On this page
The 9 Box Grid: A Complete Guide
By 9Box · Updated 2026-05-16
The 9 box grid is a talent management tool that plots employees on two axes, performance and potential, to create nine categories ranging from "underperformer" to "future leader." It's been used for succession planning, promotion decisions, and development conversations for over fifty years.
It's also, in the way most companies use it, broken. One manager filling in a grid alone produces a grid that mostly reflects that manager. This guide explains the model properly, shows you what each box actually means, walks through how to run a talent review with it, and is honest about where the traditional approach falls down.
If you just want the template, download it here. If you want to skip the theory, jump to how to run a 9 box assessment.
What is the 9 box grid?
The 9 box grid (also called the nine box grid, 9 box matrix, or performance-potential matrix) is a 3×3 framework. The horizontal axis is performance, how well someone is doing in their current role. The vertical axis is potential, how much capacity they have to take on more, grow, or move into different roles.
Each axis has three levels: low, moderate, and high. Cross them and you get nine boxes, each representing a different talent archetype.
It was developed by McKinsey in the 1970s for General Electric, who needed a way to allocate investment across 150 business units. The same logic, comparing current value against future value, got applied to people. Fifty years later, it's still the most widely used succession planning tool in HR.
The nine boxes explained
Different sources use slightly different labels, but the underlying logic is the same. Here's what each box means, working from bottom-left (low/low) to top-right (high/high).
Low performance, low potential, "Underperformer"
Not delivering in the current role, and no clear signal they'd do better elsewhere. Action: structured performance plan, role change, or, if the gap persists, exit. This is the box people are most reluctant to use, which is exactly why it often gets misused.
Moderate performance, low potential, "Effective"
Doing the job adequately, unlikely to grow much beyond it. Often dismissed unfairly. These are frequently your most stable, longest-tenured employees. Action: retain, recognise, don't push them somewhere they don't want to go.
High performance, low potential, "Specialist" or "Trusted Professional"
Excellent at the current role, no appetite or aptitude for a broader one. Action: keep them happy, pay them well, give them mastery work. Forcing them into management is how you lose them.
Low performance, moderate potential, "Inconsistent"
Shows flashes but isn't delivering consistently. Often a role-fit or motivation issue, not a capability one. Action: have the honest conversation. Wrong role? Wrong manager? Personal context?
Moderate performance, moderate potential, "Core Player"
The largest group in most organisations, and the one most talent reviews ignore. Action: targeted development, stretch projects, don't take them for granted.
High performance, moderate potential, "High Performer"
Crushing the current role, could probably take on more but not a clear future executive. Action: lateral moves, broader scope, leadership of small teams.
Low performance, high potential, "Potential Gem" or "Diamond in the Rough"
Rare and controversial. Some talent experts argue this box shouldn't even exist, if someone's consistently underperforming, by what evidence are we calling them high-potential? In practice it's used for new hires, recent role changes, or people working under bad conditions. Action: diagnose the performance gap before investing.
Moderate performance, high potential, "High Potential" or "Growth Employee"
Solid current performance, real headroom. Action: development plan, mentor, prepare for next role.
High performance, high potential, "Star" or "Future Leader"
The succession planning prize. Action: visibility to senior leadership, stretch assignments, retention focus. These people get poached.
Why the traditional 9 box grid fails most of the time
Most criticism of the 9 box grid isn't really criticism of the model, it's criticism of how it's used. The model is a structure for a conversation. The failure modes are about who has the conversation and what data they use.
One manager, alone, from memory. The most common implementation. A manager sits down once a year and places their reports in boxes based on recent vibes. Recency bias, halo effect, similarity bias, and whoever spoke up most in the last team meeting all do most of the work.
Performance and potential aren't defined. What does "high potential" mean? Two managers in the same room will mean different things. Without a written definition specific to your organisation, the grid is a Rorschach test.
Potential gets conflated with performance. "Potential" should be evidence of capacity for different or broader work, not just doing the current job well. In practice, most managers rate someone high-potential because they're a high performer. The two axes collapse into one.
It becomes a labelling exercise. Someone gets placed in "moderate performance" once and is treated that way forever. The grid is a snapshot, not a sentence. Most companies don't refresh it often enough for that to be clear.
Nobody tells the employee. The grid is filled in behind closed doors. The person it's about either never finds out or finds out badly. Either way, the development conversation that's supposed to follow doesn't happen.
The grid isn't the problem. The single-rater, undefined, one-way, once-a-year process is the problem.
How to run a 9 box assessment properly
A 9 box review that actually generates good decisions has six steps.
1. Define performance and potential in writing. Before anyone rates anyone, agree what each axis means at your company. For performance: results, behaviours, or both? Over what time period, a single year produces noise, three years is better. For potential: capacity to do what, your current next-level roles, or general growth?
2. Use sustained data, not memory. Pull together objective signals: review history, goal completion, peer feedback, 360 results, project outcomes. The grid should be the output of evidence, not a substitute for it.
3. Get more than one rater. A single manager's view is a single manager's view. The best implementations combine self-assessment (so the employee's own view is in the room), manager rating, and peer feedback. This is the core of how 9Box works: the subject scores themselves first, then peers score them on the same dimensions. The gap between the two is often where the real insight lives.
4. Run calibration. Once managers have placed their reports, get them in a room with peers and HR. Walk through each person. Challenge placements. Force consistency, your high performer and her high performer can't both be high performers if her one delivered twice as much.
5. Decide an action for every box. A grid with no consequences is a wall decoration. Every placement should produce a development action, even (especially) for the middle boxes everyone forgets.
6. Tell people. Not necessarily their box label, that can be reductive, but the underlying assessment and the development plan that comes from it. If your 9 box exercise is secret, your employees will assume the worst version of what it says.
9 box grid examples
Example 1: An engineering team of 12
A team lead places six engineers as "Core Players" (moderate/moderate), two as "High Performers" (high/moderate), one as a "Star" (high/high), two as "Inconsistent" (low/moderate), and one as a "Specialist" (high/low).
The calibration session pushes back on the Star placement, the engineer in question is the loudest in standups but their delivery metrics are merely good, not exceptional. They move to High Performer. One of the "Inconsistent" engineers turns out to have been onboarding for three months; their placement is suspended pending another quarter of data. The Specialist, a senior IC who's repeatedly declined management tracks, gets a retention conversation and a path to staff engineer rather than being squeezed toward a lead role.
Result: three placements changed, two real decisions made, one bad assumption surfaced.
Example 2: A sales team after a strong quarter
Every rep had a great Q4. The manager's instinct is to place most of the team as High Performers. Calibration forces the question: against what baseline? Three reps inherited the largest territories. Two were ramping during slower quarters and now look stellar partly because of timing. One was genuinely outperforming on a tough patch.
After calibration, only one rep moves to High Performer. The others stay at Core Player or Effective, with a note to reassess after Q2. The "Star" of the team, a senior rep with consistent multi-year performance, is flagged for the succession plan, not because of the recent quarter but because of the pattern across three years.
Result: short-term noise filtered out, real signal preserved.
Example 3: A 30-person mixed team in a scale-up
Leadership wants to identify successors for two director roles opening up next year. They place all 30 people on the grid. The result: four people in the top-right quadrant (high potential), only one of whom is currently performing at the level needed for director.
The conversation shifts. Instead of "who's ready now?" it becomes "which of the four do we invest in over the next 12 months, with what stretch assignments?" Two get cross-functional projects. One gets a coaching engagement. The fourth, on reflection, doesn't actually want the director role, they want to lead a specialism. The grid surfaced that they'd been promoted toward something they didn't want.
Result: a real succession plan, not a list of names.
When not to use the 9 box grid
The grid is wrong for some situations.
Teams under 10 people. You don't need a grid for eight people; you need a conversation.
Performance management for underperformers. If someone's struggling, they need a structured improvement plan and weekly 1:1s, not a label on a chart.
Fast-changing roles. If the work itself is shifting quarterly, last year's "high performance" rating means nothing about this quarter.
As a substitute for feedback. The grid summarises feedback; it doesn't replace it.
9 box grid alternatives
If the 9 box doesn't fit, the common alternatives are:
Talent reviews without a grid, just structured conversations about each person, no matrix. Works well for small teams where calibration is the point and the grid is overhead.
Skills matrices, plot people against required competencies, not abstract "potential." More useful when you have a clear competency framework and want to identify capability gaps.
Continuous performance management, ongoing feedback rather than annual placement. Works alongside the grid rather than replacing it; the grid becomes the annual summary of continuous data.
Competency-based frameworks, assess against defined behaviours, not against each other. Less comparative, more developmental.
Each has trade-offs. The grid's strength is that it forces comparison and surfaces gaps. Drop it only if you have something that does the same job better.
Free 9 box grid template
We've put together a free 9 box grid template in Excel and Google Sheets, the boxes, the axes, space for evidence, and an action column for each placement.
Or, if you want to run the whole review online, with self-scores, peer scores, and the report generated for you, 9Box is free while in beta. See also worked examples, how to use the 9 box grid, and 9 box vs performance review.
FAQs
What does each box in the 9 box grid mean?
Each of the nine boxes represents a combination of performance (low, moderate, high) and potential (low, moderate, high). The labels vary by source but typically run from "Underperformer" (low/low) at the bottom-left to "Star" or "Future Leader" (high/high) at the top-right. See the section above for what each box means and the typical action for each.
Who invented the 9 box grid?
McKinsey developed the 9 box grid in the 1970s for General Electric, originally to help allocate investment across GE's 150 business units. The same logic was then applied to people, and the tool became a standard for succession planning.
Is the 9 box grid still used?
Yes. It's still one of the most common succession planning tools in HR, though increasingly combined with 360 feedback and continuous performance data rather than used alone. Most large organisations still run some version of it; many smaller and more modern teams have moved to lighter-weight alternatives.
What's the difference between performance and potential?
Performance is what someone is delivering in their current role, ideally measured over multiple years. Potential is their capacity to take on different or broader work in the future. The two are often conflated, managers tend to rate high performers as high potential by default, but they're meant to be independent. Someone can perform brilliantly in their current role with no interest in or aptitude for a different one.
How often should you do a 9 box review?
Annually at minimum, semi-annually for fast-growing teams or teams going through significant change. Less frequently than that and the grid becomes a fossil; more frequently and the calibration overhead outweighs the benefit.
Can employees see their own 9 box rating?
Practice varies. Many organisations keep placements confidential to avoid demotivation, particularly for those in lower boxes. The case for transparency is that hidden placements lead to development conversations that don't happen, the employee never knows what the company actually thinks of their trajectory. The middle path is to share the underlying assessment and development plan without sharing the specific box label.
Is the 9 box grid the same as a talent matrix?
Yes. 'Talent matrix,' 'performance-potential matrix,' 'nine box,' and '9 box grid' all refer to the same tool. Some companies use slightly different axis labels (for example, 'contribution' instead of 'performance,' or 'readiness' instead of 'potential'), but the structure is identical.