The Research Sprint That Killed a Feature Before It Shipped

A subscription tier for recurring home services had already been scoped, estimated, and half-designed by the time we ran the first customer interview. Five interviews in, we recommended killing it.

Recurring intent isn’t recurring willingness to commit

Customers genuinely rebooked the same services on a predictable cadence — the usage data was real. But every single interview surfaced the same objection to a subscription: they wanted the flexibility to skip a month without feeling like they were cancelling something.

We shipped a lightweight "rebook in 30 days" reminder instead of a subscription tier — a fraction of the engineering cost, and it captured the same repeat-booking lift the subscription model was projected to deliver, without the commitment friction that would have suppressed it.