Diagnosing a sudden 20% week-over-week fall in Swiggy orders with metrics

You are a product manager at Swiggy, the Indian food and grocery delivery platform. Last week, orders came in 20% below the level of every previous week. How would you carry out a root cause analysis? The interviewer wants you to concentrate on metrics: which numbers you would look at to confirm the drop and sharpen the problem statement, and which ones would separate one explanation from another. ```hint Locate the drop before explaining it Resist listing causes straight away. First make sure the 20% is real, then break the order count into components and segments; where the loss concentrates tells you which hypotheses deserve your time. ``` ### Clarifying Questions - Does "orders" mean orders placed or orders delivered, and does the count cover every Swiggy business line (restaurant delivery, quick-commerce grocery and others) or only restaurant delivery? - Did the drop appear suddenly on a particular day, or build up gradually through the week? - Is the decline spread evenly across cities, platforms (Android, iOS, web) and customer types, or is that still unknown? - Did anything already known happen last week, such as a public holiday, a major cricket match, extreme weather, an app release or a pricing change? ### What a Strong Answer Covers - Validating the metric first: its definition, logging and pipeline changes, and whether the comparison baseline is fair given the calendar - A decomposition (users times orders per user, or a funnel from app open to delivered order) that isolates where the 20% is lost - Segmentation by geography, platform and app version, customer cohort, business line, restaurant and payment method to localize the drop - Hypotheses across demand, supply (restaurants and delivery partners), internal changes and external factors, each tied to a metric and checked in a sensible order - Next steps once the cause is found: confirming it explains the full drop, the fix, and how to verify recovery ### Follow-up Questions - The funnel shows the loss concentrated at payment and in one payment method. What do you do in the next hour, and in the next week? - Conversion at every funnel step is flat, but app opens are down by about the same 20%. Where do you look next? - How would you tell a drop caused by a competitor's promotion apart from one caused by a shortage of delivery partners?