Most candidates prepare for Amazon by studying the Leadership Principles. That's the right starting point, but it isn't what the loop tests. The loop tests whether your stories hold up when four or five different interviewers pull on them for an hour each, take detailed notes, and then compare what they heard in a debrief. This guide is about that day — how the rounds are structured, what each one feels like, how to build stories deep enough to survive probing, and where well-prepared candidates still lose the offer. If you haven't mapped your stories to the principles yet, start with our Leadership Principles guide first, then come back here.
The Amazon loop is a set of back-to-back interviews — usually four to six, each around 45–60 minutes — where every interviewer is assigned specific Leadership Principles and one of them is a Bar Raiser from outside the hiring team. Interviewers write up detailed notes and reconcile them afterwards, so shallow stories get caught. Plan for two to three stories per round and expect roughly two-thirds of your time on each story to be follow-up questions. Prepare six to eight layered stories — each with the numbers, the tradeoffs you rejected, the pushback you got, and what you'd do differently — rather than a dozen thin ones. Star Interview has Amazon-specific audio episodes and a Story Bank that turns your resume into structured STAR answers you can rehearse before the loop.
“The loop” is Amazon's name for the final interview stage — the set of interviews that happens after your recruiter screen and, for technical roles, an online assessment or phone screen. It runs as a single block, either virtually or onsite, and the people in it are deliberately not interchangeable.
Four to six interviewers, each with assigned principles
Before the loop, the recruiter distributes Leadership Principles across interviewers so the panel covers a spread rather than all asking about Ownership. You will be asked about different principles in each room, and you will not be told which ones in advance.
One of them is a Bar Raiser
A trained interviewer from outside the hiring team, brought in to keep the hiring bar consistent across the company. They have no stake in filling this specific role, which is exactly what makes the round feel different.
Everything is written down
Amazon interviewers take near-verbatim notes during the conversation and write up their assessment afterwards, tied to the principles they were assigned and the evidence you gave. Long pauses while someone types are normal, not a bad sign.
A debrief reconciles the panel
After the loop, the interviewers meet and compare write-ups. This is why consistency matters: if two interviewers heard two different versions of the same project, that discrepancy surfaces in the room you are not in.
The practical consequence: you are not preparing for one interview repeated five times. You're preparing a body of experience that has to stay coherent across five separate conversations and one written record.
The format is similar in every room — behavioral questions, deep follow-ups, a few minutes for your questions — but the emphasis shifts depending on who is across the table.
Scope, judgment, and fit for the team
This round tends to be the most forward-looking. Expect behavioral questions about delivering results and raising standards, plus probing on the scope you've owned: how big was the team, what was the budget or headcount, who did you have to convince. Have a clear sense of the level you're interviewing for and pick stories at that altitude — a tightly scoped individual project will read as junior if the role expects you to drive across teams.
Collaboration, disagreement, and day-to-day reality
These are the people who will work alongside you, and their questions skew toward Earn Trust, Have Backbone, and Insist on the Highest Standards. They're listening for how you behave when you disagree with someone whose cooperation you need. Stories where you won an argument by pulling rank land badly here; stories where you changed your own mind after seeing better data land well.
Consistency of the hiring bar, not the needs of this team
This is usually the round candidates describe as the hardest, and the reason is structural: the Bar Raiser isn't trying to fill the role, so they have no incentive to help you past a weak answer. They tend to probe one story much further than anyone else — four, five, six layers deep — until either the detail runs out or they're satisfied. Being asked “and then what?” five times isn't a sign it's going badly. Running out of substance on the third ask is.
Still partly behavioral, whatever the role
Coding, system design, case, or analytics rounds usually reserve the first 10–15 minutes for behavioral questions, because those interviewers are assigned principles too. Candidates routinely under-prepare for this: they arrive sharp on the technical material, get a Leadership Principle question in the first five minutes, and give a much weaker answer than they would have in a “behavioral” round. Treat every room as a behavioral round that happens to also cover something else.
A useful way to prepare is to budget the round rather than the answer. In a 45–60 minute behavioral round, here's roughly where the time goes:
~5 minutes
Brief, and not the place for a five-minute career narrative. Have a 60-second version of “what I do now” ready and then hand the room back to the interviewer.
~35–45 minutes, and most of it is follow-ups
Each story starts with your own telling — aim for three to four minutes — and then the interviewer spends roughly twice that long digging into it. That ratio is the whole point of this section: if you rehearse only the four-minute version, you've prepared a third of what the round asks for. If your initial telling runs past six or seven minutes, the interviewer has to cut you off to get their evidence, and you lose control of which parts they hear about.
~5 minutes, and they are scored
Interviewers reliably leave time at the end, and what you ask becomes part of the write-up. Prepare a different question for each interviewer, matched to their role — the hiring manager can speak to roadmap and team structure, a peer can speak to what the last six months actually felt like, the Bar Raiser can speak to how the bar is applied. Recycling one generic question through five rooms is a wasted signal.
Multiply that across the loop and the arithmetic is worth noticing: five rounds at two to three stories each means you may be asked to tell ten to fifteen stories in one day. Six to eight well-built stories can cover that, because several of them work for more than one principle — but three stories cannot.
A flex story is one you can angle at several Leadership Principles by changing which part you lead with. The loop adds a second requirement: the story has to have more depth than you plan to use. For each of your six to eight stories, write out these layers before the loop. You won't recite them — you're making sure the material exists so you never have to invent it live.
Layer 1 — the four-minute telling
Situation, task, your actions, result. Tight, chronological, and specific about what you personally did. This is the only layer most candidates prepare.
Layer 2 — the numbers behind the numbers
Not just the headline result, but the baseline before you started, how it was measured, how long it took, and what it cost. If you claim something improved, be ready to say what it was before, how you knew, and who else saw the data.
Layer 3 — the options you rejected
Every real decision had alternatives. Name two you considered and did not take, and why. This is where judgment becomes visible — and it is the most common place a story runs out of substance.
Layer 4 — the people friction
Who disagreed with you, what their argument was, how you handled it, and whether they ended up right. Amazon interviewers ask about disagreement in almost every round, and an answer of "everyone was aligned" reads as either a trivial project or a sanitized retelling.
Layer 5 — what you would do differently
A concrete change, not a platitude. Strong candidates can point to the specific decision they would reverse and explain what evidence would have made them see it sooner.
Layer 6 — the alternate angles
Note the two or three principles this story can serve and what you would emphasize for each. The same project can demonstrate Ownership if you lead with the mess you inherited, or Customer Obsession if you lead with the user problem you found.
One test before the loop: have someone ask you “why?” or “and then what?” five times in a row on a single story. If you reach the end of what you know on the third question, that story isn't loop-ready yet — either go back to your notes and reconstruct the detail, or pick a different project you remember better.
Amazon's follow-up questions are consistent enough that you can prepare for them directly. For each of your stories, make sure you have an answer to each of these:
The single most common probe. If you narrate in "we," you will get this question, and your answer will sound like a correction rather than a fact. Say "I" for what you did and name others explicitly when crediting them.
Be ready with the baseline, the measurement method, and the timeframe. "It improved a lot" invites three more questions; "we were at a four-day turnaround and got to same-day over two months, measured off the ticketing system" closes the loop.
They want the actual exchange: what the objection was, what you said, and what changed. Note whether you persuaded them, escalated, or changed your own position — all three can be good answers, depending on what the evidence supported.
Name what you gave up, not just what you chose. A decision with no cost was not a hard decision, and saying it was will read as either low-stakes work or low self-awareness.
Have one specific answer ready per story. "Nothing" is a red flag in every round and a fatal one with a Bar Raiser.
Amazon cares about durable results. If you left before the outcome was clear, say so plainly and share what the last data you saw showed. Guessing at results you never measured is the fastest way to lose credibility.
This tests whether you can reason under uncertainty without rewriting history. Answer from the information you actually had, then separately note what you learned later.
Notice that none of these require a new story. They require the story you already chose to be known in more detail than you first tell it — which is why depth beats breadth when you're preparing for a loop.
Several hours of back-to-back interviews is a stamina problem as much as a preparation problem. A few things that make the difference:
Track which stories you have used
Keep a short list of your stories in front of you and mark them as you go. By round four, most candidates genuinely cannot remember what they told in round one — and repeating the same story without flagging it is what creates the odd duplicates the debrief notices.
It is fine to reuse a story — say so
With five rounds and one career, overlap is expected. Just name it: "I touched on this project with someone earlier, but let me take it from a different angle." That reads as organized rather than repetitive, and it tells the interviewer to probe elsewhere.
Ask for a moment before you answer
Taking five or ten seconds to pick the right story is normal and reads as considered. Launching into the wrong story and course-correcting halfway through costs far more time.
Keep a one-page cheat sheet, not a script
In a virtual loop, one page with story titles and key numbers is a useful anchor. A full script is worse than nothing — reading from it flattens your delivery and interviewers notice immediately.
Plan food, water, and a real break
Confirm the schedule with your recruiter in advance and ask where the gaps are. Energy in the last round is scored the same as energy in the first, and you will not know in advance which round is the one that matters most to you.
Ask your recruiter what to expect
Recruiters will usually tell you the number of rounds, the rough format, whether a written exercise or presentation is involved, and sometimes which principles the panel leans on. It is a normal question and the answer materially changes how you prepare.
These aren't the beginner mistakes. They're the ones that show up in candidates who studied the Leadership Principles carefully and still got a no.
Preparing breadth instead of depth
Sixteen thin stories, one per principle, is the classic trap. Each one collapses on the second follow-up. Six to eight stories you know at the level of individual decisions, with the alternate angles mapped out, covers the same ground and survives probing.
Treating the Bar Raiser round as hostile
Sustained probing is the job, not a verdict. Candidates who read it as hostility get defensive, start hedging, or shorten their answers — all of which reduce the evidence in the write-up. Stay concrete and keep offering detail.
Inconsistent versions of the same project
If you round a number up in one room and down in another, or your role grows across the day, the debrief surfaces it. Fix your numbers and your scope in writing beforehand and tell it the same way every time.
Front-loading the situation
Three minutes of org chart and background leaves no room for your actions, and your actions are the evidence. Two or three sentences of context, then get to what you did.
Results with no measurement
Unquantified outcomes are the most common gap in otherwise strong answers. If you genuinely do not have a number, use a concrete proxy — a deadline met, a process that replaced another, an escalation that stopped — and say plainly what you could and could not measure.
Saving your best story for last
There is no ranking across rounds — each interviewer only writes up what they personally heard. Lead with your strongest relevant story in every room.
Ignoring the written or presentation component
Some roles include a written exercise, case, or presentation as part of the loop. Ask about it early. It is graded, and it is the piece candidates most often discover too late to prepare for.
Answering the principle instead of the question
Candidates who have over-studied sometimes narrate the framework — "this shows Customer Obsession because..." — instead of telling the story and letting the evidence speak. Interviewers map your answer to their principles themselves; your job is to give them specifics to map.
If you have three or four weeks between scheduling and the loop, this order tends to work better than studying the principles on repeat:
Start from your own history, not from the principles. List every project, escalation, migration, launch, and difficult conversation from the last three to five years. Pull the numbers out of old documents and dashboards while you still can — this is the step people skip and then regret in the room.
Choose the stories with the most substance and write out all six layers for each. Aim for a spread of story types — something you led, something you fixed, something you failed at, something you disagreed about, something you shipped under pressure — rather than one per principle.
Only now bring in the Leadership Principles. Mark which principles each story can serve. You'll usually find two or three with no coverage at all — that's the point of the exercise. Go back to your inventory for those rather than stretching an existing story to fit.
Say your stories out loud and time them. Then have someone interrupt with follow-ups, or record yourself answering the probes from the list above. Silent rereading builds familiarity, not fluency, and the loop tests fluency under pressure.
The two hard parts of loop prep are building stories with enough depth and getting enough out-loud practice that they sound natural after four hours of talking. Star Interview is built around both.
Story Bank builds your layered stories
Upload your resume, pick the company, and Story Bank turns your experience into structured STAR answers you can edit until they sound like you. You can keep several story banks per company — useful when you are interviewing for more than one level or team.
Amazon-specific audio episodes
Podcast-style episodes walk through Amazon’s Leadership Principles and behavioral rounds with worked examples, so you can hear how a strong answer is paced and where the follow-ups usually land.
Two-host conversational format
Hearing two people discuss and pull apart an answer makes the probing pattern obvious — which is hard to pick up from a written list of questions.
Coverage across 30+ companies
If Amazon is one of several loops you are running, company-specific episodes help you keep each interview culture straight instead of blending them together.
Prep on the commute and at the gym
Loop prep competes with a full-time job. Audio turns the dead time you already have into review time.
Speed controls and resume support
Slow down for new material, speed up to 2x for review, and pick up exactly where you left off in the days before the loop.
Build your Story Bank from your own experience, then rehearse with Amazon-specific audio episodes — so the fifth round sounds as sharp as the first.
Get Started