Cold-Start Fashion Recommendations on Myntra's Product Detail Page

Myntra is a fashion e-commerce platform. Design the recommendations shown on its Product Detail Page (PDP) for a customer who has only just signed up. Assume this customer reached the page by going from the homepage to a category page and then opening a product. ```hint Read the path A new customer has no purchase history, but the visit so far is not empty: the homepage, the category they chose and the product they opened each say something. Work out what those steps reveal, and what remains unknown, before deciding what to show. ``` ### Clarifying Questions - What should the recommendations mainly achieve: help the customer buy the product in view, find a better alternative, add complementary items, or explore more of the catalog? - What do we know about a new customer at this point: details given at sign-up such as gender or location, filters and sorting used on the category page, or the campaign that brought them in? - Is this for the mobile app, the mobile website or both, and how much of the page can recommendations take? - Should the design cover only the logic that picks recommendations, or also the page layout the customer sees? ### What a Strong Answer Covers - Why a new customer differs from a returning one: no history, little trust in the platform, and fashion-specific worries such as fit, quality and returns. - The signals available along this path and how each one shapes what is recommended. - The recommendation modules for the page, where each sits relative to the main product, and what each falls back on when signals are thin. - How recommendations improve during the session and over the first few visits without trapping the customer in a narrow set of styles. - Success metrics for the recommendations and the page, with guardrails such as returns and the conversion of the product being viewed. ### Follow-up Questions - A new recommendation module raises clicks, but add-to-bag for the product being viewed falls. What would you do? - The customer arrived through an ad for this exact product. How, if at all, would the recommendations change? - How would you measure whether recommendations help new customers or merely reshuffle purchases they would have made anyway?