Product design question

How would you improve the recommendations for Zomato?

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Product improvement for a recommendation system, requiring understanding of what signals actually predict satisfaction in food delivery.

Clarify the current gap: recommendations likely rely heavily on past order history and popularity, missing context like time of day, weather, or dietary changes.
Segment users: repeat customers with a stable pattern versus explorers who want variety and get frustrated by repetitive suggestions.
Prioritize the explorer segment's frustration, since repetitive recommendations are a common, fixable complaint that drives disengagement.
Propose incorporating contextual signals, like time of day and recent cuisine diversity, and explicitly surfacing a discover something new row alongside the usual favorites.
Add a feedback loop: a quick thumbs up or down on recommendations to refine future suggestions faster than passive order tracking alone.
Define success as click through rate on recommended items and order diversity per user over time.

Identifies a specific, real weakness, over reliance on order history causing repetitive suggestions, rather than a vague improve recommendations answer.
Segments users by whether they want variety or consistency, since one algorithm shouldn't optimize for both the same way.
Gives an illustrative number, for example assuming order diversity per user increases by 15 percent after adding a discover row.
Proposes an explicit feedback mechanism, faster and more direct than inferring preference purely from order history.

Proposing a vague better algorithm answer without naming what signal or gap is actually being fixed.
Ignoring that different user segments, repeat versus explorer, may want opposite things from recommendations.

How would you balance recommending familiar favorites against encouraging discovery of new restaurants?
How would you measure whether new recommendations are actually improving satisfaction, not just click rate?