Why 12 stories are enough
A typical onsite loop runs 4–5 rounds, and each interviewer asks 2–3 behavioral questions with follow-ups — 15–20 prompts in a single day, which sounds like a lot until you notice how few distinct situations those prompts are fishing for. “Tell me about a conflict,” “tell me about a failure,” “tell me about leading without authority” — most companies rotate through the same dozen underlying scenarios, phrased differently and mapped to different names.
Twelve stories, each covering 2–3 themes, gives enough combinations to answer almost any prompt without repeating a story inside one loop — reusing a story across interviewers the same day signals a thin bank. Fewer than 8–10 and a prompt catches you flat. More than 12 and the marginal story stops paying for its rehearsal time; polish the twelve rather than draft a thirteenth nobody asks for. Build each one in STAR + Reflection format so structure carries the story, not phrasing.
The 12 story slots
These are the twelve situations that cover almost every behavioral prompt at any company running a structured loop. For each slot: what it has to prove, which frameworks it maps to, the questions it answers, and what to write down before you forget the details.
The conflict with a colleague
What it proves: you can disagree with a peer over substance without damaging the relationship. Amazon reads this against Earn Trust and Have Backbone; Disagree and Commit. Google scores it under Googleyness — collaborating with people who see problems differently. Meta folds it into Be Open.
Questions it answers: “Tell me about a conflict with a coworker.” “Describe a time you disagreed with a peer’s technical approach.” “Tell me about working with someone difficult.”
Fill it in: Name the disagreement in one sentence — not “we didn’t get along,” the specific technical or process split. Write what you said, what they said back, and how it resolved: did one of you change your mind, or land on a third option neither started with. Skip stories where you were simply right and they were wrong; pick one where the other person had a real point.
The failure you owned
What it proves: you can sit with a mistake, name your part in it, and describe what changed in how you work. Amazon reads this against Are Right, A Lot — the “changed course” half of the principle, not the “right” half. Google scores it as Googleyness (intellectual humility). Meta folds it into Focus on Impact.
Questions it answers: “Tell me about your biggest failure.” “Describe a project you owned that didn’t go as planned.” “Tell me about a mistake and what you learned.”
Fill it in: Write the failure in one line without softening it — a missed launch, a bad estimate, a call that cost money or trust. Name the number that measures the damage. Then write the one thing you do differently now, a habit you’ve actually adopted, not a lesson you’d like to claim.
Leading without authority
What it proves: you can get people who don’t report to you to move, through influence rather than a title. Amazon maps this to Think Big, Earn Trust, and Ownership. Google calls this the “emergent leadership” half of its Leadership attribute. Meta scores it under Be Bold.
Questions it answers: “Tell me about a time you led a project without formal authority.” “Describe how you got buy-in from a team you don’t manage.” “Tell me about influencing a decision above your level.”
Fill it in: Name the people you needed to move and why they weren’t already aligned. Write the specific tactic — the doc you wrote, the demo you ran, the one conversation that turned a skeptic. Note the moment the room actually tipped, and what tipped it.
The ambiguous problem you scoped
What it proves: you can take a vague mandate — “make onboarding better” — and turn it into a scoped, sequenced plan. Amazon maps this to Dive Deep, Invent and Simplify, and Ownership. Google weighs this under General Cognitive Ability, the attribute it scores most heavily, plus Role-Related Knowledge. Meta scores it as Focus on Impact.
Questions it answers: “Tell me about a problem with no clear direction.” “Describe how you scoped an ambiguous project.” “Tell me about turning a vague goal into a plan.”
Fill it in: Write the vague ask down verbatim. Then write the two or three questions you asked to narrow it, and the scope you cut to make it shippable. Note what you left out on purpose and why nobody missed it.
The hard deadline
What it proves: you can deliver under a real time constraint without cutting a corner nobody would notice until it broke later. Amazon maps this to Deliver Results and Bias for Action. Google reads it under General Cognitive Ability and Role-Related Knowledge. Meta scores it directly against Move Fast.
Questions it answers: “Tell me about delivering under a tight deadline.” “Describe a project where the timeline got cut.” “Tell me about the highest-pressure deadline you’ve faced.”
Fill it in: Name the original timeline and what changed it. Write the two or three things you personally did to compress the work — not “the team worked overtime,” but the actual call you made. Give the ship date and what happened once it was live.
Disagree and commit
What it proves: you can lose an argument, then execute the decision as if it were your idea. Amazon has a principle named for exactly this: Have Backbone; Disagree and Commit, alongside Are Right, A Lot. Google scores it under Leadership and Googleyness. Meta scores it as Be Open.
Questions it answers: “Tell me about disagreeing with a decision but executing it anyway.” “Describe a time your manager overruled you.” “Tell me about committing to something you didn’t fully agree with.”
Fill it in: Pick a case where the decision was made above you and stuck — not the conflict story from slot one. Write your original position, the reasoning that did or didn’t change the decision-maker’s mind, and one action showing you executed in good faith rather than slow-walking it.
Mentoring or growing someone
What it proves: you invest in other people’s growth, not just your own output. Amazon maps this to Hire and Develop the Best and Strive to Be Earth’s Best Employer. Google scores it under Leadership. Meta has no named value for this; interviewers fold it into the general leadership signal.
Questions it answers: “Tell me about someone you mentored.” “Describe how you helped a struggling teammate improve.” “Tell me about developing a junior engineer or new hire.”
Fill it in: Name the specific gap the person had — not “they needed to grow,” the actual skill or habit. Write the concrete action: the pairing sessions, the stretch assignment, the feedback conversation that landed. Give the after: a promotion, a metric, a project they later owned solo.
Cutting scope under pressure
What it proves: you can protect a deadline by cutting the right things instead of grinding nights and weekends. Amazon maps this to Frugality and Deliver Results. Google reads it under Role-Related Knowledge and General Cognitive Ability. Meta scores it against Move Fast and Focus on Impact.
Questions it answers: “Tell me about cutting scope to hit a deadline.” “Describe negotiating a smaller version of a project.” “Tell me about saying no to a stakeholder.”
Fill it in: List what was in scope and what you cut. Write who you had to convince and the argument that worked. Note what happened to the cut items — shipped later, or never missed.
The production incident
What it proves: you stay level-headed during a live outage, find the root cause, and prevent the repeat. Amazon maps this to Ownership, Dive Deep, Insist on the Highest Standards, and Bias for Action — often the densest story in a bank. Google reads it under General Cognitive Ability and Role-Related Knowledge. Meta scores it against Move Fast.
Questions it answers: “Tell me about a production incident you handled.” “Describe debugging a critical issue under time pressure.” “Tell me about an outage and what changed afterward.”
Fill it in: Name the blast radius — users affected, revenue, duration. Write your role during the incident: who paged whom, what you personally diagnosed versus what others found. Then the follow-up: the runbook, the alert, the postmortem action item you drove to closed, not just filed.
Cross-team collaboration
What it proves: you can get a result that required another team’s cooperation, especially when incentives didn’t fully line up. Amazon maps this to Earn Trust, Customer Obsession (treating the other team as an internal customer), and Think Big. Google reads it under Googleyness and Leadership. Meta scores it as Be Open.
Questions it answers: “Tell me about working with a team outside your own.” “Describe a project that required cross-functional coordination.” “Tell me about a time your priorities conflicted with another team’s.”
Fill it in: Name both teams and where the incentives diverged. Write the mechanism you used to align them — a shared doc, a joint metric, a standing meeting you started. Give the shipped result and who ended up benefiting.
The unpopular decision
What it proves: you can make and hold a call the room didn’t want, when the data or the standard demanded it. Amazon maps this to Have Backbone; Disagree and Commit and Insist on the Highest Standards. Google reads it under Leadership and Googleyness. Meta scores it as Be Bold.
Questions it answers: “Tell me about an unpopular decision you made.” “Describe delivering bad news.” “Tell me about holding a line when others wanted to move on.”
Fill it in: Name the decision and who opposed it, specifically. Write the reasoning behind the call and how you delivered it to the people who disagreed. Note the fallout — did the opposition come around, and how long did that take.
Going deep on data
What it proves: you don’t accept the dashboard summary; you pull the raw numbers and find what the summary hid. Amazon maps this to Dive Deep and Are Right, A Lot. Google reads it under General Cognitive Ability and Role-Related Knowledge. Meta scores it as Focus on Impact.
Questions it answers: “Tell me about investigating a metric that looked wrong.” “Describe digging into data to find a root cause.” “Tell me about a decision you changed after looking closer at the numbers.”
Fill it in: Name the number that looked off and what the dashboard said versus what you found underneath it. Write the specific tool or query you used. Give the decision that changed once you had the real picture, and what it was worth.
Mapping stories to companies
The same twelve stories work everywhere; only the framing changes. Amazon wants the specific Leadership Principle named. Google interviewers each silently score one of four attributes: General Cognitive Ability, Role-Related Knowledge, Leadership, Googleyness. Meta reads answers against five core values: Move Fast, Focus on Impact, Be Bold, Be Open, Build Social Value. The compact version, for a fast pre-interview scan:
| Story slot | Amazon | Meta | |
|---|---|---|---|
| Conflict with a colleague | Earn Trust; Have Backbone, Disagree and Commit | Googleyness | Be Open |
| Failure you owned | Are Right, A Lot | Googleyness | Focus on Impact |
| Leading without authority | Think Big; Earn Trust; Ownership | Leadership | Be Bold |
| Ambiguous problem scoped | Dive Deep; Invent and Simplify | General Cognitive Ability | Focus on Impact |
| Hard deadline | Deliver Results; Bias for Action | General Cognitive Ability | Move Fast |
| Disagree and commit | Have Backbone, Disagree and Commit | Leadership | Be Open |
| Mentoring someone | Hire and Develop the Best | Leadership | General leadership signal |
| Cutting scope under pressure | Frugality; Deliver Results | Role-Related Knowledge | Move Fast |
| Production incident | Ownership; Dive Deep | General Cognitive Ability | Focus on Impact |
| Cross-team collaboration | Earn Trust; Customer Obsession | Googleyness | Be Open |
| Unpopular decision | Insist on the Highest Standards | Leadership | Be Bold |
| Going deep on data | Dive Deep; Are Right, A Lot | General Cognitive Ability | Focus on Impact |
For Amazon’s full 16-principle list with question patterns for each one, see the Amazon Leadership Principles cheat sheet. Google and Meta don’t publish a per-question rubric the way Amazon does, so treat this table as a framing guide, not a scored formula.
Rehearsing without sounding rehearsed
A memorized script cracks the moment an interviewer asks a follow-up you didn’t write for — you’re retrieving text, not the memory itself. Rehearse the beats instead: situation in one sentence, task in one sentence, three action beats, one result number, one reflection line. Say each story out loud five to ten times, changing the wording each pass, so what stays fixed is the structure and the facts, not the phrasing. That’s the difference between sounding practiced and sounding canned.
Practice follow-ups on purpose, apart from the story. Have someone throw the standard probes mid-story — “why that and not the alternative,” “what did the other person say,” “how did you measure that” — and answer without breaking stride. Keep an index card with only the beats and the one number that matters, and glance at it while drafting, never while telling it. Record two run-throughs and listen back once for filler words and a rushed result. Build the beats with the STAR method answer builder so each story holds its shape until it’s second nature.
Frequently Asked Questions
How is a story bank different from the Amazon LP cheat sheet?
The Amazon LP cheat sheet maps one company’s 16 named principles to exact interview questions. This story bank is company-agnostic: write the twelve stories once, then re-map each to whichever framework the interviewer that day scores against — Amazon’s LPs, Google’s four attributes, or Meta’s five values. Build the bank first; layer company-specific framing on top once you know who’s interviewing you.
Do I need all 12 stories before my first interview?
No, but get close. Six or seven stories covering the highest-frequency slots — conflict, failure, leading without authority, a deadline, an incident — carry most first rounds. Fill in the rest before onsite week, since a full loop has enough rounds to reach into every slot.
What if one story fits three slots at once?
Good — that’s why 8–12 stories cover a full loop instead of needing 20. A story about fixing a payment outage overnight can hit the production-incident, hard-deadline, and going-deep-on-data slots at once. The only rule: don’t tell the same story to two interviewers in one loop. Keep a backup story for your two or three strongest slots.
Can the same 12 stories work for Amazon, Google, and Meta?
Yes. The material doesn’t change — same situation, same actions, same result. What changes is which parts you emphasize and which vocabulary you reach for in the reflection line. A disagree-and-commit story leans on the principle name at Amazon, on “emergent leadership” language at Google, and on staying open to the outcome at Meta. Don’t build separate banks per company; build one and adjust the framing.
How long should each story take to tell?
Three to five minutes, following STAR plus a reflection line: situation and task under a minute, action as the bulk, result and reflection in the last 30–45 seconds. Time yourself out loud — most candidates run long on setup and short on result, backward from what interviewers score.
Practice the behavioral round:
