In a surprising study, Jakob Nielsen found out most (85%) of the learning gleaned from usability tests came from your first 5 user tests; 3 users tests will still give you a whopping 67% of the learning there is. He then goes on to advocate small and iterative testing over a single massive one (say, three tests with 3 users each and improvements in between instead of one lone test with 9):
You want to run multiple tests because the real goal of usability engineering is to improve the design and not just to document its weaknesses.
Combine that with Joshua Schachter’s, founder of Del.icio.us, idea of “Starbucks usability tests” (“offer someone coffee to sit down and play with your product.”) and you’ll know why I’ll be this Friday on Minerva’s Starbucks harassing
three every pretty girl I can find with Domburi (TeenVogue recently had a feature on young fashionistas being excellent for usability tests, can’t find the link though).