Usability Article Summary

This article discusses why you only need to test your website with five users. Many people belief that testing is extensive and costly, but it turns out the best thing to do is run as many small tests as possible with fewer users. The researchers used the equation N (1-(1- L ) ⁿ ) to prove their statement. When you test with only one user, this user will give you some insight, but they will not hit everything. Then, the second user will give some of the same insight as the first user, but they will offer some additional insight. Next, the third user will offer many similar thoughts, but they will still have some new things to share. This pattern of overlap will continue as you increase the amount of users. You will start learning less and less, so there is really no point in continuing past five users because it will mostly be the same information. Although the graph indicates that you must test with at least 15 users to hit all the problems, it will save you money by running three small tests of five users each.