Amazon Business Analyst Interview Guide

About Amazon

Amazon is the world's largest e-commerce and cloud computing company, and its Business Analyst roles sit inside teams that turn operational and customer data into decisions — across retail, logistics, and AWS. BAs are expected to work independently in SQL and Excel from day one, and the role is a common entry point into Amazon's wider data and product organization. Levels typically run L5 to L7, with scope and which Leadership Principles get tested shifting at each level.

Interview Process Overview

4 to 6 stages over roughly 3 to 6 weeks: an online assessment, a recruiter screen, a technical phone screen, and an onsite loop (Amazon calls it a "Superday") of 4 to 6 back-to-back interviews. One of those onsite interviewers is a Bar Raiser — someone from outside your hiring team, trained specifically to protect Amazon's hiring bar, who holds independent authority over the final decision.

Round 1: Online Assessment

Format: 45–90 minutes, remote, timed.

What they test: SQL fluency, Excel modeling, and comfort with ambiguous, under-specified data.

Typical questions:

You're given table names, columns, and sample rows with no explicit question attached — write a query that surfaces month-end account balances from sender/receiver/amount data.
Join two datasets on a shared key and compute a conversion rate.
Use a window function to rank the top-performing category per region per month.
Estimate building occupancy from badge-swipe timestamps.
A query returns duplicate rows after a join — explain why, and fix it.

Round 2: Recruiter Screen

Format: 30 minutes, phone, with a recruiter (not the hiring team).

What they test: Motivation, role fit, and basic screening against the job description.

Typical questions:

Walk me through your resume, focused on your analytics work.
Why Amazon, and why this team specifically?
Describe a project where you used data to influence a decision.
What's your real, hands-on experience level with SQL and Excel?
What are you looking for in your next role?

Round 3: Technical Phone Screen

Format: 60 minutes, video, usually with the hiring manager or a senior BA.

What they test: SQL depth beyond syntax — query optimization at scale — plus one Leadership Principle woven into a behavioral question.

Typical questions:

Write a query using a CTE to find the second-highest value per group.
This query runs fine on 1TB of data but times out at 10TB — how do you optimize it?
Tell me about a time you found an error in someone else's analysis. (Are Right, A Lot)
How would you deduplicate records with no unique identifier?
What's the difference between a window function and GROUP BY — give an example only one of them can solve.

Final Round: Onsite Loop ("Superday")

Format: 4–6 back-to-back interviews, 45–60 minutes each, mixing technical and behavioral. One interviewer is a Bar Raiser from outside your hiring team.

What they test: Everything above, plus the Leadership Principles weighted for your level. At L5, Customer Obsession, Bias for Action, and Dive Deep come up most. From L6 up, Think Big, Are Right A Lot, and Insist on the Highest Standards get more weight.

Typical questions:

Tell me about a time you dove deep into data to disprove an assumption everyone believed. (Dive Deep)
Describe a decision you made with incomplete data, and how it turned out. (Bias for Action)
Tell me about a time you advocated for the customer against internal pushback. (Customer Obsession)
Walk me through a project end to end, including a mistake you made and what you changed afterward. (Ownership / Are Right, A Lot)
A live SQL exercise: given three tables, calculate week-over-week retention and narrate your query choices out loud.

What Interviewers Look For

Structured, specific answers. Amazon officially recommends the STAR format (Situation, Task, Action, Result), and interviewers take notes against it — a good story with no clear Result still gets marked down.
Data-backed reasoning over intuition, in both technical and behavioral answers — show how you knew, not just what you did.
Ownership beyond your job description. Interviewers specifically probe for candidates who fixed something nobody assigned them.
Comfort being challenged. The Bar Raiser in particular will push back on an answer to see whether you defend it with more data or fold. Folding too fast reads as failing "Are Right, A Lot"; getting defensive without evidence reads the same way. Bar Raisers can, in rare cases, veto a "hire" decision the rest of the panel agrees on — real, but genuinely rare (one Bar Raiser has described never using it across 700+ interviews).
One story, several Principles. Strong candidates prepare 6–8 stories that can each answer 2–3 different Leadership Principles, since the same project usually demonstrates several at once.

How to Stand Out

Prepare your STAR stories in advance and specifically rehearse the "Result" line — most candidates undersell or skip the outcome.
Narrate your SQL thinking out loud in the technical rounds. Amazon interviewers are scoring your reasoning process, not just the final query.
When the Bar Raiser pushes back, treat it as a test of conviction, not a signal you were wrong — defend your position with data, or genuinely update your view. Either is a pass; visibly panicking is not.

Common Mistakes to Avoid

Treating the online assessment as a formality — a large share of BA candidates are filtered out here before a human reviews their resume.
Reusing the same 2–3 stories in every round, unadapted to the Principle being tested — interviewers compare notes afterward, and repetition without depth is noticeable.
Optimizing a query for correctness but never mentioning scale. Amazon's data volumes mean "does it work" and "does it work at 10x the data" are different questions, and interviewers expect you to raise the second one yourself.

Sources used: aboutamazon.com (Bar Raiser program, official), igotanoffer.com (Bar Raiser veto power), tryexponent.com (round structure, SQL/Excel examples, Leadership Principle weighting by level).