Why do prepared candidates draft a story's closing two sentences before writing the rest of the story?

## The ordering problem Stories prepared front to back tend to fail at the end, because the end is where preparation runs out. The candidate has rehearsed the context thoroughly, knows the actions well, and has never actually decided what the last sentence is. Under pressure the answer reaches the actions, finds nothing scripted after them, and stops: "...and then we moved on to the next thing." Writing the close first removes that failure mode structurally rather than by discipline. If the last two sentences already exist, the answer has somewhere to arrive. ## What the two sentences are **Sentence one — the outcome.** Past tense, and observable by someone other than you. From a backend example: "After the change the reconciliation job failed on none of the next 23 nights, and its runtime came down from 19 minutes to about 4." **Sentence two — the reflection.** What transferred, and where you used it. "That taught me to check the retry path before the query plan, and it is where I have started on every batch problem since." That is the whole artefact: two written sentences per story. Not a page of notes, not a script for the full answer — the close, in the words you intend to say. ## What writing it first changes ### It makes the setup selectable Once the destination is fixed, every earlier sentence has a test: does the listener need this in order for the close to land? The fact that the job existed and was failing is needed. The reorganisation that moved the service to your team six months earlier is usually not. This is the same trimming you would do anyway, but with a criterion instead of a feeling. ### It exposes hollow stories at your desk Some stories cannot produce an honest outcome sentence. You worked hard, the work was real, and nothing you can point to changed — or you no longer remember what changed. Finding that out while writing is cheap. Finding it out mid-answer is not, and the escape hatch people reach for in that moment is the conditional: "we would have seen the failures drop, but the project was reprioritised." A story that can only be closed in the conditional should not be in rotation until you have reconstructed a real ending for it. ### It makes the answer survivable when time is cut short If a listener interrupts to move on, a candidate with a written close can land it in one breath rather than losing it. Knowing the last two sentences means you can jump to them from anywhere in the story. ### It keeps the reflection from becoming a speech Written in advance, the reflection stays one sentence. Improvised at the end of a long answer, it tends to expand into general philosophy about engineering, which is the point where a listener stops taking notes. ## How to actually do it For each story you intend to use: 1. Write the outcome sentence. If you reach for a round number, stop and find the real one — awkward figures such as 23 nights or 19 minutes read as recalled, while smooth ones read as manufactured. 2. Write the reflection sentence, and make sure it names a *later* occasion, not just a feeling. 3. Read both aloud, on their own, with no story in front of them. If they do not stand up alone, the story does not stand up either. 4. Only then decide how much situation and which actions you need. A mock-interview partner is useful here in a narrow way: have them read only your two closing sentences back to you and ask what the story must have been. If they can reconstruct roughly the right shape, the close is carrying its weight. ## The limit of the technique Writing the close first is a preparation habit, not a delivery script. In the room the close should sound spoken, not recited — the same two facts, in whatever words arrive. The value was never the wording; it was having decided, in advance and in writing, that the story has an ending and what it is.