Product sense practice

Judgment is choosing and cutting. Prefer one sharp segment over boiling the ocean. Prefer company-native strengths (Meta → social graph; Google → information; Stripe → payments trust). Name what you will not build in v1 and why. When the interviewer challenges you, update or defend — do not fold instantly or dig in blindly. "I'd keep the segment because… but if growth is the only mandate I'd switch to…"

Taste shows up in solution quality: avoid feature salad; describe the key interaction in concrete UI terms for 60 seconds. "A shared trip canvas where each friend claims a responsibility and the poll for dates sits at the top" beats "a social collaborative planning experience leveraging synergies."

When stuck, narrate the stuckness productively: "I'm torn between supply-side and demand-side; given our distribution strength I'd start demand-side because…" That is judgment under uncertainty — exactly what the round measures.

Engineers meet these rounds for product-engineer tracks, 0→1 startup interviews, some tech-lead loops and internal mobility into PM. Prep overlaps with interview prep strategy: timed mocks beat passive reading. The same structure helps design reviews and PRDs.

The big idea

Practice regimen (~2 weeks):

1. Day 1–2: Learn the framework cold; watch one strong mock (Further learning) and outline it.
2. Day 3–7: One timed 30-minute prompt daily — voice notes fine; self-score on segment sharpness, solution diversity, decision, metrics.
3. Day 8–10: Partner mocks; focus on recovering from constraints ("mobile only," "no ML," "must monetize in v1").
4. Ongoing: Weekly 15-minute app critique of a product you use — mission, segment, pain, one improvement, metric.

Record yourself once; most people discover they never stated a north star or never picked among solutions until they hear the ramble played back.

Walking through it

Prompt: Design a product to help people reduce household food waste.

• Clarify: consumer app, US, mobile, startup (not Instacart unless told).
• North star: meals rescued per week per active household.
• Segments: busy dual-income families; sustainability-motivated singles; bulk shoppers. Pick busy families — high waste volume, willingness to pay for planning.
• Pain: they buy for intentions on Sunday; midweek inventory is unknown; leftovers are forgotten.
• Solutions: (1) receipt + fridge photo inventory with expiry nudges; (2) meal planner that consumes what you already have; (3) community leftover exchange. Pick (2) — clearer v1, less cold-start than exchange.
• MVP: optional barcode/receipt import; manual pantry list; "cook tonight from what I have"; waste check-in twice a week. Non-goals: delivery, social network.
• Metrics: WAU, recipes cooked from pantry suggestions, self-reported waste events; guardrail: time-to-log item.
• Trade-off: data entry friction vs automation accuracy — start manual+camera assist.

Improve YouTube for learners: north star learning session completion. Segment: career upskillers on long tutorials. Pain: losing place; no practice loop. Pick chapter-aware playlists with a practice prompt (creator-opt-in). Guardrail: watch-time quality, not clickbait.

Instagram engagement is flat — you cannot change ranking. Clarify the metric, then use surfaces you control: creation templates, collab posts, notification digest, close friends. The constraint is the point.

Common mistakes

• Ignoring the company. A generic app any startup could build. Use that company's assets.
• Frozen under pushback. Instant surrender or stubborn silence. Update aloud with a reason.
• Metric soup. One north star + 2 inputs + 1 guardrail, not every KPI you remember.

Check yourself

1. "Improve Instagram — you cannot change ranking." What is the trap?

Inventing fake levers (ranking). Move to surfaces you control: creation, collab, notifications, close friends.

2. Why record one mock?

Most people hear they never stated a north star or never picked a solution until the playback.

Key takeaways

• Choose a segment, a pain and one solution; name v1 non-goals.
• Use company-native strengths; recover from constraints by changing the surface, not the ranking you were forbidden.
• Timed mocks and weekly app critiques beat cramming frameworks.

Further learning

A full Meta/Facebook PM mock (movies product) — watch once for structure and pacing, then pause and answer yourself before hearing the candidate.

Exponent: product management interview prepAced (formerly Exponent) helps you land your dream tech career with interview courses, thousands of real questions, AI-powered practice, and expert coaching.Aced (formerly Exponent) Product Sense Interview Prep (Exponent blog)A data-backed guide to product sense interviews in 2026, with a six-step framework, real questions from Meta, Google, OpenAI, and DeepMind, and advice from interviewers and candidates who've been through it.Aced (formerly Exponent)

Back to the map: product sense interviews.