Spec: A/B Testing Events

This guide explains what data can be sent to Plainflow from A/B Testing tools. The semantic events detailed below represent the ideal for A/B Testing events.


Every A/B Testing tool is built around the concept of showing visitors an experiment with changes to the page. Each experiment can have multiple variations shown to different random samples of visitors.


The one semantic event for recording A/B Test experiments is:

Experiment Viewed

This event will be sent automatically every time a customer sees a variation of an active A/B Test when using some A/B testing tools via Plainflow. You don’t actually have to make these track calls! Note that the A/B testing tools will need to be loaded synchronously in order for these events to fire properly.


This event supports the following semantic properties:

Property Type Description experiment_id String The experiment’s ID. experiment_name String The experiment’s human-readable name. variation_id String The variations’s ID. variation_name String The variation’s human-readable name.

experiment_idStringThe agent’s ID.
experiment_nameStringThe agent’s real name.
variation_idStringThe agent’s real name.
variation_nameStringThe conversation’s ID.

Not using Plainflow yet? Get your free account here. 👈