Competency-based interviews: how panels actually score you

Somewhere in your future is a video call with three to five serious people, a shared scoring sheet, and questions that all begin the same way: “Tell us about a time when…”. The competency-based interview (CBI) is the standard selection interview across the UN system and most international organizations. It is also the stage where otherwise strong candidates fail most predictably — not because they lack the experience, but because they answer a structured assessment as if it were a conversation.

The good news: the CBI is the most learnable stage of the entire hiring pipeline. The framework is public, the scoring logic is mechanical, and preparation compounds. This guide explains all three.

What a competency-based interview is

The premise is simple: past behaviour is the best available predictor of future behaviour. Instead of asking what you would do (hypothetical) or what you believe (opinion), the panel asks what you actually did in specific past situations, then scores your account against defined behavioural indicators. Structure is the point: every candidate faces broadly the same questions, each mapped to a competency listed in the vacancy announcement, and each panellist scores independently before the panel agrees a consensus rating. This design deliberately strips out charisma, small talk and improvisation. A charming candidate with vague answers loses to a nervous candidate with precise evidence, every time.

The UN competency framework

The Secretariat’s framework — echoed with variations across funds, programmes and agencies — has three layers. Read the vacancy announcement: the competencies it lists are, with high reliability, the interview plan.

Agencies rename and regroup these — UNDP, UNICEF, WHO and others each publish their own frameworks — but the underlying behaviours barely change. If you can evidence the Secretariat’s core set, you can adapt to any variant in an afternoon by reading the agency’s published framework next to the vacancy text.

STAR, and why panels can hear it

The expected answer shape is STAR: Situation (brief context), Task (your specific responsibility), Action (what you personally did — the heart of the answer), Result (what happened, ideally verifiable). Many panels also listen for an L — what you learned. Two calibration points that distinguish practised candidates:

• Say “I”, not “we”. Panels can only score behaviour they can attribute to you. “We redesigned the distribution plan” earns nothing; “I proposed and drafted the redesign, then negotiated it with the field coordinator” earns marks.
• Spend the time where the marks are. A strong answer runs two to three minutes: roughly ten percent situation, ten percent task, sixty percent action, twenty percent result and learning. Weak candidates invert this and drown in context.

Worked example one: teamwork

“Tell us about a time you had to work with colleagues who disagreed with your approach.” (Both examples here are invented for illustration.)

Situation: “In my last role I coordinated a joint needs assessment after flooding displaced several thousand households. Our team and a partner NGO had to produce one shared report in ten days.” Task: “I was responsible for the methodology, and the partner’s lead insisted on household surveys while I believed we only had time for key-informant interviews.” Action: “Rather than escalate, I asked for a working session to map what decisions the report had to support. I drew out that donors needed severity ranking, not precise figures. I then proposed a hybrid: key-informant interviews everywhere, plus a small household sample in the three worst sites, and I volunteered my team for the harder travel. I documented the agreed method in one page and circulated it the same evening so neither side could re-litigate it.” Result: “We delivered on day nine; the ranking was adopted for the allocation decision, and the partner requested the same split method in the next assessment. I learned to negotiate from decision needs, not from methods — it turns a turf argument into a design problem.”

Notice what this answer does: names a genuine disagreement (panels distrust conflict-free stories), attributes five concrete actions to “I”, and closes with a result someone could verify by phone.

Worked example two: planning and organizing

“Describe a time when you had to manage several competing deadlines.”

Situation: “As a programme associate I supported three grants whose reporting deadlines landed in the same month — the same month our office moved premises.” Task: “I owned the data annexes for all three reports and the logistics folder for the move.” Action: “I listed every deliverable and worked backwards from each deadline, flagging that two annexes depended on field data that historically arrived late. I brought the collection forward two weeks by agreeing a simplified template with the field focal points, blocked my mornings for report work before email, and handed the two move tasks that were purely administrative to a colleague, with a one-page handover. When one dataset still arrived late, I informed the reporting officer four days ahead — not the day of — and we agreed a placeholder approved by the donor focal point.” Result: “All three reports went out on time, and my backwards-planning sheet was adopted as the office template for the next cycle. The lesson I took: escalate early with a proposed solution, and a slipped dependency never becomes a slipped deadline.”

How panels actually score

Behind the screen, each competency has behavioural indicators — phrases like “solicits input by genuinely valuing others’ ideas” or “foresees risks and allows for contingencies”. Panellists take near verbatim notes, then rate each competency on a defined scale, and the panel writes a consensus justification for the selection record. Three consequences for you:

• No evidence, no marks. A wise general statement about teamwork scores zero. The scale rewards specific, attributable, recent behaviour.
• Probing questions are help, not hostility. “What exactly did you say?” means the panel needs more scoreable detail. Give it gladly.
• One story per competency. Reusing the same anecdote for three questions caps your evidence. Panels notice.

Build an evidence bank before you need it

Preparation is not rehearsing answers; it is stocking a matrix. Down the side, eight to twelve real episodes from your work — projects, crises, conflicts, failures, deliveries. Across the top, the competencies from the framework. Fill the grid: most good stories evidence two or three competencies, and the exercise exposes your gaps while there is still time to choose different examples. For each story, write only STAR bullet points — five to seven lines, never a script — and say them aloud a few times. Keep failures in the bank deliberately: questions about mistakes and negative feedback appear in most panels, and the scoring indicator is whether you took responsibility and changed something, not whether you were blameless. The same bank, incidentally, is the raw material for your PHP achievements and your 90-second introduction — build it once, spend it three times.

Remote interview practicalities

Most first-round panels are now video calls, frequently across time zones. The mechanics are simple and worth doing properly: camera at eye level with light on your face, not behind you; a wired or reliable connection, with the call app tested on the actual account you will use; your one-page competency matrix taped beside the lens — glancing at headings is fine, reading is audible. Confirm the time zone arithmetic twice; a missed panel is rarely rescheduled. Expect the format to feel abrupt — timed questions, note-taking silences, little social feedback. That is the structure working as designed, not a sign you are failing. Pausing three seconds to pick your story is read as composure, and asking for a moment to think costs nothing on the sheet.

Questions worth asking at the end

The closing “do you have questions for us?” is unscored in most processes, but it is your only chance to gather intelligence. Skip salary — grades make it public arithmetic — and ask about the work: what the team most needs from this post in the first six months; how success would be visible after a year; what the hardest part of the role honestly is. Questions like these signal that you are already thinking from inside the post, and the answers tell you whether you want it.

A competency-based interview is a scored assessment with a published syllabus. Treat it that way: read the vacancy announcement as the exam paper, stock the evidence bank, practise attribution and timing, and walk in knowing that the panel is not looking for the most impressive person on the call — only for documented proof that you have already done what the post requires. Then go find the post: the board lists every live vacancy and each one names the competencies it will test.