Amazon · Product sense · Entry · 45 to 60 minutes

Help shoppers find missed accessories

Amazon is a large online marketplace where people browse and buy products from many categories. Answer this product management case and get it graded line by line against a weighted rubric.

product sense
user empathy
metrics
experimentation
prioritization

The situation

You are a new PM on Amazon's shopping experience team. Your director asks you to improve the path from product discovery to checkout for customers who come to buy one item but may also need a related item to complete the job, like a phone case and a screen protector, or a laptop and a mouse. Assume internal data shows that many related items are bought together, but customers often discover the second item only in a later session or after delivery, which creates extra effort for them and leaves revenue on the table. The team has access to browsing, cart, and purchase behavior across Amazon's catalog, but page space in the shopping flow is limited, and any change that feels pushy could hurt trust or reduce conversion. You need to recommend whether Amazon should add or expand a feature that surfaces related items during shopping, where it should appear, how it should decide what to show, and how to test it before a broad launch.

What to produce

Write a 1 to 2 page product brief.

Your answer must

Define the user problem and the business opportunity, and explain why this matters now.
Recommend a specific feature concept, including what kinds of items it should suggest and what customer need it solves.
Choose where in the shopping journey the feature should appear first, and justify that choice.
Explain what signals or data the feature should use to decide what to show, at a high level.
Define success metrics and guardrail metrics for an MVP launch.
Outline an MVP and an experiment plan, including key risks and how you would learn quickly.

Constraints

The solution must work across Amazon's very large catalog with little or no manual pairing for most items.
For the MVP, you can add only one new recommendation module in the shopping flow, product page, cart, or checkout.
Do not assume major changes to pricing, fulfillment, or the rest of the site design.
The proposal must protect customer trust and shopping performance, including page speed and conversion.

Format

Use short sections with bullets where helpful, for example problem, proposal, placement, metrics, risks, and test plan.

How Amazon actually played it

Explains the people-also-bought recommendation: collaborative filtering, timing, and ongoing testing. Uses a personal shopping story to show why the feature raises basket size.

Recommend at the moment of need
Behavior signals beat stated preferences
Keep testing a feature that already works

Built from Amazon Secretly Made Me Spend More Money by Sid Arora (JAPM).