logo

Experiments

Overview

In Checkmango, an experiment is a multi-variant test. Each experiment tests a single, unique hypothesis by randomly assigning participants to a variant. This could be the original (known as the control) variant or one of the other variants. You may think of experiments as a "test" or "campaign" within your website or application.

For example, when testing "does changing the placement of the checkout button lead to more conversions?" you would create an experiment with a key of CHECKOUT_BUTTON_PLACEMENT with two variants ORIGINAL_CHECKOUT_POSITION and NEW_CHECKOUT_POSITION. Your application will then decide where the button is placed based on the participant's assigned variant.

Creating experiments

When creating a new experiment, you start with an idea. What is it that you want to improve about your application or website?

Keep It Targetted

Experiments that target a small, defined area will see better results with higher accuracy.

To create an experiment in the Checkmango dashboard, login and head to the Create Experiment panel.

Once your experiment is created, you can then create the individual variants that you want to test against each other.

Updating experiments

While an experiment is in draft mode, you're free to make changes to it. Once an experiment is running, you will no longer be able to modify the experiment or variants.