Designed for Use, 2nd Edition

Keep in Mind

If you use A/B testing to compare the performance of small, incremental changes to your design, keep in mind that A/B testing will only bring you to a local maximum of usability. Making many incremental changes will get you to a good design but may not get you to the best design possible.

Imagine all possible designs for your product. Some of them are more usable; others less. If you could plot them on a graph, you might get something like this:

Partitioning Users

Partitioning users into two groups is pretty simple when testing a desktop application. Give each user one of the builds at random when he downloads your app; typically each ...