
Very little maintenance is required due to the technical structure
High customizability allows to comply with even strict data protection laws like in DE (e.g. hosting in EU/DE, encryption, logging policy, etc.)
Easy integration with Google tools
Compatible with any vendor/data recipient through "Custom Clients & Tags" (self-programmable endpoints that receive, process and send data to vendors)
It is important to understand that the Server Tag Manager is in no way an alternative to the classic "Web Tag Manager", the "GTM". The GTM is used to trigger specific tracking tags when certain situations or interactions occur on a website. The Server Tag Manager, on the other hand, receives the data sent by the tags triggered by the Web Tag Manager. The products complement each other.
GA4 and the Google Server Tag Manager work wonderfully together. In principle, however, both also work with other open-source or commercial as well as self-built tools.
Why is server-side tagging so exciting?
Intelligent Tracking Preventions (ITP)
Browser manufacturers started to use so-called "ITPs" years ago. The acronym is short for "intelligent tracking preventions" and refers to a feature that is supposed to prevent tracking by third-party services such as Google Analytics.
In reality, this usually works by blocking certain third-party requests (communication between the browser and, for example, Google servers) via a "blacklist". The domains of known trackers such as "google-analytics.com" or "connect.facebook.net" are blocked so that no more tracking data can be transmitted there. In some cases, requests are also deliberately manipulated to transmit false or highly abbreviated data.
Another possibility is to reduce the "cookie lifetime", i.e. to restrict the maximum time in which a cookie can be read again. Apple's Safari browser in particular sets new standards here with a maximum lifetime of only 24h for cookies from known tracking providers. This makes cross-day user tracking impossible.
These types of tracking prevention are currently used in Safari, Firefox and Microsoft Edge and are enabled by default. Companies that rely only on client-side tracking will be hit particularly hard by these restrictions. Depending on the distribution of users across mobile and desktop devices as well as the different operating systems, 60-70% of users can quickly be affected by ITPs.
By interposing a separate first-party server, these restrictions are unerringly circumvented, and it is also not foreseeable that ITPs will be able to reliably detect and block first-party collection servers in the future.
The thing with consent
Since the introduction of the GDPR, and in DE of course also after the introduction of the TTDSG, consent has been necessary for the use of most online marketing or web analytics tools, as these transmit personal data to third parties as well as store user identifiers in the form of cookies on the device. In order to obtain legally compliant consent, consent management solutions (often just referred to as "cookie banners") are usually used. The problem: a privacy-compliant cookie banner - even with an advanced level of optimization - often only has a consent rate of ~70%, meaning that 30% of users generally cannot be captured by marketing or analytics tools. On top of that, the trend is sharply downward. Some companies have seen consent rates drop by more than 10 percentage points over the last 12 months. The reason for this is the presumed increased interest in data protection among users.
However, consent is not necessarily required for every type of web tracking. For the recording of important metrics of a website such as number of visitors, frequently visited pages, entry and exit points or bounce rate, consent is often not necessary. It is the same with key KPIs in ecommerce such as revenue, transaction volume, exit step in the checkout process or similar. This information is not personal per se, but the problem is that with tools such as Google Analytics in purely client-side operation, personal data is also always transmitted if the data just mentioned is to be collected. In particular, the IP address must be compulsorily transmitted to the tracking provider in the case of purely client-side tagging.
With server-side tagging, on the other hand, there is only a connection from the user to the first-party tagging server. On the server, it is then decided which data is transmitted to the third-party providers such as Google Analytics without consent. This allows the most important data to be processed without consent, which significantly increases the total amount of information in the web analytics tools.
Improving Core Web Vitals
What do analytics and online marketing tools from Google, Facebook, Twitter, AWIN, AT Internet, Criteo and many others have in common? They all collect and process the same data. With client-side tagging, separate JavaScript is placed on the website for each of the providers. This leads to a multiple of identical code being executed in the user's browser. This is not only very harmful for the noticeable performance of the website, but above all leads to a worse core web vital or pagespeed score. Tools like Lighthouse, GTMetrix or other scanners notice the large amount of Javascript and penalize the website. But the bad score does not only lead to a bad feeling for the site owners: search engines like Google use such scores to adjust the SEO ranking of a website. Slowly responding and poorly optimized websites visibly disappear from the top Google results and the search engine operator has already announced that core web vitals will have an even greater impact on SEO ranking in the future.
Server-side tagging can easily help with this problem as well: The transfer of user data takes place only once: Browser to tagging server. From there, the same set of data can be expanded, reduced, formatted, or otherwise customized, and subsequently sent to a variety of third-party providers. The site feels faster, SEO scores get higher again, and at the same time you benefit from a comparable mix of data between the different vendors. And if you take a detailed look at the individual web analytics or online marketing vendors, you'll notice that even the technology providers now recommend server-side tagging as the preferred implementation variant of their tags.
Conclusion: The advantages of server-side tagging - can I benefit from the technology?
Basically, we can say that any company that works with data from web analytics or online marketing tools will benefit from server-side tagging. The more important the data is for your business strategy, the higher the impact will be by implementing a server tag manager. The key benefits are:
Adaptation to Intelligent Tracking Preventions (ITPs).
Bypassing adblockers and privacy extensions.
Capture of KPIs even without consent and also in tools such as Google Analytics
Improve data quality
Improvement of website performance by increasing core web vital scores, due to less javascript that needs to be executed in the browser
Better comparability between third-party tools by merging the database
We can only advise every business to try out the new technology. Even if the initial setup is not always easy, the implementation of server-side tagging is almost always worth it. Our customers are able to increase their data volumes by 60-70% on average, while some individual cases have even experienced improvements of up to 95%.
What about the cost?
Last but not least, the possible costs of implementing server-side tagging should of course not remain unmentioned. For the setup described in this article (GA4 + Google Server Tag Manager), the cost breakdown is quite simple:
GA4 is completely free as a web analytics and transport tool. A 360 subscription is not required.
Likewise, the Google Server Tag Manager (as the only professional solution!) can be used free of charge.
There are only costs due to the provision of an own server. These are of course variable and strongly depend on the traffic on the website, the geoscaling and other factors. When operating the tagging server in the Google Cloud, there are usually costs between 10 and 100$ per month. Google talks about a benchmark of $150 per month for a website with 50 requests per second - but most websites are likely to be well below that. If the tagging server is not operated in the Google Cloud, but in its own on-premise infrastructure, these costs can be reduced significantly.
Published on: 12.02.2023