- Home
- Manage the site
- Test variations
- A/B testing with DXPR Builder
A/B testing with DXPR Builder
Test two or more versions of a headline, a picture or a whole section, and let your site send more visitors to the version that works. You build the versions in the editor. The free RL module picks a version for each visitor and learns from what they do.
| I want to ... | Use | Needs |
|---|---|---|
| test a second version of a headline, image or section | Setting up A/B tests | the RL module |
| compare three or more versions at once | Multivariate testing | a running A/B test |
| let visitors' clicks decide the order of accordion or tab items | Smart sort | an Accordion or Tabs element |
| see which version wins | Analysing results | the View RL reports permission |
| plan tests that give clear answers | Best practices | nothing extra |
Which should I use
Use variants when you have a new idea to try: a different headline, picture or call to action. Each visitor sees one complete version.
Use smart sort when the content is fine but the order may not be. Every visitor sees all items, and the items people use most move up.
An element runs one kind of test at a time, and tests cannot be nested inside each other.
Before you start
- Install and enable the RL module:
composer require drupal/rl, thendrush en rl. - Open Reports > Status report
(
/admin/reports/status) and check that RL: rl.php accessibility says "rl.php is accessible". The RL module counts visits through this file. - DXPR Builder needs no extra submodule.
How a test works
- You open an element's More menu and choose Create variant. Variant B starts as a copy.
- You change variant B and save the page. DXPR Builder registers the test with the RL module.
- Each visitor gets one variant, chosen by Thompson Sampling.
- The RL module counts impressions and conversions, and shifts traffic towards the better variant.
Each language of a page runs its own test, so every audience finds its own winner.
Conversion types
A conversion is the visitor action that counts as success. You choose it under Experiment settings in the element's More menu.
| Conversion | Counts when | Offered for |
|---|---|---|
| Time on element | the variant is still visible after the set number of seconds | variants |
| Time on page | the visitor stays on the page for the set number of seconds | variants |
| Click on element | the visitor clicks the variant, or an element matching a CSS selector | variants |
| Click on item | the visitor clicks an item | smart sort |
| Open item | the visitor opens an item | smart sort on Accordion and Tabs |
A test you never configured uses Time on element with 10 seconds. Smart sort on Accordion and Tabs starts with Open item, other smart sort elements with Click on item.
What's next?
Setting up A/B tests step by step to continue exploring DXPR Builder.
In this section
- Setting up tests: Create a variant, choose the conversion and start the test
- Multivariate testing: Compare three or more variants of one element
- Analysing results: Read the RL reports and decide which variant wins
- Best practices: Plan, run and read A/B tests that give clear answers