I've written extensively about choosing tools for A/B testing. With the discontinuation of Google Optimize, there are very few free options left. I dare to suggest that one of the best methods for quickly (and absolutely free) launching a web experiment is using the Experiments feature in Yandex.Metrica.
In fact, after the discontinuation of Google Optimize, Yandex Metrica remains one of the few options to conduct A/B testing absolutely free. This is done in the Varioqub section. Yes, there are some limitations (such as the number of simultaneous experiments), but again – IT'S ABSOLUTELY FOR FREE.
Conducting an experiment in Metrica is very straightforward.
1. In Metrica, go to the "Experiments" section and select "Create Experiment."
Fill in the required fields based on the information gathered during the preparation stage.
2. Refer to the experiment creation documentation if you encounter difficulties.
3. Assign a unique name to the experiment and add a description.
4. Set a duration from 1 to 90 days.
5. Audience Share: Determine the percentage of the audience to participate, from 1% to 100%.
6. Specify which pages the experiment will be conducted on, using operators for precise targeting.
7. Choose the region and type of devices (mobile or desktop).
8. Select the main metric and up to 10 additional ones for analysis.
9. Choose between Visual Editor, Redirect Links, and Code Flags depending on the test's goals.
10. After configuring the variants (available through a visual editor similar to Optimize), test how different versions of the pages will be displayed.
11. Avoid stopping the test prematurely, even if the results seem obvious. Wait for the test to run its allotted time.
12. Observe the changes in metrics. It’s important not only to see the difference between variants but also to determine if it's statistically significant (look at the "p_value").
If you work with GA4 to BigQuery exports, be sure to check out my SQL cheat sheet.
