What wasn't working.
The team was building features in silos, largely driven by internal opinion and assumption. There was no structured way to hear from actual users.
Basic analytics showed what people did inside the product, but not why, or what they wished the product did differently. Roadmap debates kept circling back to whoever argued loudest in the room.
What we actually did.
I introduced lightweight in-product surveys at key moments in the user journey, rather than relying on email surveys sent after the fact.
I paired the survey responses with a repeatable review process alongside existing behavioral data, so qualitative signal, what users said, and quantitative signal, what users did, were read together, not separately.
The goal was a standing input into planning, not a one-off research project that would go stale after a single sprint.
Where the interesting calls were.
I chose in-product over out-of-product surveys deliberately. Reaching users in the moment, inside the product, produced far higher response rates and more relevant answers than email ever had.
I insisted on pairing qualitative and quantitative data every time. Neither source alone was reliable enough to act on with confidence.
I built a repeatable review cadence instead of running a single study. A snapshot tells you something once. A program keeps telling you.
The numbers after we shipped.
The team gained a continuous survey channel reaching real users inside the product, not a one-time research exercise.
Product decisions started tracing back to real behavior and stated need, rather than the loudest internal opinion.
The insights loop reshaped where the team invested its energy for the next two quarters, redirecting effort away from features nobody had asked for.
Things I'd carry into the next case.
Asking users directly, at the right moment, beats guessing every time, and it costs less than most teams assume.
Qualitative and quantitative data answer different questions. Use both before committing to a roadmap bet.
Insight programs need a repeatable cadence to stay useful. A single survey is a snapshot. A program is a compass.