From Contextual Design by Hugh Beyer and Karen Holtzblatt:

[…] Any change is a struggle. Engineers used to making what they are interested in feel constrained by having to think about what is useful and can sell. We all have to hold back the voice that tells us that producing code is progress – even if we cancel the project, even if it is the wrong code, even if we don’t know what would be useful to code. How does understanding work produce code? It is a struggle of personalities as we try to work in cross-functional teams to produce a shared direction. It is hard to remember that one smart guy working alone probably doesn’t have the whole answer. We simply have to realize that design is about people working together, and that’s what makes it hard.

I remember the first design team I worked with. I barely knew what a computer was, but I jumped in to help a team designing a very large and expensive computer. They were stuck, not on the guts of the engine, but on the control panel! So I listened to six engineers arguing about how to lay out the switches: “Won’t we crash the system by accident if the remote selection is on the same switch as off?” “Oh, they’ll only do that once.” And whether or not there should be a key on the switch: “Security is important.” “No, it isn’t.” “Yes, it is.”

As I listened, I realized that the team simply had no ground for their decisions. There was no way that reasoning and argument would get them to an answer. So I collected some data on how the panel was used: “Are you kidding? We won’t touch the remote. Someone might crash it.” “We turn the knob very, very slowly.” “Someone crashed it once, and the whole business stopped. No one touches that knob now.” And on security: “The computer is in a locked room; we don’t need it locked.” “Locking is a pain. We keep losing the key.” “We keep the key taped to the computer so we can find it.” “I catch my clothes on that lock; it sticks out.” The design was done in a day. We had a new switch for on and of and stopped agonizing about the key. I recently ran into a member of that team. He said he still talks about what happened 10 years later. The power of simply having data.

Emphases and paragraph division (partially) mine.