Repeatedly I see confusion in the community about the terms and that is why I wrote this article to explain the differences and advantages and disadvantages of sGTM.
Client-side Measurement
Principle
JavaScript code embedded into the webpage (loaded in the user’s browser) collects data about user behavior and sends it directly to the analytics tool. Naturally, this depends on the user’s consent for tracking and advertising, as well as the subsequent settings that might allow measuring without using client-side storage and avoiding the collection of personal user data, focusing instead on static (truncated) usage data.
Advantages
- Easy to implement.
- Broad compatibility with various platforms and tools.
- Extensive documentation and a large pool of professionals who know how to use it.
- No direct operating cost.
- It’s transparent to users, since they can see exactly which data is being sent and where.
Disadvantages
- Susceptible to manipulation.
- Dependent on the client’s browser and its settings (ad blockers, browser restrictions, etc.).
- Everything is public, since the user can intercept the data.
Who is it for?
Basically any website, covering about 98% of use cases.
In most instances, this data is sent directly to analytics or advertising tools. These data points can also feed Google server-side tagging.
Want to know more?
Server-side Measurement
Principle
Data on user behavior is collected by the web server. Visitors can’t block anything because nothing flows through their browser. It’s not intended to circumvent GDPR/ePrivacy requirements regarding cookie creation, IP usage, sending a user’s email to third parties, etc. To do any of that, you still need the user’s active consent.
Advantages
- Greater accuracy, as there’s no user browser that could block anything.
- Enhanced security. Generally, this data can’t be easily manipulated.
- Ability to collect data from client browsers and devices that don’t support JavaScript.
- You can transmit non-public data, such as margin. When you have consent for measurement, you can bypass ad blockers and some browser restrictions.
- It’s possible to measure statistical usage data of the service/web server without consent—i.e., data without user identifiers (no client-side storage, no device/IP info, no emails, etc.).
- You can selectively measure only certain events, like conversions to Google Ads or just Facebook conversions. You don’t have to measure everything from the start.
Disadvantages
- More complex implementation.
- Compatibility with fewer tools (a nod to Sklik here).
- Often fewer additional data points that could come from external web integrations or JavaScript-based collection.
- Combining GA4 client-side and server-side data (without sGTM) can generate very confusing results for both humans and machine learning models.
- Requires a highly knowledgeable specialist to fine-tune the measurement.
- Any change requires developers.
- You need to closely track any changes on the tool-provider side.
Who is it for?
A handful of sites—typically those sending data about margins or high-quality leads later on when they’re converted. In practice, it’s for more advanced projects.
These data points can feed Google server-side tagging or be sent directly to analytics tools.
Want to know more?
Google Ads Enhanced Conversions
Google Server-side Tagging (sGTM)
Principle
Many people assume that Google server-side tagging (sGTM) is a tool for server-side measurement. In reality, sGTM is an analytics proxy hosted on your own server that lets you forward data from client-side measurement (the browser) or server-side measurement (the web server) and modify it. By itself, sGTM measures nothing; it needs an external input. Only then can you attach additional data as if it were being sent from your own web server—just that it’s coming from sGTM. Also, it is not intended to circumvent GDPR/ePrivacy—you still need consent, depending on the scope of the data being sent. Google calls it “server-side tagging,” not “server-side tracking,” because these are two different things.
Advantages
- Reduces data and processing load on the user’s browser. For example, you can measure a conversion once via GA4 and then distribute it via sGTM to Google Ads, Facebook, Instagram, LinkedIn, and TikTok. You can even match that conversion in these tools using the visitor’s email address if consent allows.
- Helps with GDPR and ePrivacy compliance because it lets you maintain stronger control over the data. The data flows through your server, so you can, for instance, partially strip IP addresses to anonymize them before sending data further.
- You can detect spam bots and discard their data.
- Enables advanced setups like custom session handling—allowing users to move between multiple sites and apps without breaking the session, and more.
- It can bypass certain limitations of client-side implementations, and sometimes even fix data that otherwise couldn’t be changed due to vendor or platform restrictions.
- You can opt for a minimalistic implementation—for instance, sending margin data to Google Ads conversions only. This doesn’t have to be overly expensive.
- Services like Stape.io can help you host sGTM. You can also learn various tricks there.
- Combining client-side measurement with sGTM typically yields higher data quality, as it’s less likely to be blocked. By hosting measurement scripts on your own domain, you can capture 5–12% more data.
- Some privacy-centric countries require this type of separation if you want to measure anonymous data without user consent. Otherwise, the third-party service would see the user’s real IP address or device details, making the data non-anonymous. So without sGTM, anonymous measurement may not be allowed at all.
Disadvantages
- Without proper client-side or server-side measurement, sGTM by itself measures nothing.
- Beyond implementation, you also need to maintain it. Keep an eye on costs to avoid any surprises.
- Overall complexity increases—you have one more tool to master.
- It’s not free, especially at scale. The cost grows with site traffic. You can start at around 300 CZK/month for a tiny blog that measures almost nothing, up to around 9,000 CZK/month for sites with tens of millions of visits.
- If sGTM isn’t functioning or is underpowered, everything flowing through it won’t be measured.
- Combining client-side and server-side measurement via sGTM can significantly improve data quality—but this is advanced territory requiring a skilled web analyst.
- Be cautious: agencies offering sGTM sometimes don’t really understand web analytics. They might just implement it for its own sake and bill you for monthly maintenance without actually delivering better insights.
Who is it for?
For advanced projects that know exactly why they need it. If your basic client-side measurement isn’t solid, implementing sGTM can be a waste of money. Sometimes specialists set it up for clients who really don’t need it—basically to experiment at the client’s expense. It’s very useful if you know how to leverage it, but realistically only about 10% of projects truly need it.
Want to know more?
Basic documentation from Google on sGTM.
The most renowned sGTM course is by the very famous Simo Ahava. It’s not cheap, but it’s probably the best. He also has excellent articles on the topic.
Another great resource is Julius Fedorovicius and his Analytics Mania.
For the nitpickers: I’m not counting the half-baked approach of hosting an image on sGTM just to measure that.