Table of Contents

A/B testing

Lisa Fockens Updated by Lisa Fockens

A/B testing allows you to compare two or more variants of a popup to determine which one performs better. You can test different elements like targeting rules, popup design, and coupon amount to identify what works best for your audience.

Step-by-step process

Here's how to run an A/B test with Wisepops:

  1. In your Wisepops dashboard, hover your mouse over the campaign you want to test and click "A/B test" next to it.
  2. Click on one of the variants that has been created out of the original campaign
  1. Apply the modifications you want to test to your variant. In our case we will add an image:
  1. Set the percentage of traffic you want to attribute to each variant from your dashboard
  1. You can optionally add additional variants by clicking "Duplicate as new variant" next to one of the variants.

Concluding the test

Once you have collected enough data, you can check the reporting page to see if the test is significant. When significant results have been achieved, Wisepops will display a confirmation message.

To complete the test:

Option 1

select the winning variant and click "Duplicate as new campaign." This will stop the A/B test and duplicate the winning variant into a new campaign.

Keep in mind when you duplicate the campaign, you'll need to activate it, and it will have its own separate stats.

Option 2

Alternatively, you have the option to configure the traffic ratio of the chosen variant to 100%. To do this:

  1. Navigate to the variants and click on the percentage of users associated with each.
  2. Adjust the percentage to 100% for the selected variant and set it to 0% for the other one(s).
  3. Once done, don't forget to save your changes.

Best practices

Here are some best practices to keep in mind when conducting an A/B test:

  • Be patient: it can take some time to collect enough data to draw significant conclusions. The significance of your A/B test will be estimated in your reporting.
  • Keep it simple: if you want quick results, limit the number of variants. The more variants you have, the longer it will take to collect data and get significant results.
  • Test one element at a time: if your variants have a lot of differences, it will be harder to identify which elements had the biggest impact on performance.

If you have any questions, please get in touch.

Create a control group