You can read the definition of A/B testing in Wikipedia.

An A/B test is a real time test where incoming traffic is routed across 2 or more landing experiences randomly, or at a traffic distribution you define. With LiveBall's built in A/B testing feature you can split traffic across 2 or more Creatives (landing experiences) to determine which is the most effective at converting your visitors. A/B testing is great for testing an apple against an orange—like a landing page against a microsite, or different landing pages against each other. 

In LiveBall, A/B tests are set up on Traffic Sources. Traffic Sources are the URLs you create in LiveBall and send traffic to. 

In LiveBall, when you create a Traffic Source, you associate a corresponding Creative (landing page, etc) to it. Basically you link up, or 'marry' your URL to the Creatives. This gives you flexibility to serve up exactly what you want on each unique Traffic Source in LiveBall and change these out as you may need to. Just one URL let's you split traffic across any number of landing experiences, so you never have to deal with switching URLs or using multiple URLs for a test. You can associate just 1 or more Creative to your Traffic Source. Anytime you associate more than 1 Creative to a Traffic Source you are running an A/B test. Here's what else you can do:

• You can define the weight you want to traffic distributed across the various Creatives (it will default to an equal distribution, but you can use the sliders to change the weight). 
• You can have the A/B test auto-optimize or you can set it for manual optimization. 
• With auto-optimization, LiveBall will automatically 'turn off' traffic to the Creatives that are losing the test, based on your LiveBall console's statistical significance setting. Your console defaults to 80% confidence, but you can change that under your Settings tab (anywhere from 80-99% confidence). This is great because you can set up your test and know that LiveBall will take care of getting traffic to the best performing Creatives.
• If you choose manual optimization, LiveBall will not automatically route traffic to the statistically significant winner, and you will do that manually by monitoring your test results and making changes to the Traffic Source as your discretion. 

A/B testing is a great way to test different Creatives against each other to see what will convert the most traffic, and in LiveBall A/B testing is virtually effortless. Once your Creatives are published it takes just about 10 seconds to set up an A/B test and take it live—no code, no tech, no developers! Enjoy. We hope you run lots of A/B tests and get lots of conversion lift and learning from them.