Data-Driven Choice of the Sort Order for E-commerce Product Listing Pages
Company: Amazon
Role: Product Manager
Category: Product Design & Strategy
Difficulty: medium
Interview Round: Other
You are a product manager at an online store and have been asked to work out the best sort option for its product list pages, the pages that display many products at once for shoppers to compare. Explain how you would approach the problem, and list the data points you would analyze that would lead you to your conclusion.
```hint Define "best" first
Before comparing orderings, decide whose outcome you are optimizing and how you would tell that one ordering beat another. A shopper who knows exactly what they want and one who is browsing may need different things from the same list.
```
### Clarifying Questions
- Is the task to choose the default ordering shoppers see before touching the sort menu, to decide which options the sort menu should offer, or both?
- Do "product list pages" mean keyword search results, category browse pages, or both, and may the answer differ by page type or category?
- What should "best" optimize for: purchases, revenue, shopper satisfaction, or a balance, and are there commitments such as sponsored placements that the ordering must respect?
- Is there existing data on how often shoppers change the sort today and which options they pick?
### What a Strong Answer Covers
- A clear statement of the decision being made and a success metric for "best", with guardrails.
- How shopper intent differs by page type, category and user, and why that matters for ordering.
- Concrete data points: shopper behavior on list pages, product attributes, and downstream outcomes such as returns.
- An evaluation method that accounts for position bias, moving from log analysis to an online experiment, and how the final call is made.
- Trade-offs between relevance, revenue, fairness to new products and sellers, and personalization.
### Follow-up Questions
- Shoppers who switch to "price: low to high" convert better than shoppers who keep the default. Should low-to-high become the default?
- How would you keep new products with no sales or reviews from being buried by an ordering that leans on popularity?
- Your experiment shows the new default lifts conversion for electronics but lowers it for apparel. What do you ship?
Overview: An e-commerce product sense question about choosing the best sort option for product list pages and the data that would justify the choice. It tests defining success for a ranking decision, reading shopper intent across categories, selecting behavioral and product data points, and validating the result with experiments.
