An A/B testing solution that uses complex algorithms to dynamically allocate more traffic to the test version(s) performing best, while allocating less traffic to the underperforming version(s).

Because more traffic is gradually directed towards the variation that appears to be emerging a winner, multi-armed bandit tests can produce faster results.

Setting-up multi-armed bandit tests can be a good solution for marketers looking to run tests quickly and accurately.

However, this type of test set-up is complex and puts more strain on computer resources. So, it's difficult to run several multi-armed bandit tests simultaneously.

To more deeply understand the pro's and con's of using a multi-armed bandit testing methodology, see this article.