Sometimes, customers don't know what they might need. The danger of talking to potential customers when designing a piece of software lies in the problem that y

Sometimes, customers don't know what they might need. The danger of talking to potential customers when designing a piece of software lies in the problem that you're going to get a long list of convolute feature requests that, if implemented, doesn't result in anything usable. On top of that, the exemplary barber is unlikely to ever use anything besides paper. We might not be normal people, and normal people are unlikely to give meaningful design input. In my experience it makes sense to talk to them twice, once to determine general needs, and then a second time to actually have them try out the prototype.