Designing Product Search for a Large Shopping App: Shopper Intent, Ranking and Metrics

You are the product manager responsible for search in an e-commerce app. How would you design the search experience? Walk through how you would set the goal for search, decide which shoppers to focus on, find where search fails them, choose and prioritize improvements, and measure success. ```hint Sort shoppers by intent Shoppers type queries with very different levels of certainty: an exact model number, a category with a few attributes, or only a need. Ask what search has to do for each kind of query before picking features. ``` ### Clarifying Questions - Is this a large marketplace app that already has search, like Amazon's, or a new store building search for the first time? - What does the catalog look like: one vertical such as groceries or fashion, or many categories from many sellers? - Is it mobile only, and which markets and languages does it serve first? - Is the target more purchases from search, faster finding, or revenue that includes sponsored listings? ### What a Strong Answer Covers - A goal for search tied to the business, with the scope stated: platform, catalog, new or existing search - Shopper segments by intent and a search journey from query to purchase, with the main failure points - Prioritized improvements across query entry, query understanding, ranking and the results page, with their trade-offs - Enough grasp of retrieval, ranking and latency to judge feasibility and effort - Success metrics such as search conversion and failed-search rates, with guardrails and a testing plan ### Follow-up Questions - How would you place sponsored products in results without hurting relevance or trust? - A query returns nothing. What should the shopper see? - How would you check a new ranking model before and during rollout?