What does consistent mean when your answers cross several interviews?
It means the facts in a story match every time you tell it: the numbers, the dates, your title, who did what and how it ended. The wording can change, and it should, because each interviewer asks from a different angle. Your stories can change shape between rounds, and the numbers inside them should stay exactly the same.
Most loops are built so that several people hear your stories. Amazon describes its final stage as an interview loop where you meet employees individually, and says "each person will assess different aspects of your skills and experience." When notes from a loop like that are read together, a number that moved can stand out more than any single weak answer.
Why do stories drift between rounds when nobody is lying?
Memory rebuilds a story each time you tell it, so small details shift without you noticing. The psychologists Daniel Schacter and Donna Rose Addis describe memory as "fundamentally constructive, rather than reproductive," and "prone to various kinds of errors, illusions and distortions." A number you never wrote down is one of the easiest parts to rebuild differently.
Time makes it worse. A 2015 replication of Hermann Ebbinghaus's forgetting curve from 1880 got results similar to his, and Ebbinghaus derived that curve from "a gradually slowing forgetting process." A gap of, for example, two weeks between round 1 and round 3 is plenty of time for a detail to fade and be filled back in.
Practice adds a third push. Each time you rehearse a story it gets smoother, and a rounder or larger number starts to sound right. Nobody decides to inflate anything. The story just settles on the version that is easiest to say.
What is the "team of 6, team of 9" problem?
It is a fact that drifts for an innocent reason and still reads as a warning sign. For example, in round 1 you tell the recruiter you led a team of 6 engineers. In round 3 you tell the hiring manager you led a team of 9, because by then you're also counting 2 designers and a contractor.
Both numbers are true. In two sets of notes, side by side, they look like a candidate who grows the team when the stakes go up. Nobody can ask you which is right, because by the time anyone compares, you're out of the room.
The fix is one sentence you say the same way every time. "I led 6 engineers, and 9 people in all once you count 2 designers and a contractor." Now the 6 and the 9 both appear, each with its meaning attached, and the next interviewer hears the same thing the last one did.
What is the 5-item consistency check?
It is a check of five facts after every round: your numbers, your dates, your role, who did what, and the outcome. I check these five because each one is easy to change without noticing. For each, compare what you said today with what you said before, and decide which version is the true one.
| # | Check | What to compare | Drift, for example | The fixed version, for example |
|---|---|---|---|---|
| 1 | Numbers | Team size, budget, revenue, percentages | "A team of 6" in round 1, "a team of 9" in round 3 | "6 engineers, 9 people counting design and a contractor" |
| 2 | Dates and durations | When it started, how long it took | "Six months" to the recruiter, "about a year" to the manager | "From January to July 2024, so about six months" |
| 3 | Your role | Your title and who you reported to | "I ran the project" in one round, "I was on the project" in the next | Your actual title, and the decisions that were yours to make |
| 4 | Who did what | Where you say "I" and where you say "we" | "I built the model" became "we built the model" | Name your part and name the team's part, in that order |
| 5 | The outcome | The result and the figure attached to it | "Cut churn by 20%" became "cut churn a lot" | The figure, what it measured and over what period |
Write the fixed version once, on a single page, one line per story. That page is your facts sheet. Before each round you read it, and after each round you check what you said against it.
How do you run the consistency check after each round?
Do it the same evening, in about 10 minutes, after your debrief notes. List every number, date and title you said today. Put the list next to your facts sheet and your notes from earlier rounds. Anything that differs gets a decision on which version is true and how you'll say it from now on.
The post-interview notes template has a part for exactly this, called the numbers you said, and the printable 10-minute debrief template has the same prompts. If you haven't done a debrief on yourself before, start with what an interview debrief is.
The US Army runs the same kind of review after training. Its circular on after-action reviews, TC 7-0.1, lists follow-up actions that include "revising standard operating procedures," according to the Army's announcement in 2025. Your facts sheet is your standard procedure, and each round is a chance to revise it.
Keep this as analysis after the call, never during it. In the room you want your attention on the person asking, and a sheet in front of you pulls it away.
If you record your rounds with everyone's consent, Prept's debrief flags it when a number you give changes from one round to the next and asks you which one is true. The first debrief is free and needs no card.
What should you do if your story already changed?
Pick the true version and use it from now on. If the change was small and nobody asked about it, you don't need to raise it. If an interviewer asks, or the number moved a lot, correct it in one calm sentence in your next round. A plain correction sounds careful, while a long defense sounds like you were caught.
Here is the sentence, for example: "Earlier I said a team of 9. To be precise, I managed 6 engineers, and the wider group was 9 with design and a contractor." Say it once, then go on with your answer.
Plan the correction as part of preparing for round two, so you aren't working out the words under pressure. If the drift came up in the room and the interviewer's tone changed, the page on how to tell you lost the room will help you judge how much it cost.
Is it bad to tell the same story to different interviewers?
No. Reusing a strong story across rounds is normal, and it often makes sense, since each interviewer may be scoring something different. What matters is that the facts inside it match. Change the angle to fit the question, and keep the numbers, dates and your role exactly as they were.
A good story has a fixed core and a flexible frame. The core is the facts sheet. The frame is which part you lead with, and that changes with the question. The STAR method guide covers how to build the Result around a number, so the core holds up to a follow-up.
What you want walking into round 3 is a small set of facts you can defend, told the same way to everyone who asks.