A 9 box assessment is the act of placing a person in one of nine boxes by performance and potential. It sounds like scoring and it is not: there is no scale that produces the box, only a judgement two or three people have to agree on. That is why the same company can run the assessment twice in a year and get different distributions, and why the interesting output is never one person's box but the shape of the whole set. This page is about how the assessment is made in a small team that has one HR person and no assessment budget, and what the free 9-box grid worksheet on this site returns once the placements are in.
Who makes the assessment
In a small business the honest answer is the person's manager, with the HR lead in the room. That is fine, and it is better than the alternative most small teams reach for, which is to buy an assessment instrument and let it decide. The manager knows the delivery. What the manager does not have is a view across the company, which is why the placement made alone is nearly always generous on potential and inconsistent between departments. The second pair of eyes is not a formality; it is the whole control. If only one person can be in the room, make it the same person for every department so the inconsistency is at least uniform.
What potential means in a 9 box talent assessment
Potential in a 9 box talent assessment is not ambition, tenure or how much the manager likes the person. The workable definition for a small company is narrow: could this person hold a materially bigger role within a defined horizon, usually two or three years, given development the company can actually afford. Every clause matters. Bigger has to be defined, the horizon has to be stated, and affordable development rules out the answer that assumes a budget nobody has. A definition this tight produces fewer high potential placements than managers expect, which is the point.
Calibration is where the assessment becomes usable
Calibration is the meeting where placements from different managers are put side by side and moved. It is uncomfortable and it is the only step that makes the grid comparable across the company. Run it on the counts rather than the names first: if one department has placed a third of its people in the top row and another has placed none, the disagreement is about the standard, not about individuals, and it is faster to settle it that way. Then walk the individual placements that changed. The CIPD performance factsheet cited below treats calibration as part of the process rather than an optional extra, and for a company small enough to fit its managers in one room there is no reason to skip it.
What the counts say once the nine box assessment is done
Enter the nine counts in the worksheet and read three things. The star share tells you how much of the company is both delivering and ready to grow, and whether that is concentrated in one team. The high potential pool tells you whether succession has anyone to work with at all. The at risk share, the box that is neither performing nor projected to, tells you how much of the payroll is in a conversation nobody has had. None of those figures has a right answer this site would publish; they are your distribution, and their value is that next cycle you can see which way they moved.
Questions people ask about 9 box assessment
Is a 9 box assessment a test the employee takes?
No. It is a placement made by the person's manager and reviewed with at least one other person. There is no instrument the employee sits. Some companies feed assessment data into the potential judgement, but the box itself is always a decision somebody signs.
Should employees be told their box?
That is a policy choice this site does not make for you, and it has employment law implications worth checking with your own adviser. What matters for the arithmetic is that the placement is recorded honestly, because a grid softened to be shareable stops measuring anything.
How often should the assessment run?
Most small companies run it once or twice a year, usually alongside the performance cycle. More often than that and the placements do not have time to change; less often and the succession plan built on them goes stale before it is used.
What if a manager places everyone high?
That is the normal first result, and it is what calibration is for. Compare the counts by department before you discuss any individual. A department with a third of its people in the top row against another with none is a standards problem, and settling the standard moves more placements in ten minutes than arguing person by person does in an hour.