The hard part — Talktrack

Leadership had already decided. Build the Loom integration.

We didn't think it solved the real problem, so instead of building it we ran longitudinal studies and surveys to find out what async work actually breaks on. It came back clearly: constant context switching, having to stop your own work, and the missing emotional register that makes feedback legible.

Then we built both — a Loom-style implementation and our version — and tested them against each other rather than arguing about it. Users preferred ours decisively. That's what moved leadership, and it moved our team from platform into Miro's core product org.

Pushing back cost us weeks and could have gone badly. It went well because we came back with evidence instead of an opinion.