Implementation
Amplitude through Google Tag Manager, with the trade-offs stated up front
GTM is the fastest road to Amplitude events when engineering bandwidth is the bottleneck - and the road has documented potholes. Here's the official-template setup and every limitation Amplitude's own docs state, so you choose it with open eyes.
What's the official setup path?
Amplitude ships an official client-side template - 'Amplitude Analytics Browser SDK' in the GTM template gallery, running Browser SDK 2.0 since version 240a7 (Aug 2023). The documented flow:
- Add the template from the GTM gallery.
- Create an Initialize tag carrying your project API key.
- Fire init FIRST - on the All Pages / Initialization triggers - before any event tag.
- Add event, user-property, and revenue tags that map your dataLayer to your tracking plan.
Setup steps: Amplitude's GTM source docs (Sources). The init-first rule is where most broken GTM installs start (field observation).
What limitations do the docs themselves state?
- Web container only - the client-side template doesn't serve app containers.
- GTM caps custom templates at 100 fields, so newer SDK options ride a documented 'New configuration options' workaround.
- Plugins load only when the template tag fires - so plugin features require tag-sequencing chains.
- Data-layer timing: rapid successive pushes can hand tags stale values.
All four from Amplitude's own GTM docs page - these are their statements, not our editorializing. Our judgment on top: GTM fits marketing surfaces and dataLayer-shaped events; product surfaces with identity nuance and offline behavior are native-SDK territory. Many stacks run both.
Is there a server-side GTM option?
Yes - a separate official template for the GTM server container, which sends data through Amplitude's HTTP V2 API and supports IP-address suppression and UTM capture. It lives only in the server-container gallery. Server-side GTM buys you first-party data flow and ad-blocker resilience at the price of running the container infrastructure - the standard sGTM trade, not an Amplitude-specific one (that framing is ours).
The decision in one paragraph
Choose GTM when the events you need already live in a disciplined dataLayer and shipping speed beats depth; choose the native SDK when identity, sessions, plugins, or mobile parity matter; choose both when marketing pages and product surfaces have different owners. Whatever the route, the tracking plan comes first - GTM makes it faster to ship events, including wrong ones.
Sources
- Amplitude docs: Google Tag Manager (client-side template, setup, limitations)
- Amplitude docs: Google Tag Manager server-side template (HTTP V2 API)
Product capabilities change - every claim above links to the primary source it came from. Judgment calls and field observations are ours and labeled in-line.
Frequently asked questions
Is there an official Amplitude template for Google Tag Manager?+
Yes, two: a client-side 'Amplitude Analytics Browser SDK' template in the GTM gallery (running Browser SDK 2.0 since Aug 2023), and a separate server-container template that sends data via Amplitude's HTTP V2 API. Both are documented in Amplitude's source catalog.
What are the downsides of implementing Amplitude through GTM?+
Amplitude's docs state four: web container only, GTM's 100-field template cap (workaround documented), plugins requiring tag-sequencing because they load when the tag fires, and possible stale values under rapid dataLayer pushes. Our added judgment: identity-sensitive product surfaces are better served by the native SDK.
Should I use server-side GTM with Amplitude?+
If you already run an sGTM container for first-party data flow, Amplitude's official server template (HTTP V2 API, IP suppression, UTM capture) slots in cleanly. If you don't, adopt sGTM for stack-wide reasons, not for Amplitude alone - the container is real infrastructure to operate.
Get a free, scored audit of your Amplitude instance
Send us read-only access and get a scored findings report within 48 hours: taxonomy health, duplicate events, governance gaps, and the three fixes with the highest data-trust payoff. No commitment.
Request the free audit