I deleted our AI scribe analytics module and kept one activation deadline

I build HIPAA compliant AI for healthcare, and last month I deleted a feature I had just finished. It was a journey analytics module in our internal dashboard: stage by stage funnel, cohorts, drop off between every step. The module worked as designed. I removed it anyway.

The reason is that I could not name one decision it had changed. I opened it, read the stages, agreed with them, closed the tab. It described the past accurately and left me where I started, because a funnel tells you where people stopped and never tells you which of the eleven things on that screen to cut.

Two failures made it worse. Our event data disagreed with our billing system about how many customers we had, and once a number can be argued about in a meeting it stops being used, no matter which source is right. And the stages were named after our product instead of the user's day, so a drop between two of them was a fact about our UI, not about a doctor's schedule.

What replaced it is a deadline. A new user has to get their first clinical note out of the product in under ten minutes, and anything sitting between signup and that note is a candidate for deletion. That one number has killed onboarding steps, changed a default and settled two arguments, which is more than the dashboard did in its whole life.

The mechanism I had missed: analytics answers what happened, and a deadline answers whether we broke a promise we made. Only the second one names the thing to change, because a missed deadline points at a specific step while a drop off points at a stage.

Where does this break? If you kept your analytics and it earns its place, what decision did it change that a single deadline could not have. And if you run a number like mine, is it a deadline or an average, because averages have never moved anything for me.