what 85 hospitals taught me

i built the first version of Savior for one hospital, from inside that hospital. i didn't interview the front desk. i sat behind it.

you cannot understand a system from the outside. you can only describe it.

from the outside, hospital operations look like a process problem. from behind the desk, you see it's a people-under-pressure problem wearing a process costume. the official workflow and the real workflow share almost no steps. the gap between them, the workarounds, the sticky notes, the 'we don't actually do it that way', that gap was the entire product.

what changed in how i do research

i stopped trusting anything a user could tell me in a conference room. people are honest and wrong about their own behaviour, not because they lie, but because the real process is invisible to them too. it's just the water they swim in.

so now i go to the water. i'd rather spend a week behind the desk than a month in interviews. the insight is never in what people say about the system. it's in the small ugly workaround that's been quietly load-bearing for years. find that, and you've found the product.