9-Box Talent Review: How Many Performance Rating Levels Should You Use?

Key Takeaways

• Pave data shows that as companies grow, they adopt more granular performance rating systems: four-rating systems are most common at smaller companies, while five and six-or-more rating systems dominate at organizations with 1,000+ employees.
• The 9-box talent review combines performance and potential into a single framework, making it more useful for succession planning and talent development than a standard performance rating alone.
• The right rating scale depends on what you're using the ratings for. If ratings drive merit decisions, scale granularity directly affects pay differentiation. If ratings feed a 9-box, the performance axis needs enough precision to place employees meaningfully.
• There is no universally correct number of rating levels. The best system is the one your managers can use consistently, and your comp program can actually act on.

‍

The number of performance rating levels a company uses shapes everything downstream: how merit budgets get allocated, how managers calibrate their teams, and whether a 9-box talent review produces meaningful distinctions or a muddled middle.

Most HR and compensation leaders inherit a rating scale rather than design one deliberately, which means the scale they're using may not match the program it's supposed to support.

Pave's data on rating-scale adoption across company sizes, combined with the tradeoffs of each system, gives compensation and HR leaders a more grounded starting point for that design decision than anecdote or convention.

What Is a 9-Box Talent Review and How Does It Work?

A 9-box talent review (also called a 9-box matrix, 9-box talent grid, or 9-box talent matrix) is a talent management framework that maps employees on a three-by-three grid across two axes: current performance (low, medium, high) and future potential (low, medium, high).

The combination produces nine possible placements, each carrying different implications for development, succession planning, and compensation. You may also see it labeled a 9-box performance review, a 9-box talent assessment, or simply a 9-box assessment; the names differ but the mechanics are the same.

The framework is most commonly used during annual talent reviews for workforce planning, succession discussions, and identifying high-potential employees.

Where a standard performance rating answers "How is this person doing now?" the 9-box adds a second question: "Where could this person go?" Answering both requires two separate judgments that need to be calibrated consistently across managers, which is where most implementations break down.

How Do You Place Employees on the 9-Box Grid?

Performance is typically anchored to observable outcomes: goal attainment, quality of work, and how an employee operates relative to the expectations of their role and level. Most organizations use their existing performance rating as the performance axis input, which is why rating-scale design and 9-box design are connected decisions.

Potential is harder to define and more prone to bias. The most defensible potential assessments focus on specific, observable indicators rather than vague impressions. Common criteria include learning agility (how quickly someone picks up new skills), leadership behaviors (how effectively they influence peers and develop others), and scope readiness (evidence that they're operating above their current level).

The practical challenge is calibration. Without explicit definitions and a structured calibration process, the 9-box reflects manager perception rather than a useful talent map.

How Many Performance Rating Levels Should You Use?

Pave analyzed rating-scale adoption across 200+ companies running merit cycles in its Compensation Planning tool and found a consistent pattern: as companies scale, they move toward more granular rating systems.

Among 200+ companies running merit cycles in Pave's Compensation Planning tool, rating-scale distribution shifts clearly by size.

At 1–200 employees, four-rating systems lead at 44%, with three-rating systems (the category that includes the 9-box, since its performance axis has three levels) and five-rating systems at 24% each.

Four ratings stay in front at 200–499 employees at 46%, then five-rating systems take the lead at 500–999 employees (32%) and 1,000–2,999 employees (54%).

At 3,000+ employees, six-or-more rating systems become the most common choice at 54%. The main trend: as companies mature, they adopt more granular performance rating systems.

The sharpest shift comes past 1,000 employees, where five and six-or-more rating systems together account for 85% of companies at 1,000–2,999 and 92% at 3,000+, while three-rating systems hold at just 8%.

That reflects a practical reality: larger organizations need finer distinctions to allocate merit budgets across more employees, levels, and functions.

The Case for Each Rating System

Three Ratings

Force clear distinctions and direct conversations: below expectations, meets expectations, exceeds expectations. The limitation is compression.

In a merit cycle, "meets expectations" can describe the majority of your workforce, leaving little room to reward within that band. Only 8% of companies with 1,000+ employees still use a three-rating system, reflecting how quickly that compression becomes a problem at scale.

Three levels are also exactly what a 9-box uses for its performance axis, which is why companies running a 9-box sit inside this category in Pave's data.

Four Ratings

Eliminate the neutral middle option. That's the main argument for this scale: it forces managers to decide whether an employee is performing above or below expectations rather than defaulting to the center. It adds differentiation over three ratings without significantly increasing calibration complexity.

Five Ratings

The leading choice at companies between 500 and 3,000 employees for good reason. They offer enough granularity to distinguish meaningfully between performance levels, give merit matrices room to differentiate increases across a realistic spread of performance, and can be banded down to the three performance levels a 9-box uses when talent reviews call for it.

The tradeoff is that calibrating five levels consistently across managers requires more effort and clearer definitions.

Six or More Ratings

Unlock the precision needed for organizations with highly specialized roles or mature pay-for-performance programs where small differences in rating translate to meaningful compensation differences. The risk is that beyond five levels, rating distinctions can feel arbitrary to managers and employees, which erodes trust in the system.

No Ratings

A deliberate choice some organizations make, particularly large enterprises outside of tech, and some pre-IPO companies. In these systems, managers typically have more discretion during merit cycles rather than operating from a rating-driven matrix.

This approach can reduce the gaming and recency bias that formal ratings introduce, but it places more weight on manager judgment and makes it harder to run a consistent 9-box process.

How the 9-Box Connects Performance Ratings to Compensation Decisions

For compensation teams, the connection works like this: an employee's performance rating feeds the performance axis of the 9-box. Their position in the grid then informs not just their merit increase but also whether they're a candidate for an equity refresh, an off-cycle promotion, or accelerated development investment.

Pave's Merit Cycle State of the Union 2026 report shows how leading companies are structuring those connections between performance outcomes and compensation decisions in practice.

When talent reviews and merit cycles run as parallel processes without a shared framework, managers make promotion recommendations in one and merit recommendations in the other without connecting the two. That inconsistency shows up as unexplained pay variation and erodes employee trust.

If you're using a 9-box and a separate performance rating, the most important design question is whether the rating scale you use produces enough differentiation to place people meaningfully on the grid. A three-rating system may not give managers enough information to distinguish the nine cells. A five-point rating system typically does.

Design a Rating System Your Comp Program Can Act On

The rating scale you choose shapes how merit budgets get allocated, how the 9-box performs as a talent tool, and whether managers can calibrate consistently enough for the output to be trusted.

Pave's Compensation Planning product connects performance ratings directly to merit cycle workflows, so the link between rating and pay decision is explicit, auditable, and consistent across managers.

Book a demo with Pave to see how the AI compensation platform helps you design and run a merit cycle that reflects how your organization actually thinks about performance.