The half of the interview that isn't the case.
Candidates rehearse cases for months and their stories on the train. Then the offer turns on the stories. This is how to build one that survives 20 minutes of questions — with a worked example you can pull apart sentence by sentence.
Why it decides offers
10–20 min
one story, deep
10–15 min
inside every case round
~8 questions
breadth over depth
The asymmetry is the point: you cannot bluff a story the way you can waffle through a framework. Every claim you make invites a question you either can or cannot answer.
Tell it in 60 seconds
Run it with a timer and no notes. The first attempt will overrun and you will discover which 40 seconds you were spending on scene-setting nobody asked for. That is the output of the exercise — not the recording.
Setup, action, impact, recap
Orients the listener fast. The cheapest part of the story and the easiest place to lose 30 seconds you needed later.
The part they are actually assessing. Nearly every sentence here should have you as its subject.
Turns a nice anecdote into evidence. A number, a decision that stuck, or a person whose mind changed.
Hands the turn back cleanly and tells them what you think the story proves, so they probe your framing rather than hunting for the point.
A note on length. Published McKinsey guidance puts the opening answer at roughly 3 to 4 minutes, because the interviewer will then spend another 10+ minutes on follow-ups — a 2-minute answer can leave a PEI feeling thin. BCG and Bain want it tighter. Build the 60-second version first either way; it is the only version that tells you what your story is really about.
One story, annotated
In my second year at a logistics firm I was the analyst on a warehouse that was missing its dispatch window 4 days in 5. It had been escalated twice, and everyone I spoke to said the same thing: we were understaffed.
Rather than trust the pick-time data, I spent 2 shifts on the floor with a stopwatch, because I suspected the times were being rounded. They were. The real bottleneck wasn’t headcount — the two busiest aisles shared a single scanner handover point. I rebuilt the pick sequence over a weekend and took it to the site manager as a 2-week trial on one shift, rather than as a recommendation to his director.
Dispatch reliability went from 20% to 86%, and the sequence was rolled out to 4 more sites that quarter. So: a staffing problem that turned out to be a layout problem — and what made it stick was letting the person who owned the floor own the fix. What I took from it: when everyone agrees on the cause, go and measure it yourself.
Hover any part of the story, or click to pin it so it stays while you read — click anywhere else to let go. You get the beat it belongs to, the job it does, and why it earns its place. Consecutive sentences often share a beat and still earn it for completely different reasons.
Read the two versions back to back. The long one is not the short one with sentences added — it is the same story told properly at each length, which is what you will actually have to do. What the minute version loses is the site manager, the cost of the detour and the moment he presented it himself: all excellent material, and none of it needed for the story to be understood. That is the distinction the drill exists to teach.
Impact means you, specifically
We restructured the process and reliability improved a lot.
I rebuilt the pick sequence and took it to him as a 2-week trial; reliability went from 20% to 86%.
Named actor, named action, named number.
I was responsible for the analysis.
I spent 2 shifts on the floor with a stopwatch instead of trusting the reported data, because I suspected the pick times were rounded.
Responsibility is a job title. A decision is evidence.
It was a great learning experience for me.
I’d push for the trial a week earlier — I spent that week building consensus I could have earned with results.
Specific, costed, and it does not flinch.
If a sentence would still be true had you called in sick that day, it is not impact. Cut it or rewrite it with yourself as the subject.
The four dimensions
“Describe a time you built a relationship that changed an outcome.”
Whether you can move people who don’t report to you — and whether you noticed how they felt about it.
“Tell me about something you started that nobody asked you to start.”
Initiative with teeth: you saw a gap, you owned it, and you carried it past the point where it stopped being fun.
“Tell me about a time you led a group through something difficult.”
Not the title. Whether the group got somewhere it wouldn’t have got without you, and how you handled the people in it.
“Describe a time you changed your mind, or something that didn’t work.”
Real reflection. They are listening for what you would do differently, said without flinching and without over-apologising.
That is 8 stories, which sounds like a lot until you realise one strong experience often covers two dimensions from different angles. Build the story once, then decide which beat you lead with depending on what they asked.
Shorter, and it starts before you notice
- Why you
- Why consulting
- Why BCG
Rehearsed but not recited. A memorised paragraph is audible.
BCG asks about failure and ambiguity noticeably more than McKinsey or Bain. Have a real failure ready — one with a cost you can name and a lesson you actually applied afterwards.
Breadth, and prove it moved
Results orientation
Ambitious goal, measurable outcome. Bain wants the number.
Teamwork
Especially through disagreement, not just alongside people.
Leadership
Mobilising others toward something difficult.
Entrepreneurial spirit
Building or improving when nobody asked you to.
Aim for 3 to 5 flexible stories, each mapped to at least two of these; that beats 8 rigid ones. With 8 questions coming, you will need the same experience to answer more than one of them from a different angle.
The follow-up storm
What were you actually worried about at that point?
Who disagreed with you, and what did they say?
What would you have done if he’d refused the trial?
How did you know the pick times were being rounded?
What did that cost you personally?
What would you do differently?
Notice that none of these are about the outcome. They are about your reasoning, the people, and the moments you were uncertain — which is why a story you lived beats a story you selected for how it sounds.
How good stories die
The setup ate the clock
Org chart for the first 90 seconds. They stopped listening 40 seconds in.
“We” all the way through
A team did something admirable and your part is unrecoverable from the telling.
No number, no changed mind
It ends with “it went well”. Nothing distinguishes it from a story that didn’t.
Too polished to be true
No hesitation, no cost, no moment you got it wrong. Follow-ups find the seams fast.
Chosen for the title, not the content
The most senior-sounding project, where you did the least.
The failure story with no failure
“My weakness is I care too much.” BCG in particular will keep digging until it finds a real one.
Your story bank
Every dimension needs a lead story and a spare
Fully built: the long version, the 60-second cut, and 20 follow-ups you have already been asked out loud by someone else. This is the one you open with.
A different experience entirely — other people, other setting. It needs the 60-second cut and about 5 follow-ups, not a second 4-minute showcase. You are proving you have more than one life, not performing twice.
Why you get asked
“Do you have another example?” — sometimes a genuine probe, sometimes the first one did not land.
They interrupt: “that is more of a teamwork story — anything where you led?”
The question they actually ask fits your second story better than the one you planned.
Later round, same dimension, and the notes from the first interview are in front of them.
Draw the spare from a different part of your life — if the lead is from work, take the spare from university, a side project, or something you ran outside your job. Two stories from the same team 6 weeks apart get heard as one story, and you will have spent the preparation without buying the cover.
Building the bank
Draft it long, then cut to 60 seconds
The cut is the exercise. Whatever survives is the spine.
Put a number on the impact
Or a decision that stuck, or a person who changed their mind.
Map each story to at least two dimensions
One experience, different lead-in depending on what they asked.
Write the version where you got something wrong
Every firm asks. BCG asks hardest.
Have someone ask you 20 follow-ups
You will find the seams faster than by rehearsing alone.
Never reuse a story inside one round
Interviewers compare notes between rounds.
One last thing, and it is the one candidates resist: pick the story where you were genuinely uncertain, not the one with the most impressive logo. The uncertainty is what you get to talk about for 15 minutes.
Sources: McKinsey's own careers guidance on the PEI and its four dimensions, plus published 2026 guides on the BCG fit interview and the Bain experience interview. Firms adjust their process — check your invitation email for the format you have been given.