Prime Now recommendations that steer demand toward a lean, space-efficient assortment

Design a product for Amazon's Prime Now that suggests items for customers to buy, where the business objective is to optimize how many SKUs (distinct stock-keeping units) the service carries and how it uses its warehouse space. Walk through the users and their problems, the product you would build, how you would prioritize it, and how you would measure success. Assume the service's publicly known model: a limited selection of everyday items delivered within hours from small fulfillment sites close to customers. ```hint Link what customers see to what the shelves hold Start from what makes a SKU expensive to keep in a small warehouse, then ask what the items a customer is shown could do about that cost without making the selection feel worse. ``` ### Clarifying Questions - Does "optimize" mean carrying fewer SKUs, carrying a better-chosen set of SKUs, or getting more sales out of each unit of warehouse space? - Where do recommendations appear: the home screen, search results, the cart, substitutes for unavailable items, or all of these? - Are perishables such as fresh produce in scope alongside household goods, given how differently they age and how much space they take? - Can the product change what each warehouse stocks, or only influence what customers buy from the current stock? ### What a Strong Answer Covers - A clear reading of the dual objective: customer value from fast delivery and efficient use of limited SKUs and space in small warehouses - Users and stakeholders on both sides (customers with different shopping missions, inventory planners and warehouse operations) and their pain points - A recommendation design that accounts for local stock, substitutes, sales speed, item size and shelf life, and that feeds what it learns into assortment decisions - Metrics that pair customer outcomes with operational ones such as sales per unit of space, sell-through and stockouts, plus guardrails against hurting customers - Risks of letting operational goals steer what customers see, and how to test a change that affects shared warehouse inventory ### Follow-up Questions - Customers increasingly accept a recommended substitute for one product. How would you decide whether to stop stocking the original? - How would you run an experiment when every customer served by a warehouse draws on the same inventory? - How should the design change for perishable items such as fresh produce?