What the interviewer is checking
You start with users, not features
Strong candidates name who the product is for and what that person is trying to get done before anything gets drawn.
You choose
Listing ten ideas shows range; picking one and saying why shows judgment. The second is what senior roles are hired for.
You see the trade-off
Every good option costs something: build time, a second user group, revenue. Saying what you'd give up is the senior signal.
You'd know if it worked
A single success metric, and the counter-metric that would tell you it backfired.
How to run the answer
Clarify the goal in one or two questions
Ask what the company wants from this product (growth, engagement, revenue) and any constraint. Then stop asking and start working.
Pick one user segment and say why
Name two or three segments, then choose the one with the sharpest pain or the most strategic value. Say it out loud: "I'll focus on X because…"
Name the problems, rank them
Two or three real frustrations for that user, ranked by how often they happen and how much they hurt.
Offer a few solutions, then commit to one
Sketch two or three, from cheap to ambitious. Choose one, describe the first version, and name what you'd cut.
Close with the metric and the risk
One number that says it worked, one that says it hurt something else, and the first experiment you'd run.
Common misses
Jumping to the solution
Opening with a feature idea skips the part the interviewer is grading.
Designing for everyone
"All users" isn't a segment. Without one person in mind, every later choice is a guess.
No trade-off
An answer where nothing is given up sounds like a pitch, not a decision.
Worked example
How would you improve a grocery delivery app?
- Asks whether the goal is order frequency or basket size, then picks frequency.
- Chooses busy parents ordering weekly, because a missed weekly order is a lost habit.
- Ranks problems: rebuilding the same list every week beats slow delivery windows.
- Commits to a one-tap "repeat last order, edit what changed" flow and cuts personalized recommendations from the first version.
- Measures weekly repeat orders, watches basket size as the counter-metric.