How to Crush the Product Sense Interview by Starting with the User, Not the Feature
"How would you improve Instagram for creators?" Jennifer Okafor leans back. Her pen is down. She shipped products used by hundreds of millions of people. She has asked this question to 200 PM candidates. She already knows who is going to fail — the ones who start talking about features. "I would add a scheduling tool..." Dead. "What about a collab marketplace..." Done. "We could improve the algorithm to..." Next. The candidates who pass this round do something counterintuitive: they stop talking and start asking. "Which creators? What's their biggest pain point?" Jennifer's pen comes back up. Now there's a conversation.
Why This Conversation Goes Wrong
You jump to features immediately. "I would add a content scheduling tool and a creator monetization dashboard." You just proposed a solution without defining the problem. Jennifer now knows you think like an engineer building a feature list, not a PM solving a user problem. The interview is functionally over.
You define the user too broadly. "Creators on Instagram..." — that's 200 million people. Which creators? Micro-influencers with 5K followers? Full-time content creators with 500K? Professional photographers? Each segment has radically different needs. "All creators" is the PM equivalent of saying "everyone is our customer."
You skip metrics entirely. You propose three features but never say how you'd measure success. Jennifer is thinking: "How would you know if this worked?" If you can't define the metric, you can't defend the investment. And if you can't defend the investment, you can't prioritize.
You only add — you never subtract. "We should build A, B, C, and D." But Jennifer's follow-up will be: "You have engineering resources for one. Which one?" If you can't kill three of your own ideas, you don't understand prioritization — the core skill of product management.
The User First
Product sense interviews test whether you think like a PM. The single biggest differentiator: PMs who start with the user pass. PMs who start with the feature fail. The User First framework structures your answer as a funnel: define the user → identify pain points → generate solutions → prioritize ruthlessly → define metrics → articulate trade-offs.
Segment before you solve
"I'd like to focus on mid-tier creators — 10K to 100K followers — who create content as a significant part of their income but aren't full-time yet." Picking a specific segment shows intentionality. Explain why: "This segment is growing fastest, has the highest churn risk, and represents Instagram's best opportunity to prevent creator exodus to TikTok." Now you're thinking like someone who owns a P&L, not someone who read a product blog.
Name pain points from the user's perspective
"For mid-tier creators, the three biggest pain points are: inconsistent reach making it impossible to predict income, no direct way to convert followers into paying customers, and content fatigue from the volume required to stay relevant in the algorithm." Notice: these aren't feature gaps. They're lived experiences. Jennifer is checking whether you can empathize with the user or only analyze the product.
Propose solutions tied to specific pains
"For inconsistent reach, I'd test a 'creator boost' feature that guarantees minimum impressions for one post per week. For monetization, a native 'subscriber-only' content tier. For fatigue, AI-assisted content repurposing from long-form to Stories." Each solution maps directly to a stated pain point. No orphan features. No "it would be cool if..."
Kill your own ideas before she does
"If I had to pick one, I'd go with the subscriber tier. Here's why: the 'creator boost' could cannibalize ad revenue, which is a political non-starter. The AI repurposing is technically complex with uncertain ROI. The subscription model has proven demand on Patreon and Substack, and it directly solves the income unpredictability problem." Prioritization requires you to kill things you just proposed. That's the skill.
Define what success looks like — and what you'd watch for
"North star metric: percentage of mid-tier creators earning over $500/month through the subscription feature. Leading indicators: subscriber conversion rate and creator adoption in the first 90 days. Anti-metric I'd watch: if overall engagement drops because creators gate too much content, we'd need to adjust the content ratio." Metrics prove you think beyond the launch. The anti-metric proves you think about second-order effects.
The moment that changes everything
Jennifer isn't grading your idea. She's grading your thinking.
Here's what most PM candidates miss: Jennifer doesn't care if your answer is the right answer. There is no right answer. She's watching your process — do you define before you solve? Do you prioritize by killing ideas? Do you state your assumptions explicitly? The candidate who says "I'm assuming this segment represents 15% of monthly active creators — I'd want to validate that with data" scores higher than the one who presents a perfect feature proposal without acknowledging a single assumption. Product sense isn't about having the best idea. It's about having the clearest thinking. The PMs Jennifer promotes are the ones who can say "I don't know, but here's how I'd find out" without flinching.
What to Say (and What Not To)
Instead of
"I would add a scheduling tool for Instagram creators."
Try this
"Before features — which creators? I'd focus on mid-tier, 10K-100K, because..."
Instead of
"All creators would benefit from better analytics."
Try this
"The biggest pain point for this segment is inconsistent reach making income unpredictable."
Instead of
"We should build A, B, and C."
Try this
"If I had to pick one, it's the subscription tier. Here's why the others don't win..."
Instead of
"I think this would be really successful."
Try this
"Success metric: percentage of mid-tier creators earning over $500/month. Anti-metric: engagement drop."
Instead of
"That's my answer."
Try this
"I'm assuming this segment is 15% of MACs. What would you push back on?"
The Bigger Picture
An analysis of 1,200 PM interviews at top tech companies by Exponent found that candidates who started their product sense answer by defining a user segment were 4.2x more likely to receive a "strong hire" rating than those who started with a feature proposal. The correlation was so strong that interviewers could predict the outcome within the first 60 seconds based solely on whether the candidate asked "which users?" or said "I would build..."
Product sense interviews at companies like Google, Meta, and Stripe are the most heavily weighted round in PM hiring, accounting for 35-40% of the overall evaluation. They test a skill that cannot be taught in a prep course: the instinct to understand the user before building for them. Interviewers consistently report that the #1 mistake is feature-first thinking, and the #1 signal of a strong PM is the ability to kill their own ideas.
The meta-skill being tested in a product sense round is intellectual honesty. Can you say "I don't know the actual churn rate for mid-tier creators — I'd want to look at that data"? Can you admit that one of your three ideas is weaker than the others? Candidates who present a polished, flawless answer are actually scoring lower than those who show their work, including the dead ends. Jennifer is hiring a thinking partner, not a presentation machine.
Practice This Conversation
15 minutes · AI voice roleplay with Jennifer Okafor
Reading about this is step one. Practicing it changes everything. Sonitura lets you rehearse this exact conversation with Jennifer Okafor, a realistic AI group product manager at a top consumer tech company who reacts to your words in real time. It takes 15 minutes. The next time an interviewer asks "How would you improve X?" — you'll start with the user, not the feature.
Practice This Scenario Free →Related Guides
Job Interview
How to Ace the Startup Culture Fit Interview Without Sounding Rehearsed
7 min read · 12 min practice
Job Interview
How to Ace the Coding Interview by Thinking Out Loud
8 min read · 15 min practice
Job Interview
How to Solve the Analytics Case Interview by Asking "So What?" at Every Step
8 min read · 15 min practice