Getting started with Event Tracking

Alec Moloney  
Edited

Event tracking is the capture of user actions as they occur. This rich form of data allows us to understand better why people make the decisions they make, and build solutions to make improvements and recommendations to better meet user or customer needs.

 

Tracking considerations

Before you begin, it's good to understand that how you design your event tracking has an impact on how events are processed through the pipeline. This includes how long it takes, the data structure of the data warehouse, and also how you meet your compliance obligations.

If you're wanting any guidance or advice when designing your tracking solution don't hesitate to reach out to your onboarding specialists or our Support team for advice.

 

Events v Entities

Tracking design comes down to capturing events and their related entities.

Event: An action, generally performed by a user.
Entity: Attaches to an event to add context or state of the environment when it occurred. Entities are added by some trackers and enrichments, or manually.

Screenshot 2024-10-05 at 2.29.19 PM.png

 

The canonical event model

Every Snowplow event comes with 130 default properties. We call this the canonical event model. Canonical properties capture a broad range of details including:

  • Event identifiers
  • Timestamps
  • Location info
  • User/session identifiers
  • Device information
  • Referral information
  • Marketing parameters
  • Page/URL info
  • Tracker & platform info
  • Event metadata

A full breakdown of canonical event model properties can be found in our product documentation.

 

Out-of-the-box events

Snowplow has a range of out-of-the-box events packaged with our trackers. You can learn more about these in the product documentation for your specific trackers.

 

Custom events

If you're wanting to capture an event that doesn't have a corresponding out-of-the-box event definition, you can create a custom event. There’s two custom types of Snowplow events.

Structured Events: Our older, legacy type. These are similar to older industry solutions, providing several out of the box fields which you fit your implementation to.
Self-describing Events: Our recommend type. These are described by JSON Schema which can be used to communicate and validate the properties and data tracked against the event.

Screenshot 2024-10-05 at 3.31.56 PM.png