The 9-box grid: how to read each box and four mistakes that make it useless

Echipa HR 365 · reviewed 2026-09-02 · 5 min read

The 9-box grid places each person on two axes: current performance and the potential to take on a bigger role. Its value is not the classification but the fact that it forces a conversation about each person between several managers — if that conversation does not happen, the grid stays labelling with no effect.

The two axes, and why the second is the hard one

Performance is measured on the results of the last 6–12 months, in the current role. That is the easy part: there are reviews, objectives, figures. Potential is something else — not “how good are they”, but “how likely are they to succeed in a significantly larger or different role”. An excellent specialist who neither wants nor is able to lead people has high performance and low growth potential, and that is not a criticism.

The nine positions and what you do with each

The top-left box is the one most often missed: somebody with high potential and weak performance is usually a person in the wrong role, not a weak person. The cost of replacing them is higher than an internal move, and the move often also resolves the problem that triggered the assessment.

The bottom-right box — high performance, low growth potential — is the most valuable and the worst handled. These are the people the operation leans on. Promoting them into management “as a reward” loses an expert and produces an unhappy manager, in the same gesture.

Four mistakes that make the exercise useless

1. A single assessor. A grid completed by one manager alone reproduces their opinions. The functional minimum is two or three managers who work with the same people and have to agree out loud.
2. A forced distribution. “At most 10% in the top box” turns a conversation about people into a quota negotiation. If a team genuinely has four strong people, the figure has to show it.
3. The result stays in a file. Without a written action for each box and a deadline, the exercise repeats next year with the same positions and the same conclusions.
4. People learn where they were placed, indirectly. The 9-box is not communicated individually as a label. What is communicated is the action that follows from it: a new task, a plan, a corrective conversation.

The calibration meeting: where the value is actually produced

What a session worth three managers’ time looks like

1. HR — sends the grid pre-filled by each manager, separately, 3 days ahead
2. Managers — discuss only the people they placed differently — the rest are approved silently
3. The manager proposing an extreme position — brings the two concrete observations; without them, the position moves towards the centre
4. HR — records for each person one action and a deadline, in the meeting, not afterwards
5. HR — revisits the list in six months and checks what happened to each action, not to each position

The second step saves the most time: in a team of thirty, managers agree on twenty-five and argue about five. Only those five deserve a meeting; the others consume hours without changing anything.

The minimum succession plan the grid produces

The most concrete use of the grid is not individual development but covering critical roles. It takes three steps and fits on one page:

From the grid to a coverage plan

1. Leadership — marks the critical roles: those whose month-long absence visibly stops or slows something
2. HR — for each critical role, looks in the grid for who could take over in six months and who in two years
3. Manager — gives the likely successor a task from the target role in the next quarter, so the test happens now rather than at the moment of need

Critical roles with no name in the “six months” column are the only genuinely urgent output of the whole exercise. There are usually two or three in a mid-sized organisation, and identifying them justifies the time spent on the grid by itself.

When not to use it at all

The grid assumes you already have performance data you can stand behind. In an organisation with no reviews and no written objectives, it adds no rigour — it merely gives a respectable shape to impressions. In that situation, the first step is the review form, not the grid.

The second case where it is not worth it: when the decision is already made. A grid built to justify a promotion or a restructuring decided in advance is spotted immediately by managers and compromises the instrument for years.

From what team size does it make sense?

From around 20–25 people, and from the point where there are several managers who can compare. Below that, the conversation can be had directly, without an instrument.

How often is it redone?

Once a year is enough for the positions; the actions are checked at six months. Redone quarterly it becomes bureaucracy, and people move between boxes from noise rather than real change.