If I did it again, I’d bring engineers in earlier. The calendar shipped, but only after a late debate about scope. Showing them the coordinators’ evidence sooner would have shortened it.

I’d also ship the dashboard in smaller pieces. With one engineer, waiting for the whole thing meant shipping none of it. The research team had already told me what mattered most: alerts, and knowing that participants were syncing consistently, especially new ones. People who keep up with a study early on are more likely to keep up the whole way through, so spotting a new participant who’s falling behind lets the team nudge them, or find out why and help, before the habit slips.