Amazon Revenue Falling 10% a Month: Building a Root-Cause Metric Diagnosis
Company: Flipkart
Role: Product Manager
Category: Analytics & Experimentation
Difficulty: medium
Interview Round: Other
Amazon's revenue is falling by 10% every month. You are the product manager asked to find out why. Explain which metrics you would look at, how you would break the problem down, and the order in which you would work through the possible causes.
```hint Find where the money went
Amazon earns revenue in several very different ways, and each has its own volume and price drivers. Locate where the missing revenue sits before you propose a cause.
```
### Clarifying Questions
- Does "10% monthly" mean each month is 10% below the month before, compounding, or that each recent month is 10% below the same month a year earlier?
- When did the decline start, and did it appear as a sudden step on a particular date or build up gradually?
- Is the drop spread across every business line (first-party retail, third-party seller fees, advertising, subscriptions, cloud) and every country, or concentrated in a few?
- Is the figure reported revenue in a single currency, so that exchange-rate swings or an accounting change could account for part of it?
### What a Strong Answer Covers
- Pinning down what the 10% means and when it began, including what a compounding monthly decline would imply over a few months.
- Ruling out measurement, accounting and currency explanations before looking for product causes.
- Breaking revenue down by business line and region, and each line into volume, frequency and price or fee drivers.
- Separating external causes (competitors, the economy, regulation, seasonality) from internal ones (pricing or fee changes, outages, product and policy changes).
- Ranking hypotheses and naming the data cut that would confirm or rule out each one.
### Follow-up Questions
- The drop turns out to sit almost entirely in third-party seller fees. What do you look at next?
- Orders per customer are flat, but average order value is falling. What could explain it, and how would you tell the explanations apart?
- Once the cause is found, which leading indicators would you add to a dashboard so the next decline is caught earlier?
Overview: A metrics-diagnosis prompt asking which measures a product manager would examine if Amazon's revenue kept falling 10% every month. It tests clarifying what the decline means, ruling out data and currency effects, breaking revenue into business lines and drivers, and ordering hypotheses sensibly.