Owning the story: "I", results, reflection and the follow-up drill

Engineering is collaborative, and describing collaborative work accurately in the first person feels like taking credit. Getting this wrong in either direction is expensive.

The frame from the previous lesson tells you where each part of a story goes. This lesson is about the parts that decide whether the story is believed: whether the interviewer can find you in it, whether the result holds up when questioned, whether the reflection shows judgement — and whether the whole thing survives the follow-up questions, which is where most of the evidence in a behavioral round actually comes from.

Why "we" is fatal to a score

The rubric scores a person. An interviewer who cannot locate your contribution has nothing to write in the row.

An invented but typical exchange:

Candidate: So we decided to move off the shared database. We looked at a few options and we went with per-service databases, and we did the migration over about four months with dual-writes.

Interviewer: What was your role in that?

Candidate: I was part of the team that worked on it.

Interviewer: Which part did you own?

Candidate: I mean, we all kind of worked across it.

Three questions in, and the interviewer has nothing. The write-up will say: "Unable to establish individual contribution across two stories. Recommend no hire at senior; possible at mid."

Note the interviewer had to ask twice. That drilling is itself recorded — "had to be prompted repeatedly to describe own contribution" is a line that appears in real write-ups.

Why "I" alone is also a problem

The opposite failure is rarer and reads worse when it happens:

"I designed the whole thing, I wrote the migration tooling, I ran the cutover, I fixed the issues afterwards."

For a four-month, multi-service migration, this is not credible. The interviewer's follow-up is "who else was involved?" and the answer either reveals a team you erased or reveals that the project was smaller than it sounded. Both are bad.

At senior levels there is a second problem: a story where you did everything yourself is evidence that you cannot delegate or influence, which is the job.

The pattern that works

"We" for the frame. "I" for the decisions and the actions. Named others for their parts.

"We had agreed to move off the shared database — that decision was made before I joined the effort. I owned the write path: the dual-write layer and the consistency checks. I proposed we shadow-read for two weeks before cutting over, which wasn't in the original plan; the team had planned a straight cutover with a rollback. I built the comparison job that caught the discrepancy in currency rounding — that would have shipped as a real money bug. My colleague Anouk owned the schema design and Dev owned the read path, so I can talk about their pieces at a high level but the write path is where I can go deep."

That last sentence is a strong move. It is an honest boundary, and it pre-empts the follow-up that catches people out.

Three phrases worth having ready

• "I owned X, which meant I decided Y" — attaches a decision to an ownership claim.
• "The team's plan was A; I argued for B because —" — locates you inside a group decision.
• "That part I can only speak to at a high level; my piece was —" — sets an honest boundary before you are caught at one.

Result and reflection

Once the interviewer knows what you did, the last forty-five seconds of the story carry a disproportionate share of the score, and they are the part most often improvised.

Quantifying honestly

Every result should answer: what changed, by how much, measured how.

When you have no number

Three honest routes, in preference order:

1. Quantify the input. "I migrated 40 services off the deprecated library ahead of its end-of-support date." You could not measure the benefit; you can count the work.

2. Name what became possible. "That unblocked the same-day delivery launch, which had been waiting on it for two quarters."

3. State the qualitative change precisely. "The Sunday-night manual restart stopped being a thing anyone had to do." Precise and unquantified beats vague and quantified.

Reflection: judgement, not apology

Reflection has one job: show that you evaluate your own work. It is not a confession.

Three shapes that work:

The earlier decision. "I'd have gone to the controller directly in week two instead of arguing through my manager for a month. I was treating it as an org problem when it was a conversation."

The changed model. "I went into it thinking the problem was ranking. It was tokenisation. Since then I read a hundred real failing examples before I form a hypothesis — that habit came out of this."

The trade-off I would take differently. "We chose the dual-write for safety and it cost us six weeks. Knowing what the discrepancy rate turned out to be, I'd have shadow-read for one week rather than two and shipped a month earlier."

Two shapes that do not:

"Honestly I don't think I'd change anything, it went really well." — closes the door on the highest-value part of the answer. There is always something.

"I felt terrible about it. I really let the team down and I still think about it." — the interviewer is assessing judgement, not remorse. Emotional disclosure reads as instability under pressure, particularly for roles with on-call or incident command.

The distinction that matters

Reflection is about the decision, not the feeling. "I decided X; given what I know now, Y was better, and here is the general rule I took from it" is the shape. If your reflection sentence contains no decision, rewrite it.

Surviving the follow-up drill

Everything so far is the story. But most of the evidence in a behavioral round comes from follow-ups, not from the story. The story is the interviewer's entry point; the drilling is the interview.

The three moves

Almost every follow-up is a variant of three questions:

1. Why that choice? Tests whether the decision was reasoned or inherited. 2. What was the alternative? Tests whether you considered one at all. 3. What went wrong? Tests honesty and whether you were close enough to the work to know.

They compose. "Why that choice" leads to "what would have changed your mind", which leads to "did you test that assumption".

The drill tree

Notice: the tree is three layers deep and the leaves converge. Interviewers reach the same place — did you actually check, or did you assume? — from any starting branch.

Preparing two layers below the surface

For each story in your bank, write the answers to these six, in full sentences:

1. Why did you choose that approach over the obvious alternative?
2. What was the alternative, and what would it have cost?
3. What assumption were you making, and did you verify it?
4. Who disagreed with you, and what was their argument at its strongest?
5. What went wrong that you did not mention in the story?
6. What would have made you abandon the approach?

Question 4 is the one candidates skip and the one that separates scores. Stating an opponent's argument at its strongest is a trust signal, and it is very hard to improvise.

What to do when you do not know

Say so, then give the shape of the answer:

"I don't know the exact number — that dashboard is gone. What I can tell you is the order of magnitude: it was hours per week, not minutes, because it was blocking a person's morning."

Or:

"Honestly, I didn't consider that at the time. Thinking about it now, I'd have —"

Both are strong. Both are far stronger than a confident invented figure that falls apart when the interviewer asks how it was measured.