Here is a question that exposes a problem in most organizations: if you asked five managers to rate the same employee's performance on a 1 to 5 scale, would they all pick the same number?
In most companies, the answer is no. Not even close.
The issue is not the scale itself. A 1 to 5 system works well when it is properly defined and calibrated. The issue is that most organizations hand managers a number range with no clear definitions and expect consistency. The result is that one manager's 3 is another manager's 4, and the entire system loses credibility.
Defining Each Level
The most important thing you can do with a rating scale is write clear, behavioral definitions for every level. Here is a framework that works.
1 - Does Not Meet Expectations
Performance is significantly below what the role requires. The employee is missing deadlines, producing work that requires frequent correction, or failing to meet basic responsibilities. This rating should include a documented performance improvement plan with specific milestones and a timeline of 30 to 60 days.
2 - Partially Meets Expectations
Performance falls short in one or more key areas. The employee handles some responsibilities well but is inconsistent or struggles with important parts of the role. Development support is needed, and the manager should set clear short-term goals with weekly check-ins.
3 - Meets Expectations
The employee is performing at the level expected for their role and experience. They consistently deliver quality work on time, meet their goals, and are a reliable contributor to the team. This is a solid, respectable rating. A team where most people earn a 3 is a healthy, well-functioning team.
4 - Exceeds Expectations
Performance is clearly above what the role requires. The employee consistently delivers beyond their goals, takes on additional responsibility, and positively influences the people around them. They are ready for stretch assignments or expanded scope.
5 - Significantly Exceeds Expectations
Performance is exceptional and rare. The employee has delivered results that materially impacted the team or company in ways that go well beyond their role. A 5 should represent the top 5 to 10 percent of performers and should be supported by specific, measurable evidence.
The Problem With Grade Inflation
The biggest threat to a rating scale is grade inflation. In most organizations, the average rating drifts toward 4 over time. Managers are uncomfortable giving 3s because employees interpret "meets expectations" as negative. The entire scale compresses into 3.5 to 5, which makes it nearly useless for distinguishing performance.
There are three ways to fight this.
Normalize the 3. Explicitly and repeatedly communicate that a 3 is a good rating. It means the employee is doing their job well. Use team meetings, manager training, and review kickoff communications to reinforce this message. When managers and employees both understand that "meets expectations" is a positive outcome, the resistance to giving (and receiving) a 3 drops significantly.
Require justification for 4s and 5s. Any rating above a 3 should be supported by specific examples that demonstrate performance beyond what the role requires. If a manager cannot point to concrete evidence, the rating should come down. This is not about being stingy with high ratings. It is about making sure those ratings mean something.
Calibrate across teams. Calibration sessions are where managers present their ratings and reasoning to a group of peers or senior leaders. The goal is to ensure that a 4 on Team A represents the same level of performance as a 4 on Team B. Without calibration, rating inflation happens unevenly, and employees on stricter teams are disadvantaged during promotion and compensation discussions.
Calibration in Practice
A calibration session typically works like this:
- Each manager presents their ratings and a brief justification for anyone rated 1, 2, 4, or 5
- Other managers ask questions and challenge ratings that seem inconsistent
- The group reaches consensus on borderline cases
- Ratings are adjusted as needed before being finalized
These sessions take time, usually one to two hours per department, but they are the single most effective tool for maintaining rating integrity. Companies that skip calibration almost always end up with a meaningless scale within two or three review cycles.
Common Pitfalls
Using half-points or decimals. Resist the temptation to add 3.5 or 4.2 ratings. This defeats the purpose of having defined levels and introduces subjective micro-distinctions that cannot be consistently applied.
Rating the person instead of the performance. A well-liked employee who underdelivers should not get a higher rating because of their personality. Ratings should reflect outcomes and behaviors against the role's expectations, not likability.
Averaging across competencies without weighting. If you rate five competencies and average them, a critical weakness in a core skill can be hidden by high scores in less important areas. Weight competencies based on their relevance to the role, or provide both individual competency ratings and a separate overall rating with its own definition.
Changing the scale mid-cycle. If you need to redefine your scale, do it between review periods, never during one. Changing definitions while reviews are in progress creates confusion and inconsistency.
Making the Scale Work
A rating scale is a tool. Like any tool, it works well when it is maintained and used correctly, and it fails when it is neglected. Define every level with clear behavioral descriptions. Train managers on how to apply those definitions. Run calibration sessions every review cycle. Communicate to employees what each rating means.
Platforms like Culture Wheel build these safeguards into the review process with guided rating frameworks and built-in calibration tools, so managers do not have to rely on memory or guesswork when assigning ratings.
When everyone in the organization shares the same understanding of what a 1 through 5 means, the performance review becomes a tool people actually trust.