Data Analyst interview
at Amazon.

Tell-me-about-a-time stories on ownership, conflict, and impact — scored against this company's real values.

1. 01 Tell me about a time your attention to detail in an analysis caught an error or insight that others had missed (Dive Deep).
2. 02 Describe a time you delivered an analysis that directly improved the customer experience, even though it was not what the business originally asked for (Customer Obsession).
3. 03 Give me an example of when you had to push back on a stakeholder who wanted a conclusion the data did not support, and how you handled it (Have Backbone; Disagree and Commit).
4. 04 Tell me about a time you automated or simplified a recurring report to save the team significant time (Invent and Simplify / Frugality).

The hard screen for the craft itself — talked through out loud, not whiteboarded in your head.

1. 01 Given Amazon order and shipment tables, write a SQL query using a CTE and window functions to compute the rolling 7-day average of late deliveries per region.
2. 02 How would you measure whether a change to the product detail page increased add-to-cart rate, and what confounders would you control for?
3. 03 Write a query to find customers whose order frequency dropped by more than 50 percent quarter-over-quarter, to flag potential churn.
4. 04 A weekly business metric suddenly spiked; walk me through the steps and SQL you would use to determine whether it is real or a data-quality artifact.

Open-ended design of the systems this company actually runs, with the interviewer probing your tradeoffs.

1. 01 Design a self-service dashboard and underlying data model that lets Amazon category managers track sales, returns, and inventory health.
2. 02 Design the reporting data pipeline that aggregates raw clickstream events into daily business metrics for the retail website.
3. 03 Design a data-quality monitoring system that alerts when a key Amazon KPI table arrives late or with anomalous values.
4. 04 Design the metrics layer for tracking Prime membership funnel conversion from sign-up to first order.