This article was written in collaboration with Florian Chabert.
Server-side tagging has highly intensified in the last 12 months. We’ve been talking about it a lot for the past few years, but in 2022 we noticed many organizations developing projects in this area, especially companies with mature analytics capabilities or who were comfortable with the technological uncertainty the new paradigm brings.
There are many reasons the analytics world is interested in server-side tagging. It’s used particularly for performance optimization, to enrich the data available to a business (through the addition of primary data), and for more developed masking options for sensitive data (enabling them to be scrambled before being sent to analytics and media platforms, amongst other things). But it is particularly being used to mitigate restrictions related to cookies (whether third- or first-party), which is the main advantage that’s driving our clients to resort to this tracking method. But be careful—while some completely legal and legitimate techniques can allow you to get around current restrictions in effect on data tracking, playing a game of cat-and-mouse with browsers will likely not pay off in the long term.
Over the years, our team has had many opportunities to guide organizations down this road, whether supporting companies in realizing “little” projects (to solve simple issues) or major project development involving the complete transformation of a data collection ecosystem.
Based on the diverse range of experience we’ve acquired, we would like to offer an overview of the restrictions related to using cookies, a few examples of situations where server-side data collection can be a wise move, and a few possible solutions that you can explore before getting started with server-side tagging.
The issue with cookie expiration
For many years the analytics world has relied on using cookies as its main source of information. These browser cookies enable third-party libraries installed on websites to store or collect data from users of their platform.
To avoid the dishonest use of these techniques, browser providers have slowly tried to restrict the capacities of cookies (scope, lifespan, etc.) to limit them to their intended uses. As the game of cat-and-mouse continued, the restrictions became more and more onerous, even for owners of sites that were careful about respecting the privacy of users.
The most restrictive and widespread feature without doubt came from Apple and its Safari browser in the form of ITP (Intelligent Tracking Prevention). This program, in constant evolution since September of 2017, notably includes
- the disappearance of third-party cookies;
- restrictions on the lifespan of cookies dropped in JavaScript (and therefore all those dropped by third-party libraries);
- restrictions on the lifespan of all information inserted into browsers’ temporary storage such as localStorage.
Other browsers implement similar programs, but they don’t all have the same popularity as Safari.
What is the real problem?
The biggest limitation on lifespan is related to link “decoration,” meaning links containing URL parameters. All advertising links are generally constructed in the manner presented below:
Diagram showing typical loss of attribution related to Safari’s ITP restrictions.
What is the result?
Logically, if the advertising attribution from media platforms is cut off, you can expect two major effects on the activities of a marketing team as a result:
- It becomes harder to measure the impact of campaigns that are more removed from the customer’s conversion (awareness or consideration campaigns, for example, are heavily affected by this loss).
- Media platforms will no longer be able to learn as easily based on campaign performance nor optimize their automatic adaptation to campaign targeting as well as they did before. This means the performance of campaigns won’t be as good.
Ways to mitigate the damage from the client side
Platforms that sell advertising have therefore found an easy solution to continue to connect people who have seen or clicked on an ad with conversions to the maximum possible extent. This simple technique is based on primary data.
Take, for example, the email address of a user. The platform knows who it has shown advertising to and sometimes also knows their email address. Such is the case when someone buys a product or fills out a form. In addition, if at the time of completing a purchase a person indicates their email address, the platform is able to make a connection between these two distinct actions.
This process of associating data has a different form and even a different name depending on the platform. On Facebook and TikTok, it’s called Advanced Matching, on Google, Enhanced Conversions, and on Pinterest, Enhanced Match. But no matter what it’s called, it’s still the same idea, even if these platforms accept different kinds of data for associating the first and last name of the user, their email address, street address (with all of its elements) and phone number.
17%
improvement in conversion rates on YouTube action campaigns (on average).
According to a Google study conducted in 2022
These data association techniques are very effective and, as part of a good attribution optimization strategy for conversions, they should be the first to be adopted. According to Google, they increase conversion rates for “Search”-type campaigns by 5%, on average, and those for “YouTube for Action”-type campaigns by 17%.
Of course, to be able to employ these techniques, you need to have established methods of primary data collection. It’s fairly common to collect some of these data during the purchase process. On the other hand, it’s not as obvious how you would obtain such information as part of other types of conversions or intermediary micro-conversions, such as a “drive-to-store” for example.
There is also another more technical option to gain additional conversion points, among other things, by prolonging the lifespan of cookies: server-side tagging.
Server side tagging
Usually, during a classic analytics integration, the user’s computer will contact many data collection servers, as shown in this diagram:
Diagram of an integration using the Google Tag Manager client.
This generates many calls from the user’s browser. In some situations, this will simultaneously execute a significant amount of JavaScript code on the browser. Each library (including pixels in most marketing situations) will drop its own cookies, which are subject to many limitations.
In an integration made from the server side, you will use a data collection server such as Google Tag Manager (GTM Server), for example.
Diagram of an integration using the Google Tag Manager server.
In this kind of use case, the GTM server will be required to execute several actions, such as:
- Receive the most “raw” request possible from the user’s navigator
- Remodel and enrich the data
- Drop a cookie from the server (these cookies have fewer lifespan restrictions)
- Communicate the full request to each service
Unlimited advanced possibilities:
Server-side tracking gives you access to many more advanced cases for companies whose marketing activities rely heavily on this operation. Here are a few advantages to this way of working:
- If you keep your pixels, you can duplicate all your conversion events through the APIs of media conversion platforms (Facebook, Google Ads, Snapchat, TikTok, Pinterest, etc.).
- If your data layer is sometimes missing complete data and your tagging tool (GTM, Adobe Launch) uses a lot of JavaScript, implementing all of that from the server side will give you access to a technologically more powerful tool and will considerably limit the code loaded on the site.
- You can save all the tracking errors you identify in a database (BigQuery, Firestore, Google Spreadsheet, etc.) quite easily, which will facilitate problem identification (and resolution
- You can send email alerts for specific cases, such as following each completion of a form.
Lastly, server-side tracking is an extremely flexible system. It basically offers the possibility of contacting any other service that has an API (to send Slack notifications or emails, or to contact a database, for example).
What is the benefit?
When properly integrated, server-side tracking will provide:
- better conversion measurement, mainly when performed later (not right after a click), on another device or in another browser;
- a good technological direction, since trends indicate the market is heading increasingly towards server-side tagging (Adobe, Amplitude, etc.);
- better management of complex ecosystems.
A good game plan
If your main goal is to achieve better measurement of your campaign results to optimize media investments, your game plan should begin with the activation of improved conversions directly through client-side pixels.
This strategy requires less effort in the short term and can often be deployed very quickly. It will be even more effective if you have a good data layer and a good strategy for collecting primary data. Once you have activated the improved conversion features, you can start thinking about the next step, which is setting up server-side tracking.
Don’t forget that no matter what tactics you use to get around or overcome any obstacles, measuring attribution on the Internet is becoming an increasingly difficult operation to perform.
To meet this challenge, remember that arriving at a more holistic picture of your sources of conversions (through cause and effect analysis, for example) can give you a better perspective on your performance.
Finally, if your goal is instead to strengthen your measurement ecosystem and establish advanced techniques (anomaly detection, data enrichment, etc.) as suggested above, server-side tagging is perfectly adapted to this.
*
Our team has a lot of experience with guiding organizations in establishing server-side tracking projects, from simple implementations to more ambitious projects that completely transform a data collection system.
Don’t hesitate to get in touch to learn more about the advantages your company could gain from this technology.
-1.png)
