By: Stormly

A/B Testing

A/B Testing is a method to determine which of two variations is the most effective in an experiment, where one variation is control and the other experimental. The Experimentation can be done on event conversion, retention, sessions, revenue, viral coefficient, and much more. This A/B testing tool automatically determines sample size and how many days your experiment should run. It also provides conversion lift (relative conversion improvement) and helps decide if the economic impact is big enough to implement the new variation

  • Shows which A/B version is the winning one and with how much significance.

  • Experiment on event conversion, retention, sessions, revenue, viral coefficient, and much more.

  • Automatically determines sample size and how many days your experiment should run

  • Shows conversion lift (relative conversion improvement), and helps decide if the economic impact is big enough to implement the new variation

To track experiment data, use the Set-Up Data option in the projects page.

To A/B test viral coefficients, you also need to track user referrals. For more information see Referring Users / Virality in the Setup Data page of your project.

Performs action one or more times
For example adding an item to the shopping cart at least 2 times or completing a payment.
Reaches minimum amount
For example having at least $99 in sales for all items bought over lifetime of user.
Comes back one or more days later
For example coming back at least 8 days after signing up.
Has one or more sessions
For example having at least 2 sessions, where a session is defined as playing a song at least 8 hours apart.
Completes the last step of a funnel
For example installing the app, signing up, playing a song and then paying for the premium subscription.
Invites one or more other users
Conversion when someone invites one or more other users.
Want to build your own Insights? Visit the Developer Portal.