The nine-box grid is one of the most used talent management tools, and one of the worst fed.
The structure is simple. One axis measures performance, the other measures potential, and each person lands in one of nine quadrants. Succession, development, and investment decisions come out of it.
The problem is not the grid. It is that one of the two axes almost never has data behind it.
The two axes are not equal
Performance is measured. Targets hit, reviews from recent cycles, observable results. It can be noisy, but it exists.
Potential is an opinion. At most companies it gets filled in with the direct manager’s impression of how far the person can go. No instrument, no written definition, and no comparability across teams.
Crossing a data point with an impression and treating the output as a decision grid is why many talent reviews end up confirming what the committee already believed.
The three biases that fill the potential axis
Performance disguised as potential. The most frequent error and the most expensive. Someone excellent in their current job gets marked high potential, and ends up promoted into a role that demands something else. It is the classic mechanic of promotion to the level of incompetence, and it starts in this box.
Visibility disguised as potential. Whoever presents to the committee, travels, or works near leadership gets seen more. Being seen more is not being more.
Similarity disguised as potential. The evaluating leader recognizes their own trajectory in a candidate and reads it as a signal. It is the best documented bias in promotion decisions.
None of the three gets corrected by asking the evaluator to be objective.
What potential actually means
Before measuring it you have to define it. And the useful definition is not “how far they can go,” which is not measurable.
Potential is the probability of performing well in a role that is different and more complex than the current one. Different matters. It almost always means more ambiguity, more people, and less direct control over the outcome.
Framed that way, the question changes. It is not “how well do they do this today?” but “what does this person have that works in a context they have not lived yet?”
Measuring the part that can be measured
Stable personality traits predict performance in roles a person has not held. That is exactly the question the potential axis asks, and it is where an instrument contributes something performance reviews cannot.
Three Five Factor dimensions have known relationships with performance in higher-complexity roles. Openness connects to operating in unfamiliar contexts. Emotional stability sustains output under prolonged pressure, which is what changes as levels go up. Conscientiousness predicts performance across essentially every occupation studied.
That does not turn potential into an exact number. It turns one impression into two data points you can set side by side.
How the exercise changes
A talent review with two measured axes runs differently.
The committee arrives with performance from prior cycles and with a profile measured using the same instrument for everyone in the exercise. Each leader still brings their read, and that read still counts. What changes is that there is now something to check it against.
The interesting cases are the disagreements. Someone the manager marks low and the profile places high is usually someone with little visibility. Someone the manager marks high and the profile does not support is usually excellent performance mistaken for potential. Both conversations are more useful than any box.
What to do afterward
The grid is not the deliverable. The deliverable is what happens in each quadrant.
High performance with high potential calls for a concrete succession plan, with an identified seat. High performance with medium potential calls for retention and technical depth, not a promotion into management nobody asked for. Low performance with high potential is usually a fit problem with the current role, and there it pays to look at the job profile before looking at the person.
How the second read gets built is covered in potential mapping, and how it becomes a plan in succession planning with data.
Related OCEAN+ profiles
Discover which personality dimensions to look for in each role.
Related Articles
The Job Profile: How to Define It With Data, Not a Template
The job description you publish and the profile you decide against are two different documents. Most companies have the first and none of the second.
Performance Reviews: The Methods and How to Choose
Rating scales, 360s, OKRs, forced ranking. Each answers a different question, and picking wrong is the most common reason the whole process changes nothing.
Competency-Based Interviews: The Guide That Changes the Decision
The structured competency interview is one of the few formats with real validity evidence. Most companies implement half of it and lose the benefit.