Went through the Snap PM loop earlier this year for a senior PM role on the ads team. This is a full breakdown.
Recruiter screen: Standard. Why Snap, comp range, timeline. They asked if I'd used Snap recently which, look, you should have the app on your phone and actually use it before any Snap PM interview. I demo'd a specific feature I thought was underutilized during this call and the recruiter seemed to appreciate the specificity.
PM technical screen (45 min): This surprised me. A PM from the team ran me through a product case: "Snap is seeing a 15% drop in daily Story viewers in the 25-34 age bracket over 6 months. Walk me through how you'd diagnose and respond." Classic PM case, but they pushed hard on metrics. How would you break down DAU? What's your funnel hypothesis? If it's a content supply problem vs. a distribution problem, how do your interventions differ?
They did NOT ask me to design a new feature from scratch in this round. It was diagnostic, not generative.
Onsite rounds (4 rounds): Product sense: Design a feature for Snap to increase revenue from creators. Open-ended. I framed around the creator monetization gap vs. TikTok and YouTube and proposed a tipping feature with Snap-specific twists. They wanted to know how I'd prioritize what to build first. Strategy: How would you evaluate whether Snap should enter a new vertical (they gave me AR glasses as context). Framework-heavy round. I used a market sizing + competitive moat lens. Execution: Tell me about a time a launch didn't go as planned. Classic PM behavioral. They wanted to see ownership and learning, not blame-shifting. Leadership cross-functional: How do you work with engineering when priorities conflict? This is a values alignment round as much as a skills one.
One thing I noticed: Snap PMs care a lot about whether you understand the consumer side, not just business metrics. They'd push back when I went too B2B in my framing. Know your consumer intuition.
Didn't get the offer. Feedback was that my strategy round was strong but product sense felt too feature-list-y. That was fair.