With how many people you should test with

With how many people you should test with

Official chart from NNgroup
Official chart from NNgroup

The best results come from testing no more than 5 users and running as many small tests as you can.

This number has been defined from NNGroup, the world leaders in research-based user experience design.

Even if 5 is the ideal number, the truth is that even 1 is a good way to start, get feedback and iterate.

Here is is why:

  • 0 test = 0% of usability issues uncovered
  • 1 test = around 25% of usability issues uncovered
  • 2 tests = around 50% of usability issues uncovered
  • 3 tests = around 70% of usability issues uncovered
  • 4 tests = around 75% of usability issues uncovered
  • 5 tests = around 85% of usability issues uncovered

Treat the people you test with as credits, and spend them wisely

My approach when designing a product is to start by designing a quick dirty prototype. Once I feel I have something I can test, I do a user testing session to get some feedback, and I either iterate on my prototype and fix the problems I found or I do another user testing session.

If I have to do 5 testing sessions for every prototype iteration I create, then it can get really time consuming.

I treat the people I test with as credits and I use them wisely. I don’t spend all of my credits immediately. I don’t want to test one prototype many times. But instead I want to test many prototypes, a few times.

The goal is to design with user testing. To let people show you the right path to follow as early as possible. That's why I am such a big fan of "dirty" prototypes and rapid user testing. The sooner you do the testing, the sooner you know what's wrong and the sooner you find the right path to follow.

Don't spend days or weeks designing something, and then testing it with many people. Instead try to do many small iterations by letting user feedback to guide you.