Rejected from Amazon after the onsite loop, even with good feedback from most interviewers? It was probably the Bar Raiser. Here's what they actually score, and why strong coding doesn't save you.
By Adarsh Sharma · 20 June 2026 · 4 min read
After you map the weak Leadership Principle, run an interview debrief on that round so the next loop targets the real gap.
You finished the Amazon loop. A few of the interviewers seemed to like you. Then the rejection landed anyway, with no breakdown of what happened. If most of your rounds felt fine and you still got cut, there's a good chance the same person sank you that sinks a lot of strong candidates: the Bar Raiser.
What the Bar Raiser actually is
Every Amazon loop includes one interviewer who doesn't work on the team you're interviewing for. That's the Bar Raiser, and on Blind the description that comes up again and again is blunt: their job is to reject you. They're trained to keep Amazon's hiring bar from dropping, and the explicit mindset is that it's better to reject a good candidate than to risk hiring a bad one.
They also carry outsized weight. Along with the hiring manager, the Bar Raiser is one of the two most powerful voices in the debrief. Positive feedback from the other three interviewers does not save you if the Bar Raiser isn't convinced. That's why "but most of them liked me" rejections happen so often here.
Leadership Principles outweigh your code
This is the part candidates from a pure-LeetCode background get wrong. At Amazon, the behavioral evaluation against the Leadership Principles can matter more than your coding rounds. A pattern engineers report repeatedly: you can get in with mediocre coding and a genuinely strong LP showing, but the reverse — great coding, weak LP — almost never works, especially at senior levels.
So if you walked out thinking "my coding was solid," that may be true and completely beside the point. The loop wasn't only testing whether you can code. It was testing whether your past behavior is evidence of how Amazon wants people to operate.
The two ways LP answers get you rejected
From candidate debriefs, failed Leadership Principle answers tend to fail in one of two ways:
- The story didn't clear the bar. The situation was too small. No real scope, no hard tradeoff, no measurable outcome. "I fixed a bug and the team was happy" is not evidence of Ownership or Bias for Action. The Bar Raiser is looking for genuine scale and difficulty.
- The story felt too good to be true. This one surprises people. Over-rehearsed, suspiciously clean stories where you're the flawless hero read as fabricated. Real ownership includes the messy parts, the thing that went wrong, what you'd do differently. Polished-to-perfection makes a trained interviewer skeptical.
What to fix before you try again (Amazon lets you reapply)
- Pick bigger stories. For each core LP — Ownership, Bias for Action, Dive Deep, Deliver Results — choose a situation with real stakes, a real conflict or failure, and a number at the end.
- Say "I," not "we." The Bar Raiser is scoring your behavior. "We decided" tells them nothing about you. Use "we" only for context, "I" for every action you took.
- End every story with impact. A percentage, a latency number, revenue, customers, time saved. No number means no evidence.
- Keep one real failure ready. A believable story where something broke and you owned it beats a flawless one every time.
How to find out which Leadership Principle actually sank you
Ask the recruiter on a call — they won't email it, but many will tell you which competency came back weak. If you can't get that, you have to reconstruct it from what you remember: which LP questions you got, which stories you told, where the follow-ups dug in and you ran out of depth.
That reconstruction is exactly what PortLume's Interview Debrief does. You tell it the company, the stage, and the questions you recall, and it gives you a read on the most likely Leadership Principle you missed, what the Bar Raiser was really testing, and which stories to rebuild before you reapply. It's free to try, no account needed.
The reapply is real — use it
Amazon recycles candidates after a cooldown. The people who get in on the second attempt usually aren't better coders the second time. They figured out which LP they were weak on, rebuilt two or three stories with real scope and real numbers, and walked back in with evidence instead of hope.
Next steps
Put this into practice with one of these free tools and guides.