What are transformations in server Google tag manager

Before transformations existed, changing the data a tag sent meant editing the tag, or building a variable for every field. If five tags needed the same fix, you made the fix five times. A transformation lets you make the change once, in one place, and decide exactly who it applies to. This guide covers the three types, the scope rules and some practical recipes. The surrounding parts are described in clients, tags, triggers and variables.

Where a transformation sits

The server container handles an event in a fixed order: the client builds event data, transformations run, triggers are checked, tags fire. A transformation therefore sees the event exactly as the client produced it, and the tags see the event exactly as the transformation left it.

For online stores

Orders from your store, back to the ad that earned them

Your store sends the order once. Open Tracked passes it on to each ad platform, cleaned and deduplicated.

Open Trackedorder tracked

+1 order

The three types

Allow parameters

You list the parameters to keep. Every other parameter is removed from the event the tag sees. This is the strictest option and the safest one for a vendor that should receive very little, for example an advertising network that only needs an event name, a value and a click ID.

Exclude parameters

You list the parameters to remove. Everything else stays. Use it when one field must never leave, for example the visitor IP address or an internal customer number.

Augment event

You add or overwrite a parameter. The value can be fixed text or the result of a variable. Common uses are adding a currency when the site forgot it, setting a hashed e-mail address, normalising an event name, or copying a value from a cookie into the event.

Match conditions and scope

Every transformation answers two questions:

Scope is the point most people miss. A transformation with no scope set can affect every tag, which usually is not what you want. Setting it to one tag keeps the change local and makes the container easier to reason about when you return to it six months later.

Six practical recipes

01

Keep the IP away from GA4

Create an Exclude transformation for ip_override, scoped to the GA4 tag. The same event still reaches Meta with the IP it needs for matching.

02

Send Meta only what it needs

Create an Allow transformation scoped to the Meta tag, listing event name, value, currency, content IDs and the user data fields. New fields added to the site later do not leak by accident.

03

Fix a currency once

Create an Augment transformation that sets currency to USD when it is missing. Every tag now sees a currency.

04

Hash an e-mail once

Augment the event with a hashed e-mail variable. Meta, TikTok and Google Ads all read the same prepared value.

05

Rename an event

If your site sends order_complete and a vendor expects purchase, Augment the event name for that vendor only.

06

Add the country

Augment the event with the visitor's country from a request header. With Open Tracked, the Visitor Location power-up supplies it.

Transformations and privacy

Transformations are the right tool for data minimisation: send each vendor the least it needs to do its job. They also leave a clear trail, because the container lists every transformation and what it does. When a reviewer asks "which vendors receive the visitor's e-mail address", you can answer by reading the scope of the transformations. See GDPR and server-side tracking and personally identifiable information.

Testing a transformation

1. Open Preview in the server container and send an event.
2. Click the event, then the Transformations tab.
3. Check that your transformation is listed as "applied" for the event.
4. Open the tag and read the event data it received. The removed field must be missing or the added field present.
5. Send an event the match condition should skip, and check the transformation is shown as not applied.

Common mistakes

• No scope. The transformation hits tags you did not mean.
• Allow list too short. A tag stops working because a field it needs was not listed. Check the vendor's required fields first.
• Order assumptions. Two transformations on the same event run one after the other, so the second sees the output of the first.
• Augmenting with an empty variable. The field is added but empty, and the vendor rejects the event.

In Open Tracked

The container files from the Open Tracked Setup Wizard use variables for the power-up data (Visitor Location, Device Info, User ID and the Enricher's stored values), so you can read them in a transformation or a tag without writing code. Keep your own rules in transformations, so they survive when you download fresh files.

Transformations compared with other ways to change data

A checklist before you publish a transformation

The short version

3 things to remember

Scroll and the points stack up.

01

A transformation edits event data before triggers and tags run. It has three types: Allow, Exclude and Augment.

02

You can scope a transformation to one tag or one client, so Meta can get a field that GA4 does not, and the reverse.

03

Use them to keep personal data away from vendors that do not need it, and to fix names and values once instead of in every tag.

Frequently asked questions

Fix your tracking at the source

Open Tracked hosts your server-side Tag Manager, keeps first-party cookies alive and shows what each ad is really worth.

Start freeMore in Glossary →

Written by

The Open Tracked team

We build Open Tracked: server-side tracking hosting, analytics and the tools around them, for stores that spend on ads. Questions about this guide, or an idea for the next one?

Write to us Read the docs