Amazon Leadership Principles Interview Guide (2026)

Company Guide · Updated April 2026 · Reviewed by a former Amazon Bar Raiser (8 years active)

Amazon doesn't have behavioral questions. It has Leadership Principles — sixteen of them — and every story you tell in a loop gets scored against one (sometimes two) of them. Interviewers are trained to downgrade answers that don't map cleanly to a principle, even when the story is technically impressive.

This is the rubric that makes Amazon behavioral interviews harder than Google's, Meta's, or Microsoft's. It's also the reason this guide exists. Most English-language articles cite the 14-principle list from 2018, or a stale version of the 16-principle list from 2021. For a 2026 loop you need the current list and the current scoring shape — including the Bar Raiser pattern and the two newest principles added in the 2021 rewrite.

Pair this article with our full behavioral interview guide for STAR mechanics, then use the rubric below to structure your Amazon-specific story index.

The 16 Leadership Principles, as of 2026

Amazon added two principles in 2021 ("Strive to be Earth's Best Employer" and "Success and Scale Bring Broad Responsibility") and tightened the language on four others in a 2023 revision. None have been removed. The list is:

1. Customer Obsession
2. Ownership
3. Invent and Simplify
4. Are Right, A Lot
5. Learn and Be Curious
6. Hire and Develop the Best
7. Insist on the Highest Standards
8. Think Big
9. Bias for Action
10. Frugality
11. Earn Trust
12. Dive Deep
13. Have Backbone; Disagree and Commit
14. Deliver Results
15. Strive to be Earth's Best Employer
16. Success and Scale Bring Broad Responsibility

What interviewers actually do with this list. Every question is prefixed in the interviewer's script with "Tell me about a time you…" followed by a stem that maps to one (or occasionally two) principles. The interviewer then listens for concrete evidence that you lived the principle — not a restatement of it.

Generic stories lose points. Scored answers name a customer, name a decision, and name a measurable result.

The Bar Raiser: what they look for

Every Amazon onsite loop includes one Bar Raiser — an independent senior interviewer, trained at an internal program, who is not on the hiring team. Their job is to protect the long-term hiring bar regardless of how urgently the team needs the headcount.

Three things to know:

• They have veto power. A "not inclined" vote from the Bar Raiser almost always blocks the offer, even when the hiring manager pushes to override.
• They go long on one story. Bar Raisers typically pick one story you told and drill for 45–60 minutes — follow-ups about tradeoffs, people who disagreed, what you would do differently now. Every story in your index needs at least three layers of depth.
• They debrief separately. After the loop, the Bar Raiser writes their own note and presents in a room that includes the hiring manager but is led by the recruiter. The tone is "protect the bar," not "close the req."

If you under-prepare for exactly one thing in the loop, don't let it be the Bar Raiser.

All 16 principles, with one scored STAR example each

Each principle below has a sample stem and a 60–80 word scored STAR answer at the mid-SDE (L5) expectation level. Read them as shape, not script — replace with your own stories.

1. Customer Obsession

Stem: "Tell me about a time you went beyond what your team asked of you to fix something for a customer."

Scored answer: "I ran analytics for an e-commerce client shipping to EU warehouses. Their ops lead flagged that returns were spiking but couldn't tell me why. Our contract didn't cover analysis, but I built a five-query dashboard in two days, isolated the issue to one mislabelled SKU family, and pushed the fix upstream to our catalog team. Returns dropped 34% the following month. The client renewed at 2× the original contract value."

2. Ownership

Stem: "Describe a time you took responsibility for a problem that wasn't yours to fix."

Scored answer: "An on-call alert for checkout latency fired during a holiday sale. It was our payments team's oncall, but they were paged five times and couldn't triage. I paused my feature work, paired with their engineer, isolated a slow join, and pushed a hotfix behind a feature flag within 90 minutes. The sale hit target. Afterwards I wrote a runbook for the join pattern that prevented the same incident in the next two holidays."

3. Invent and Simplify

Stem: "Tell me about a time you simplified a complex process."

Scored answer: "Our deploy pipeline required three separate merges and two manual approvals. Engineers skipped deploys on Fridays. I proposed a single-button canary rollout with automated rollback, shipped it as an RFC, built the MVP over three weeks, and got adoption on four services before going wider. Median deploy time dropped from 47 minutes to 4, and Friday deploy frequency tripled."

4. Are Right, A Lot

Stem: "Tell me about a decision you made that turned out to be wrong. What did you learn?"

Scored answer: "I argued for a custom event bus over SNS for a new service because our team had burned time on SNS limits before. Six months later, ops cost was 3× SNS and our event-bus maintenance was eating a full engineer. I wrote a retro, migrated us back to SNS over a quarter, and now run a 'pick the default' check before advocating for custom infrastructure. The migration paid back in five months."

5. Learn and Be Curious

Stem: "Tell me about a time you had to learn something unfamiliar quickly."

Scored answer: "I was asked to own the observability migration to OpenTelemetry with no prior experience. I read the spec, built a 20-service test cluster in a week, then wrote an internal tutorial that became the team's adoption guide. Migration shipped in 10 weeks, a month ahead of schedule, and the tutorial was cited in two other team adoptions."

6. Hire and Develop the Best

Stem: "Tell me about a time you developed a teammate."

Scored answer: "A new hire on my team struggled with scoping. I pulled a recurring design-doc template from three senior engineers, drilled the new hire through five scope-definition exercises in their first month, and had them co-own the next quarterly roadmap. Their first solo launch shipped on time; they were promoted to L5 within 14 months — six months faster than the team average."

7. Insist on the Highest Standards

Stem: "Describe a time you pushed back on quality you felt wasn't good enough."

Scored answer: "A vendor delivered a UI component at 60 fps on their demo but dropped to 20 on our customer devices. PM wanted to ship anyway. I ran a measurement harness across 120 real customer devices, shared the distribution, and refused sign-off. We held the release two weeks, the vendor shipped a fixed build, and we hit 58 fps median. The customer complaint rate stayed at baseline instead of the 6% spike we'd have seen otherwise."

8. Think Big

Stem: "Tell me about a time you proposed something much larger than your remit."

Scored answer: "Our service owned checkout, but I saw that the post-checkout email pipeline had 12 teams building near-duplicate templates. I wrote a memo proposing a shared email service, scoped it at two quarters, got buy-in from three directors, and drove the first launch. Three years later it handles 400M+ emails/month across every Amazon business unit I know of."

9. Bias for Action

Stem: "Tell me about a time you made a decision without complete information."

Scored answer: "A sev-2 dashboard showed customer conversion dropping 8% during a marketing launch. Our analytics pipeline wouldn't confirm the cause for 24 hours. I pulled a sample of the last 1,000 sessions, grepped for the new promo code, and found it was being applied twice. I pushed a guard in 40 minutes, caught the double-apply on 11k orders overnight, and we issued corrections before any customer saw a wrong charge."

10. Frugality

Stem: "Tell me about a time you did more with less."

Scored answer: "Our ML inference bill was $220k/month. Instead of requesting more budget, I profiled the workloads, found 62% were cold-starting unused models, and shipped a tiering system that spun models up on demand. Bill dropped to $78k/month with no latency regression. The cost savings funded two additional headcount for the team."

11. Earn Trust

Stem: "Describe a time you changed your mind based on someone else's data or argument."

Scored answer: "I was set on migrating our data store to DynamoDB. A principal engineer walked me through three production outages they'd debugged in DynamoDB that I hadn't considered. I redesigned the proposal to keep PostgreSQL for transactional writes with a DynamoDB read replica, wrote it up publicly, and delivered the split system on schedule. The principal engineer later cited the revised design as the cleanest migration they'd reviewed that year."

12. Dive Deep

Stem: "Tell me about a time you went deeper on a problem than anyone expected."

Scored answer: "Customers reported intermittent payment failures. PM wanted to close as 'can't reproduce.' I pulled 200 failure traces, correlated them against the upstream retry pattern, and found a 1-in-2400 race condition in the token refresh. I wrote the fix, a regression test, and a 300-word memo diagramming the race. Failure rate dropped from 0.04% to 0.001%. The memo became the onboarding read for the payments team."

13. Have Backbone; Disagree and Commit

Stem: "Describe a time you disagreed with your manager but committed to the decision anyway."

Scored answer: "My manager wanted to launch a referral feature in two weeks. I had data showing the attribution logic would double-count shared links. I escalated with a 400-word memo, flagged the specific failure mode, and proposed a three-week plan. Manager held the date. I committed, shipped on the original schedule, monitored closely, and caught the double-counting on day 3 with the guardrails I'd pushed for. We rolled back cleanly, fixed, and re-launched a week later. Trust with my manager went up, not down."

14. Deliver Results

Stem: "Tell me about a time you delivered a project that had slipped."

Scored answer: "A cross-team migration was eight weeks behind at handoff to me. I cut scope to the three highest-value services (from nine), renegotiated SLAs with three partner teams, and shipped daily demo builds. We delivered the reduced scope on the original date; the remaining six migrated over the next two quarters without a single post-migration regression."

15. Strive to be Earth's Best Employer

Stem: "Tell me about a time you improved the day-to-day experience of your team."

Scored answer: "Our oncall rotation was hitting engineers with 20+ pages a week. I built an alert-quality scorecard, merged six duplicate alerts, and routed three noisy ones to async tickets. Pages/week dropped to 6. Two engineers who had asked to leave oncall stayed. Retention across the team that year was 100%, up from 78% the year prior."

16. Success and Scale Bring Broad Responsibility

Stem: "Tell me about a decision where you considered the broader impact of a choice, not just your team's outcome."

Scored answer: "We had a chance to reduce checkout latency by 40ms using a third-party data broker. I flagged the data-handling posture — the broker resold customer purchase signals. We had the contract in hand. I pushed for an internal-only alternative that took two quarters longer. We absorbed the delay. A year later the third-party was cited in a regulatory review we would have been dragged into."

The 32-story index

You won't be asked 16 questions in your loop. You'll probably be asked 12–14, drawn from the full list, across 4–6 interviewers. Some questions will touch two principles; some interviewers will ask about two different principles in the same round. If you have one story per principle and the interviewer asks back-to-back Ownership and Bias for Action questions in the same loop, you'll reuse a story — and every Bar Raiser is trained to clock that.

The fix is simple. Prepare two stories per principle — 32 stories total — tagged in a spreadsheet with:

• Primary principle
• Secondary principle (if dual-use)
• Measured outcome (number or percentage, never a verb)
• Scope (individual, small team, cross-team, org)
• Year

Before the loop, drill each story to four-minute spoken length, then cut to three minutes. The longest answer a Bar Raiser will listen to before cutting you off is around four minutes. Practice with a stopwatch.

Common rejection patterns

Five things the Bar Raiser downgrades on, in rough order of frequency:

1. "We" slips. If you cannot narrate what you personally did in 80% of your Action step, the story doesn't score.
2. Missing Result metrics. "We launched the feature and it went well" reads as a non-answer. Name the number.
3. Under-diving on Dive Deep. When asked for a system-level detail, "I trusted the team" is scored as avoidance.
4. Conflict stories without real disagreement. If the other party in your Backbone story never actually pushed back, the answer is a non-example.
5. Stale stories. Anything older than three years reads as "no recent scope." Keep at least two stories from the last 12 months.

A 7-day Amazon prep schedule

• Day 1. Read Amazon's public Leadership Principles page end-to-end. Mark the three you've had the least exposure to in your career. Those need the strongest stories.
• Day 2. Write 32 story skeletons in a spreadsheet — one-sentence Situation, one-sentence Action, one-sentence Result, the principle(s), the number.
• Day 3. Expand your six weakest skeletons into full four-minute answers, out loud, with a timer.
• Day 4. Expand the remaining 26. Record yourself on at least eight of them.
• Day 5. Run a mock loop — five questions, no breaks, with a friend or an AI coach. InterviewPilot's Amazon-preset practice loop runs this end-to-end with scoring against each principle.
• Day 6. Drill Bar Raiser follow-ups on your five strongest stories. What were the tradeoffs? Who disagreed? What would you do differently?
• Day 7. Light review only. Sleep.

Frequently Asked Questions

How many Amazon Leadership Principles are there in 2026?

Sixteen. The two most recent additions are "Strive to be Earth's Best Employer" and "Success and Scale Bring Broad Responsibility", both added in 2021 and still active on Amazon's 2026 interview rubric. Every behavioral question maps to one of the sixteen.

Do I need one story per Leadership Principle?

Aim for two per principle — so 32 stories indexed — because interviewers often ask back-to-back behavioral questions in the same round and you can't reuse a story. The index lets you retrieve the right one under pressure.

What does a Bar Raiser do at Amazon?

The Bar Raiser is an independent senior hire-trained interviewer assigned to your loop who is not on the hiring team. They have veto power on the offer. Their job is to protect the long-term hiring bar regardless of how urgently the team needs the headcount.

How long should an Amazon behavioral answer be?

Two to three minutes spoken. Under 90 seconds reads as shallow; over four minutes and the interviewer runs out of time for the follow-ups that score the most rubric points.

Can I reuse the same STAR story for two different LPs?

Reluctantly. If you must, explicitly tell the interviewer: "I already used the warehouse-rollout story for Ownership; for Bias for Action, here's a different one." Reusing without calling it out is scored as a lack of depth.

Which LPs are most commonly asked for SDE interviews?

Customer Obsession, Ownership, Bias for Action, Dive Deep, and Earn Trust come up in almost every SDE loop. Invent and Simplify and Insist on the Highest Standards are the next tier. Prepare those seven first if time is tight.

What happens if I fail the Bar Raiser round?

The Bar Raiser's "not inclined" vote in the debrief typically blocks the offer even if the hiring manager pushes for it. It is not appealed. Your best response is to request feedback and re-apply after a strengthened story set in 6–12 months.

Keep reading

• The Behavioral Interview Guide: STAR, Stories, and How to Actually Win — the pillar guide
• 10 Common Interview Mistakes That Cost You the Job — what to avoid in any loop
• The Complete Guide to AI-Powered Interview Preparation — how to practice with AI scoring

Ready to run a scored practice loop against Amazon's rubric? Start a free trial — company preset and Bar Raiser-style follow-ups included.