The STAR method, properly

Everyone knows STAR. Situation, Task, Action, Result. It is the first thing a careers service teaches and the last thing most people do well.

The usual failure is easy to spot. Someone spends four sentences setting the scene, one on what they did, and none at all on what changed. The answer sounds busy. It proves nothing. A hiring manager reading forty of these in a morning remembers none of them.

STAR is not the problem. Half-finished STAR is the problem. This guide is about the half almost everyone skips: a Result stated in numbers, and a fifth letter that turns a competent answer into a senior one.

The two letters people can do

Situation and Task get plenty of attention, usually too much. Keep them short. The reader needs enough context to understand the stakes, and no more.

• Situation. Where you were and why it mattered. One sentence, sometimes two. “Our onboarding took six weeks and a third of new hires quit before month three.”
• Task. What you owned. Not the team’s remit, yours. “I was asked to cut the drop-off without adding headcount.”

If you find yourself explaining the reader’s own industry back to them, stop. That is padding. They know their world. They want to know what you did in it.

Action: what you did, not what the team did

Action is where most CVs go passive. “Was involved in”, “supported”, “helped to deliver”. Those verbs hide you. Recruiters read them as a signal that you cannot name your own contribution.

Use plain verbs that point at a decision you made. Built. Rewrote. Renegotiated. Cut. Merged. Each one implies a choice, and a choice implies judgement, which is the thing being assessed.

One honest constraint helps here. Real work has friction. If the first version broke, say so, then say how you fixed it. “The initial rollout stalled on a data issue, so I sequenced the migration in batches.” That single admission does more for your credibility than three polished sentences, because it reads like memory rather than a script.

Result: the letter that carries the answer

The Result is the whole point, and it is the one most people leave vague. “It was a big success.” “Feedback was positive.” “It made a real difference.” None of these survives contact with a sceptical reader.

A Result needs a number. Before and after, or a rate, or a saving. “Drop-off fell from a third to under a tenth in two quarters.” “We closed the role in eleven days instead of the usual six weeks.” The number is not showing off. It is the evidence that the story is true.

Two rules keep the Result honest. Pick the story for relevance first, then reach for the number. A smaller number on the right competency beats a bigger number on the wrong one. And never trade the truth of the number for its size. An estimate you can defend is worth more than a round figure you cannot.

The L: the letter that signals seniority

Here is the extension almost no one uses. After the Result, add a Lesson.

STAR becomes STAR-L: Situation, Task, Action, Result, Lesson. The Lesson is one sentence on what you took from the work and did differently afterwards. “I now pilot any process change with one team before a full rollout.” “It taught me to agree the metric with the sponsor before I start, not after.”

The Lesson does something the other four letters cannot. It shows reflection. Anyone can describe a win. A senior candidate can tell you what the win taught them, which is the difference between someone who did a thing once and someone who knows why it worked and could repeat it. Interviewers are listening for exactly that. The Lesson hands it to them.

Keep it to a single line. A Lesson that runs long turns into a lecture, and a lecture undoes the modesty that made the reflection land.

Fermi-estimating the metric you never wrote down

Now the hard part. You know the work made a difference. You never measured it. The number is gone, if it was ever written down at all.

Do not abandon the story, and do not invent a figure. Reconstruct one. This is Fermi estimation, named after the physicist who was known for good order-of- magnitude guesses from almost nothing. You build the number from things you do remember.

Say you cleaned up a reporting process. You do not know the hours saved. You do know the shape of it:

• The report ran weekly. That is about fifty times a year.
• Three people used to spend most of a morning on it. Call it three hours each.
• After your change it took one person about twenty minutes.

Before: roughly nine hours a week. After: about twenty minutes. You saved on the order of eight hours a week, or something like four hundred hours a year across the team. You did not measure that. You reasoned it from parts you are sure of.

Two disciplines make this safe. Round down when you are unsure, so the figure is a floor rather than a boast. And label it as an estimate when you say it out loud: “roughly”, “on the order of”, “about”. A reader trusts an honest estimate. They distrust a suspiciously precise one, and they bin an invented one the moment a follow-up question exposes it.

The line you never cross: estimate the size of a real effect, never the existence of one. If the work happened, reconstructing its scale is fair. If it did not, no amount of arithmetic makes it true.

Three worked examples

Here is the method applied. Each starts with the version people actually write, then the same story finished properly.

Operations

Before. “Responsible for improving our warehouse processes. Worked with the team to make things more efficient and reduce errors. This was well received by management.”

Nothing is provable. No number, no decision, no lesson.

After. “Our pick-and-pack error rate was running at about one in twenty orders, and returns were climbing. I owned the fix. I rebuilt the picking sequence around location rather than order number and added a two-second scan check at dispatch. Errors dropped to roughly one in a hundred over three months, which cut returns handling by an estimated fifteen hours a week. It taught me to instrument a process before changing it, so the improvement is a measured fact and not a claim.”

The error rate is reconstructed, not invented. The decision is named. The Lesson signals someone who would do it again, better.

Customer-facing

Before. “Handled a portfolio of key accounts and built strong relationships, consistently delivering excellent service and hitting my targets.”

Every candidate says this. It reads as a description of the job, not of the person.

After. “I inherited a portfolio of twelve accounts, two of which had given notice to leave. I sat down with both and rebuilt their onboarding around the outcomes they actually cared about rather than our feature list. Both renewed, and one expanded its contract by about a third the following year. I learned to ask a leaving customer what they were promised, because the gap is usually the whole problem.”

Concrete numbers, a specific action, and a lesson about diagnosis over persuasion.

Individual contributor, no obvious metrics

Before. “Wrote internal documentation and helped colleagues understand our systems, which improved the team’s knowledge sharing.”

Soft work, softly described. It vanishes on the page.

After. “New engineers were taking about six weeks to make their first solo change, mostly lost in undocumented systems. I wrote the missing runbooks and reorganised them by task rather than by system. The next three joiners shipped their first change in around two weeks. I stopped writing reference docs and started writing ‘how do I do X’ docs, because that is the question people actually ask.”

There was no dashboard for this. The six-weeks-to-two figure is a defensible estimate from onboarding memory, and the Lesson is a genuine change in approach.

The short version

Cut Situation and Task to the bone. Put yourself back into the Action with plain verbs and one honest constraint. Finish the Result with a number, reconstructing it from what you remember if you have to. Then add the Lesson, in one line, and let it do the work of proving you would do it all again.

Do this for your best eight or ten stories and you have an evidence bank. Every application, every interview, every “tell me about a time when” draws from it. The work is in building it once.

---

That building-it-once is the tedious part, and it is the part we made a tool for. Championed runs a forensic interview that pulls these stories out of you, presses for the number when you have skipped it, and holds each one as reusable STAR-L evidence. It frames what you actually did. It will not invent a result you never had. Pilots open soon, and the waitlist is the only thing to join for now.