9 Box Grid

What the 9 box grid is

The 9 box grid, also written nine box grid, is a display tool that places employees’ names in nine boxes. The boxes are formed where three levels of performance in the current role cross three levels of the capacity each employee is judged to have for a larger role.

What the grid produces is a position, not a score. An employee does not leave the exercise carrying a new number. They leave it with a location, and the location raises a question: what should be done with this employee over the coming year? Taking the position as a verdict on the person uses the tool for something it was not designed to do.

The labels attached to the nine boxes vary between the versions in circulation, and the definition of the 9 box grid does not rest on any one set of them. What defines the tool is the crossing of the two axes, not the names written in the boxes.

The two axes of the 9 box grid are separate judgements

Performance in the current role and capacity for a broader role are two dimensions that are kept apart when people are assessed. The grid is not that principle; it is the principle drawn on paper. It obliges whoever uses it to give each dimension an independent assessment, because a box cannot be filled without two values.

The two dimensions differ in three respects, and losing sight of any one of them collapses them back into a single dimension:

• Time. Performance is a judgement on what has already happened, and it has evidence. Potential is an estimate of what has not yet happened, and it has only indications.
• Source. Performance is assessed from results measured against published performance standards. Potential is assessed from behaviour in new situations, and from an ambition that has to be asked about rather than assumed.
• Grounds for challenge. A performance rating can be challenged with evidence. An assessment of potential can only be challenged with a competing judgement. That is why documentation on the potential axis is more necessary, not less.

Converting a five level scale for the 9 box grid

An appraisal can rate performance on five levels, as a graphic rating scale numbered from 1 to 5 does, while the grid needs three. So there has to be a rule that turns five levels into three. That rule is a decision that changes the whole picture, but a policy can leave it unwritten.

Take 90 employees whose ratings fall across five levels as follows: 8 at the top level, 22 at the next, 44 in the middle, 12 below it and 4 at the bottom. The total is 90.

• First rule: the top two levels count as high performance. The top row then holds 8 plus 22, which is 30 employees; the middle row holds 44; and the bottom row holds 12 plus 4, which is 16.
• Second rule: only the top level counts as high performance. The top row then holds just 8; the middle row holds 22 plus 44, which is 66; and the bottom row holds 16.

The top row has moved from 30 to 8 without a single employee’s performance changing. That is a ratio of 3.75, close to fourfold, and it comes from a line in a policy rather than from any result at work.

The arithmetic shows that the rule changes the outcome. It does not show which of the two rules is the right one, and the definition of the 9 box grid does not favour either. Choosing the rule is a decision that the organisation writes into its policy before the grid is used for the first time.

The conversion rule and the 9 box grid box that decisions rest on

Suppose potential has been assessed on the other axis with this result: 12 high, 40 moderate and 38 lower. The box that is examined in succession meetings is the one where high performance meets high potential.

• Under the first rule, with 12 of the 90 rated high on potential, about 13.3%, the expected number in that box is 30 multiplied by 0.133, or about 4 employees.
• Under the second rule the expected number is 8 multiplied by 0.133, or about 1 employee.

The gap between four and one is a gap in a decision. An organisation that sees four names builds a plan on them, but an organisation that sees one name concludes that it has no successor. Both are looking at the same employees.

The multiplication above assumes that the two axes are independent. That assumption makes the illustration simpler; it does not describe the situation. Both assessments can come from the same manager, and then an employee rated high on performance can tend to be rated high on potential as well. We found no measurement of how strong that tilt is. Its possible presence is enough reason to take the box as the output of a procedure rather than as a picture of reality.

The figures in both examples are assumed, chosen to show the effect of the conversion rule on the count. We found no published measurement of the effect of the 9 box grid on the quality of promotion decisions on which to base a figure.

How the 9 box grid differs from forced ranking

The grid can be confused with forced ranking, and the difference lies in where the constraint falls. In forced ranking the constraint falls on the distribution, not on the judgement: the percentages are written before the results are seen, and the categories are then cut at those percentages whatever the results turn out to be.

The grid has no prescribed percentage at all. A box can be empty, and half the names can gather in a single box, without anything being breached. Whoever adds a percentage for each box has stopped using the grid. They are running a forced ranking on two dimensions under the grid’s name, and they then need the rounding rules and the decision on the unit of application that forced ranking requires.

The 9 box grid as nine boxes and three decisions

The grid invites someone to write an action for every box, which produces nine tracks that cannot be managed. A more useful approach takes the boxes in three groups:

• Where potential is high: the question is which role comes next, and when. This connects to succession planning and to career pathing, and it cannot proceed without an expected vacancy.
• Where performance is high and potential is lower: the question is how the employee is retained and recognised in the current position, not how they are moved.
• Where performance falls short of what is required: the question is not settled in the grid at all. Addressing underperformance has a separate process and a separate document, and the grid is a display tool, not a route for remedying performance.

The high performance boxes of the 9 box grid where potential is lower

The boxes that combine high performance with moderate or lower potential can be large, and they can hold people who do their work properly and who neither want a broader role nor suit one. Approaching those boxes as a problem in search of a remedy can damage the tool.

These are people on whom today’s work depends, and losing one of them is not made good by promoting someone with more potential. The sound use is to take these boxes as a question about retention and recognition in the current position, and to look at the role each person in them occupies. An employee in a critical role whose performance is high deserves more attention than one placed in a higher box whose position can be filled by someone else.

When a team is too small for the 9 box grid

The grid has nine boxes, so it divides the people placed in it into nine groups. An organisation of 12 employees spread across nine boxes averages one and a third names per box. Some boxes can then stay empty, and a filled box can hold a single name.

A box holding one name is not a classification. It is a named judgement on a particular person, recorded in a table. That reverses the nature of the tool: something designed to show the picture of a group becomes a document about an individual, one the individual does not know about and cannot answer, and one seen by everyone who attends the meeting.

So the grid suits settings where the numbers are large enough for a box to be a group. Below that, the two questions the tool serves are asked directly, in a conversation about each employee individually: how is this person performing, and how far could they progress? Those two questions do not need a chart to be put to twelve people.

What the 9 box grid is not

• Performance calibration. It is a procedure through which managers come to a common understanding of the rating scale. The grid is a picture displayed once the assessment has settled. The grid can be shown in a calibration meeting, but showing a tool in a meeting does not make it the meeting.
• Performance appraisal. The grid consumes the outcome of the appraisal and does not produce it. Whoever fills the performance axis with a fresh judgement made during the meeting has created a second assessment of the employee, one the employee does not know about and cannot challenge.
• A talent pool. That is a group of people assembled to fill future positions. The grid is a sorting tool from which such a group may be drawn, but the grid is not the group.

What undermines a 9 box grid

• A single rater. The potential axis rests on an estimate, and an estimate by one person about a direct report carries whatever rater bias that person brings. On this axis there is no evidence to correct it, as there is on the performance axis.
• Standing still. A grid filled in once and never revisited turns into a permanent classification of the people placed in it.
• Telling employees their box. The grid names nine categories instead of one, so announcing a box hands the employee a judgement about their future that rests on no evidence they could answer. What is published is the criteria and the path.
• Use without vacancies. A box full of names with no roles ahead of them creates an expectation with nothing to meet it, and that can lead high potential employees to leave the organisation.

In what we read of the Saudi Labor Law (نظام العمل), its Implementing Regulation (اللائحة التنفيذية) and the model work regulation (النموذج الموحّد), we found no mention of tools for classifying potential. So the 9 box grid cannot be described from those texts as required or as prohibited, and a decision that those documents govern under a written provision of their own is not to be based on the grid by itself.

Before a 9 box grid is filled in

The grid shows what the organisation has already decided; it decides nothing on the organisation’s behalf. If performance standards are published, and potential is assessed through described behaviour and through an ambition the employee states and stands behind, the grid gives a useful picture for a meeting of an hour. If either of those is missing, the grid shows impressions arranged in boxes. That is more dangerous than impressions left unarranged, because it looks like measurement.

The question to ask before the first session is this: if a name is placed in a box, and whoever placed it is asked six months later for the reason behind its position on the second axis, is there anything written down to point to? If there is not, what the tool lacks is its material, not its form.

This is an explanation of the concept, not legal advice.

Related terms

• Succession Planning
• Performance Calibration
• Peer Appraisal
• Performance Standards
• Graphic Rating Scale