Webclat / Tealium Practice
Webclat / Tealium Practice  /  capability guide

Creating and connecting a Tealium data source: the platform data source, start to finish

A data source is the connection point between one specific system and the server-side layer. Getting the setup sequence right the first time is what stops a connector from becoming a second implementation project six months later.

What a platform data source is

A Tealium data source connects a specific system - a SaaS platform, a cloud warehouse, a webhook, a file feed - to the server-side layer. "Platform data sources" are the pre-built connectors for common systems; a generic webhook or File Import source is what you configure by hand when no pre-built connector exists.

When you need one

Any time server-side collection needs to originate from something other than a client-side tag on a webpage - a CRM firing a webhook on a status change, a cloud warehouse holding an attribute worth syncing, a partner system pushing conversion events directly.

How to set one up correctly

Pick the source type deliberately - a pre-built platform connector when one exists is almost always less work and less risk than a hand-built webhook doing the same job. Authenticate per that source's actual requirement (API key, OAuth, signed request - this varies by source and by Tealium release, so confirm current requirements rather than assuming). Define the mapping from the source's native fields to Tealium events or attributes explicitly; do not accept a default mapping without reading what it actually does. Decide the downstream destination before going live - a specific EventStream, a data layer attribute, an AudienceStream visitor attribute - and set access controls to match who should be able to see or edit the source.

How to verify it worked

Trigger or send a real test payload from the source system itself, not a synthetic one built to look convenient. Confirm the trace or log on the Tealium side shows a received event with the fields you expect, and confirm the mapped fields actually populate on a test profile - not merely that the source shows a "connected" status.

Common mistake: treating "connected" as the finish line. A data source that authenticates successfully but maps the wrong field, or none at all, will report healthy indefinitely while quietly producing nothing useful.

Common questions

What's the difference between a data source and a connector?

In practice the terms overlap - a data source brings data in, a connector (or webhook connector) typically sends data out. Some platform integrations do both. The distinction that matters operationally is direction: inbound versus outbound.

Can a data source feed more than one destination?

Generally yes - once data lands through a source it can be routed to multiple downstream connectors or attributes, subject to your license tier. Confirm the specific routing options against current documentation.

Get the data source mapped right the first time.

Source selection, authentication, field mapping, and a verification pass against a real test payload before anything goes to production volume.

Scope My Integration