Metrics question

Amazon has been seeing an increase in their volumes for the past one year. Their logistics team is facing challenges now. You are the PM to help them manage logistics. What would you recommend as a product?

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Structured problem-solving on an operational scaling challenge, checking you connect the metric to a concrete product recommendation.

Clarify the specific challenge, whether it is delivery delays, warehouse capacity, or last-mile routing, since the fix differs by bottleneck.
Identify likely bottlenecks given rising volume: warehouse throughput, delivery route optimization, and demand forecasting accuracy for staffing and inventory placement.
Prioritize the bottleneck most likely to be causing customer-visible pain, typically last-mile delivery delays, since that is what drives satisfaction and complaint volume.
Propose a product recommendation, for example a dynamic route optimization and load-balancing tool that better predicts delivery windows and redistributes volume across facilities.
Address a secondary lever, improved demand forecasting to pre-position inventory closer to demand spikes, reducing strain on any single facility.
Define success as on-time delivery rate and cost per package delivered, and confirm which specific pain point the interviewer wants prioritized.

Names a specific likely bottleneck, last-mile delivery and routing, rather than proposing a vague 'improve logistics' answer.
Proposes a concrete tool, dynamic route optimization with real-time load balancing across facilities, tied directly to the stated volume growth problem.
Adds demand forecasting as a complementary lever, addressing the root cause of strain, not just the symptom.
Sets measurable success metrics, on-time delivery rate and cost per package, that logistics teams would actually track.

Giving a vague answer like 'build better logistics software' without naming a specific bottleneck or mechanism.
Addressing only one part of the problem, like routing, and ignoring forecasting or capacity planning.
Not defining any measurable success metric for the recommendation.

How would you prioritize between fixing routing versus fixing forecasting first?
What data would you need to validate which facility is the actual bottleneck?