How Amazon's Leadership Principles Are Actually Scored

Every candidate preparing for an Amazon interview knows about the Leadership Principles. Most can list them. A lot can recite rehearsed stories for each one. But very few understand how an interviewer actually evaluates what they hear, and that gap is where most rejections come from.

I spent 5 years recruiting at Amazon. I sat in hundreds of debrief sessions listening to interviewers explain why a candidate was a hire or a no-hire. After a while, patterns become very clear. Here is what is actually happening when an interviewer scores your behavioral answer.

Interviewers are listening for evidence, not stories

A common mistake is treating LP questions like a storytelling exercise. Candidates focus on building a narrative that sounds compelling. Interviewers are doing something different. They are listening for specific evidence that a behavior actually happened, at the right level of impact, with the kind of judgment the role requires.

The difference matters. A compelling story about a difficult project tells an interviewer something. Specific, verifiable evidence that you drove a decision, owned an outcome, or resolved a conflict in a particular way tells them something much more useful. They are not judging how well you tell stories. They are judging whether your examples are real and whether the behavior you demonstrated matches what Amazon needs at your target level.

Each LP question is tied to a specific principle, and the scoring is calibrated to level

When an interviewer asks a question, they have already decided which LP they are probing. They have follow-up questions ready. They know what a strong answer looks like at L4 versus L5 versus L6, because the expected scope and depth of ownership changes significantly by level.

At L4, strong answers usually involve your own direct work and clear ownership of a task or problem. At L5, interviewers want to see you influencing across a team or dealing with ambiguity. At L6 and above, the expectation is that your examples reflect decisions that shaped a system, an org, or a strategy. The same story can be a strong hire signal at one level and a signal that someone is not ready for the next.

Follow-up questions are not small talk

One of the most common things I saw in debriefs was interviewers noting that a candidate "fell apart" on follow-ups. The candidate gave a polished initial answer and then could not hold up under a few basic probing questions. That is a red flag, because it suggests the story was rehearsed rather than real.

Interviewers are trained to dig. "What would you have done differently?" "What was the data that drove that decision?" "Who else was involved and how did you influence them?" If your story does not hold up to those questions, the initial answer does not matter much.

Weak signals interviewers actually flag

In debriefs, these phrases came up over and over as reasons a behavioral round scored low:

• "We" instead of "I" throughout the answer, making it impossible to identify what the candidate personally did
• Outcomes that were vague or unmeasured, like "it went well" or "the team was happy"
• No evidence of pushback, difficulty, or genuine challenge, suggesting the story was cherry-picked to avoid complexity
• LP answers that described what the candidate should have done, rather than what they actually did
• Stories that were relevant but too junior for the level being interviewed for

What a strong answer actually sounds like

A strong LP answer at any level has a few things in common. The situation is specific enough that it could be verified. The candidate's individual contribution is clear. The outcome includes something measurable or concrete. And the candidate can speak to why they made the choices they did, including what they considered and what they ruled out.

The best answers are not the most dramatic ones. They are the ones where an interviewer leaves the room feeling like they understand exactly how this person thinks and operates, based on something real that actually happened.

That is what you are trying to demonstrate. Not a great story. A real signal.