What to expect, what they’re really testing, and what a strong answer looks like — scored.
Customer obsession, working backwards from the customer, data-driven decisions, and operational excellence. Amazon uses a press release / FAQ process (PRFAQ) and expects PMs to write clearly, define success metrics before building, and reason about long-term customer value over short-term metrics.
The question below was asked by Amazon interviewers. The answer is graded on the five dimensions real PM interviewers use: structure, specificity, reasoning, decision quality, and delivery.
“How would you improve Amazon Prime's value proposition for existing members?”
I want to be specific about the problem before proposing solutions. 'Improve Prime' is broad — so I'd start by asking: is there a retention problem? Is engagement declining on specific Prime benefits? Or is this about fending off competition from Netflix, Walmart+, and others?
Let me assume the business question is: how do we reduce Prime cancellation rate among members who have been subscribed more than 24 months, since long-tenure members are most at risk of 'subscription fatigue' — they've already gotten the Prime benefits they signed up for.
I'd analyze which benefits long-tenure members actually use versus which ones they've stopped engaging with. My hypothesis, based on publicly available data: long-tenure members primarily use Prime for free shipping and have low engagement with Prime Video, Music, and Reading. That means the multi-benefit pitch that acquired them is no longer retaining them.
My proposal: a 'Prime Snapshot' — a personalized annual summary (like Spotify Wrapped) that quantifies the value a member captured. 'You saved $340 in shipping this year. You watched 48 hours of Prime Video. Here's what you'd pay without Prime.' This addresses subscription fatigue by making the invisible value visible. Cost to build: low. This is a data assembly + UI project, not new infrastructure.
Success metrics: open rate on the Snapshot email (leading indicator), and 90-day retention rate post-Snapshot versus control group (primary). Guardrail: NPS — I don't want users to feel surveilled by seeing their usage data.
Scopes the problem, states a hypothesis about the specific segment at risk, proposes one concrete solution, and defines metrics.
Names the 24-month segment, the $340 shipping savings framing, and the Spotify Wrapped analogy — all concrete and memorable.
Subscription fatigue insight is real and the 'make invisible value visible' logic is clearly stated.
Commits to one low-cost high-confidence solution rather than a menu of options.
Tight and well-paced; the Spotify Wrapped analogy is efficient communication.
This is a strong answer because it resists the urge to brainstorm ten Prime features and instead focuses on a specific retention problem with a specific user segment. The Snapshot idea is creative, plausibly low-cost, and has a clear success metric. The one gap is that the answer doesn't acknowledge the risk that Snapshot could backfire if users see they've used Prime less than expected — the NPS guardrail touches on this but could be more explicit.
Add one sentence acknowledging that Snapshot has a backfire risk (users realize they're not using Prime enough to justify the cost) and explain how you'd mitigate it.
Paste any Amazon PM interview question and your answer. Get scored on the same five dimensions — instantly, free, no signup.
Grade my answer free →First grade is free. No account needed.