A/B testing is a powerful tool for improving website performance. It involves comparing two or more versions of a webpage to determine which variation performs better.

​
Key Concepts

​
Baseline

The baseline is the default version of the element you’re experimenting, serving as the reference point against which other variations are measured. You might keep the baseline unchanged to assess whether the experiment you’re running will yeild any improvements. That is why Gleef does not enables baseline removal.

​
Variation

A variation is an alternate version of the baseline. In an A/B test, you typically experiment one or more variations to see if they outperform the baseline.

We highly recommend not to make more than 1 variation at a time, meaning the baseline and another variation. This is to ensure that the results are statistically significant quickly and accurately.
Therefore, if you want to experiment more than 1 variation, you should experiment each variation at a time, and whenever the experiment becomes significant, launch a new experiment with the baseline defined as the success of the previous experiment.

​
Control

The control group refers to the users who are shown the baseline version, allowing you to compare their behavior against those exposed to the variations. These visitors see the baseline.

You can conduct A/A test to make sure that everything is working well on your website.

​
A/A test

A/A test are conducted to make sure current experimentations are reliable. The idea is to run an experiment without any change, meaning that the 2 variations will be the same. You should not see any difference in the results.

​
Why A/B Test?

A/B testing helps companies optimize user engagement, conversion rates, and overall business performance by making data-driven decisions. Typical objectives include:

  • Increasing click-through rates (CTRs)
  • Improving conversion rates
  • Reducing bounce rates
  • Testing new headlines or CTA text

Netflix runs approximately 20 A/B tests per user at any given time!
Learn more at the Netflix Technology Blog

​
Benefits of A/B Testing

  1. Data-Driven Decisions: Base your changes on actual data, rather than assumptions.
  2. Improved User Engagement: Test different elements to see which engages users the most.
  3. Optimized Conversion Rates: Identify changes that can increase sales or signups.
  4. Reduced Risk: Test small changes before rolling them out widely.

Was this page helpful?

Sucess eventsGoals & Metrics