What is A/B testing?
A/B testing compares two versions of a page by showing each to a share of visitors and measuring which converts better.
Definition
Visitors are split at random between version A and version B, and the difference in conversion rate shows which performs better, if enough people take part.
Why it matters
It replaces opinions with evidence about what actually works for your visitors.
How to get it right
- Test one meaningful change at a time.
- Decide the goal metric in advance.
- Wait for enough visitors.
- Start with the headline and the call to action.
Mistakes to avoid
- Stopping a test early
- Testing tiny details on low traffic
- Changing several things at once
Related terms
Orlume, the design-first AI website builder, gets the details on this page right for you: describe your site in one prompt and it decides the layout, type, colour and structure, builds it in React and Next.js and puts it live in one click.
Free to start · No card required · Live in one click
More conversion terms
Questions, answered.
The things people ask before they start building.
What is A/B testing?
A/B testing compares two versions of a page by showing each to a share of visitors and measuring which converts better. Visitors are split at random between version A and version B, and the difference in conversion rate shows which performs better, if enough people take part.
Why does A/B testing matter?
It replaces opinions with evidence about what actually works for your visitors.
How do you get a/b testing right?
Test one meaningful change at a time. Decide the goal metric in advance. Wait for enough visitors. Start with the headline and the call to action.
Still stuck? Email usBuild something that doesn't look AI-made.
Describe the site you want. Watch it get designed, not just generated. Then publish it to its own orlume.site address.
- Free to start
- No card required
- Publish to a live URL
- You own the code