9-box grid

What do the two axes mean?

Performance looks backwards and is the more defensible axis: what did this person deliver against what was agreed, over a defined period? If you cannot point to the objectives, you are rating a feeling.

Potential looks forwards, and is where the framework gets slippery. It should mean something specific and testable - the likelihood of succeeding in a materially bigger or broader role within a defined window, say two years. It should not mean how much you like working with somebody, how ambitious they seem, or how much they resemble the people already senior.

Is "low potential" a bad thing?

No, and reading it that way is how the grid earns its bad reputation. Low potential means "unlikely to move up two levels in the near term", which describes the majority of any workforce - including people who are excellent, deeply expert, and completely irreplaceable in the job they do. The bottom-right box holds some of the most valuable people in most organisations.

If your grid treats "not going to be a director" as a failure state, the grid is wrong, not the people.

When should the 9-box grid not be used?

The first exclusion matters most. Redundancy selection criteria need to be objective, measurable and capable of being evidenced - and a subjective forward-looking judgement about "potential" is close to the opposite of that. Using a 9-box position to decide who goes hands a tribunal both an unfairness argument and a discrimination argument in one move. Selection criteria belong in a scored matrix built for the purpose, applied to a properly defined pool: see the redundancy process guide, and the redundancy cost calculator to cost it.

Where does bias get in?

Through the potential axis, almost always. Performance can be argued about with evidence; potential is a prediction, and predictions drift toward familiarity - people who communicate the way the rater does, who are visible rather than effective, who are available for extra work because of circumstances outside the job. Part-time and flexible workers are routinely marked down on potential for reasons that have nothing to do with capability.

Three defences, in order of how much they help:

• Calibrate with more than one manager. A rating no-one else has to hear justified out loud is not a rating, it is an opinion. This is the highest-value hour in the whole exercise.
• Check the distribution against your demographics. If the top row looks nothing like the workforce, that is a finding about the process, not about the people.
• Date it and expire it. A grid position from eighteen months ago is a memory. Redo it or drop it.

Should you tell people their box?

This is the question that decides whether the exercise is talent planning or a secret ranking - and there is no version of "secret ranking" that ends well. It leaks, and what leaks is always the label rather than the reasoning.

The workable middle: share the outcome, not the coordinates. Nobody benefits from being told they are a 2-2. Everybody benefits from "we think the next step for you is leading the migration project, and here is why" or "we would like you to go deeper rather than broader this year, and that is a genuine choice, not a consolation." If you cannot describe someone's box as a constructive next step, you have not finished the exercise.

Keep the underlying conversation going between reviews with regular 1:1s - an annual grid built from one manager's twelve-month recollection is far weaker than one built from notes taken as things happened.

A simpler alternative

If nine boxes feels like more machinery than your team needs - and under about thirty people it usually is - the useful signal comes out of two questions per person, asked honestly at the same time each year: is this person growing? and what would we do if they resigned tomorrow? That surfaces the same succession risks and development gaps without asking managers to score their colleagues on a nine-point scale they will then feel obliged to defend.