Research calls are for listening, not for showing

For years I opened every research call by sharing a prototype. It felt efficient. It was also the fastest way to learn nothing, because the moment people see your design they start being polite about it.

Now the first twenty minutes have no screen at all. I ask people to walk me through the last time they did the thing we are designing for: what they opened, who they messaged, where they got stuck.

People are very bad at predicting what they will do, and very good at remembering what they did.

Only at the end do I show anything, and by then I know which parts of it matter.