Product Manager Interview (+ AI PM) · Product Sense & Design · Lesson 3 of 3
Improve an existing product & prioritize ideas
'How would you improve Instagram / Google Maps / our checkout?' is product sense with an existing surface. Same spine — goal → user → pain → solutions → prioritize → recommend — but you start from what exists and where it falls short.
- Anchor the goal: improving for engagement, retention, new-user growth, or revenue? Different goals → different ideas.
- Segment + find friction: pick a user, map their journey, find where they drop off or feel pain.
- Generate 3–5 ideas spanning easy wins and bold bets — show range.
- Prioritize explicitly: rank by impact vs effort (or reach/impact/confidence), and pick.
- Define success: name the metric that would tell you the improvement worked.
Always land a recommendationA pile of ideas is a fail; a prioritized pick with a success metric is a pass. Interviewers want a decision-maker. End with: 'I'd build X first because it has the highest impact for the effort, measured by Y.'
◆ Lock it in
- Improvement questions use the same spine, starting from existing friction.
- Generate a range of ideas, then prioritize explicitly (impact vs effort).
- Always land a recommendation plus the metric that proves it worked.
Feynman drill — say it out loudGiven 'improve Google Maps for tourists', outline how you'd move from goal to a single prioritized recommendation.
Step 1 rate your confidence · Step 2 pick your answer
The most common failure on an 'improve product X' question is:
Step 1 — how sure are you?