9-box grid

How the grid works

Each person is rated on current performance (against expectations in the present role) and assessed potential (capacity to grow into bigger or different roles), each on a three-point scale. The intersections create the familiar cells: high/high stars earmarked for stretch and succession, high-performance/low-potential deep experts to be retained and honoured rather than promoted sideways into misery, low/low cases requiring an honest conversation, and the corner everyone forgets, high-potential/low-performance, which is usually a person in the wrong role rather than a wrong person.

Where it goes wrong

The performance axis is at least anchored in results; the potential axis is where bias moves in. Unstructured "potential" judgements reliably favour people who resemble their assessors, and a label assigned in one calibration meeting can shadow a career for years. The grid also tempts organisations into secrecy, ratings people are never told about, driving decisions they never understand, which corrodes exactly the trust talent processes are meant to build.

Using it defensibly

Treat the grid as a conversation structure, not a verdict. Define potential behaviourally before the meeting (learning agility, capacity for complexity, leadership evidence), bring data to both axes, including assessment evidence rather than recollection, calibrate across managers to iron out easy and harsh raters, and pair every placement with an action: development, stretch, retention risk plan or role change. A grid that produces actions is a planning tool; one that produces labels is a filing system for bias.