An assessment centre is not a place. It is a method: several candidates, several exercises, several trained observers, and a rule that every judgement about a candidate must be tied to behaviour that was actually seen. The easiest way to see how this works is to follow one candidate through one day. Our imaginary employer is a retail chain hiring first-time shift managers, and the candidate is Divya, a store supervisor from another company.
Before the day: deciding what to measure
Nothing happens at the centre until someone has done a job analysis of the shift manager role and settled on a short list of dimensions, the behaviours that separate good from poor performers. Here the team chose three: planning and organising, communication, and influencing others. Three or four is wise. Observers cannot reliably track ten things at once, and when organisations list ten, the scores tend to collapse into a single general impression.
Then the exercises are designed so that each dimension is visible in at least two of them. The mapping is a simple grid, dimensions down the side and exercises across the top. For this day: an in-tray exercise, a leaderless group discussion and a role play with an upset customer played by a trained actor-assessor.
Morning: the in-tray
Divya gets forty-five minutes and a pile of simulated mails, messages and a rota: two staff calling in sick, a delivery delayed, a manager’s request for a report by noon. She must decide what to handle, what to delegate and what to defer, and write a note explaining her choices. The observer is looking at planning and organising. Does she spot that the delayed delivery and the absent staff collide on the same Saturday? Does she give reasons? She scores 4 on planning in this exercise.
Midday: the group discussion
Six candidates, a brief about allocating a limited festival-season bonus pool across stores, thirty minutes, no appointed leader. Two observers watch Divya and a third candidate. They write down what she said and did, not what they felt about it. “Asked the quiet candidate for her view; proposed a time split for the three options; summarised twice” is an observation. “Seemed like a natural leader” is not. After the exercise, each observer classifies notes under the dimensions and gives ratings against behavioural anchors, such as “3 = contributes relevant points but does not check understanding”. Divya scores 3 on planning, 4 on communication and 3 on influence.
Afternoon: the role play
She is told a customer is angry about a refund refused by a colleague. The role player follows a script with set escalation points so each candidate faces a similar situation. The observer watches for communication and influence: does she listen before offering solutions, stay calm, find a workable resolution? She scores 3 on communication and 2 on influence; she conceded too quickly and did not hold the store’s policy limit.
The integration meeting
This is the step that separates a proper centre from an afternoon of games. Observers do not vote on a gut feeling. Each reads out their evidence and rating for each dimension in each exercise. The grid for Divya looks like this.
| Dimension | In-tray | Group | Role play | Dimension score |
|---|---|---|---|---|
| Planning and organising | 4 | 3 | – | 3.5 |
| Communication | – | 4 | 3 | 3.5 |
| Influencing others | – | 3 | 2 | 2.5 |
Dimension scores are averages of the exercise scores. If the team decided in advance that influencing matters most for this role, they might weight it: 30 percent planning, 30 percent communication, 40 percent influence. That gives 0.3 x 3.5 + 0.3 x 3.5 + 0.4 x 2.5 = 1.05 + 1.05 + 1.0 = 3.1 out of 5. Unweighted, the mean is about 3.17. The weighting changes little here, but with a different profile it could decide the outcome, which is why it is fixed beforehand and not argued afterwards.
The rating alone is not the decision. Observers also discuss the pattern: Divya is organised and communicates clearly but is weaker when she has to push back. The hiring manager must decide if that is trainable in the first six months or a deal-breaker. That conversation is better informed than it would have been after a single interview.
Where it goes wrong
- Untrained observers who rate on personality or on who spoke most.
- Exercises not linked to the real job, so candidates are judged on their skill at games.
- Too many candidates per observer, causing shallow notes.
- Cost: a centre needs rooms, role players and several observers’ days. It makes sense for roles where a wrong hire is expensive, such as management trainees and leadership pipelines, and little sense for high-volume entry roles.
Used well, it gives candidates a fair chance to show the behaviour the job demands, and gives the organisation evidence that does not depend on one interviewer’s impression.
Figures and names in the examples are illustrative.
Leave a Reply