Solving an identified problem – A/B testing

When the designers have a proposed solution, it can be A/B tested.

A/B testing is a method where you have two versions of a page/site to present to the user.

Version A, usually the current version, and version B with a proposed solution.

The A/B testing tool presents 50% of users with the new B version while keeping the other 50% on the A version.

The test is deployed to the site for a period of usually 2 weeks.

During this time, you compare which version converts better.

This kind of testing at scale is very useful if you believe you have a solution.

And your site already has users.

Making it easy to A/B test increases the likelihood that new solutions are tested.

Instead of deploying changes that you think help the customer, you may as well test if it actually does.

– Samy

Component Analytics daily

Every weekday, I publish about Component Analytics and analytics collection strategies.