A calibration meeting can fail before the first rating is discussed. If managers arrive with different definitions of strong performance, incomplete evidence, or a preference to defend their own team, the outcome becomes a negotiation rather than an assessment.
A well-designed calibration meeting agenda template creates the controls that conventional review meetings often lack. It gives leaders a common standard, separates evidence from opinion, and records why a rating changed or stayed the same. For HR teams, that means a more defensible process. For managers, it means less time spent debating impressions and more time making clear decisions.
What a calibration meeting agenda needs to achieve
The purpose of calibration is not to force every department into the same rating distribution. It is to test whether comparable performance is being assessed consistently across roles, teams and managers.
That distinction matters. A high-performing specialist may not look like a high-performing people leader, and a team working through a difficult operating period may have different delivery conditions from one that had stable priorities. Calibration should account for context without allowing context to become an excuse for vague or uneven standards.
A practical agenda therefore needs to do four things: establish the assessment criteria, review evidence before advocacy, identify material rating gaps, and capture decisions with clear ownership. The meeting should also distinguish between a rating issue and a performance management issue. Not every difficult case needs to be resolved in the room, but every unresolved case needs a next step.
Calibration meeting agenda template
Use the following structure for a 60 to 90-minute meeting. For larger groups, run separate sessions by function or job family, then hold a cross-functional moderation session for roles with comparable scope.
1. Open the session and confirm decision rules - 5 minutes
Start with the meeting owner, usually HR or the relevant people leader, confirming the purpose of the session. State the rating scale, the role level or employee group under review, and the evidence threshold expected for any proposed change.
A useful opening script is: “We are testing consistency against agreed role expectations. We will discuss observable outcomes, demonstrated behaviours and documented feedback. We will not change a rating solely because a manager is more persuasive, a team has historically rated higher, or a distribution looks uncomfortable.”
Confirm who can make final decisions and who is responsible for recording them. If senior leaders are present, make clear whether they are moderators or decision-makers. Ambiguity at this point can turn a structured discussion into a hierarchy-driven one.
2. Reconfirm role expectations and rating definitions - 10 minutes
Review the definition of each rating before individual employees are discussed. Keep this practical. A rating of exceeds expectations should be tied to sustained delivery beyond the core requirements of the role, not simply effort, long hours or one high-profile project.
Where possible, use the actual position description and agreed responsibilities. This is where AI-assisted role analysis can be useful: it turns broad job descriptions into assessable tasks and focus areas, so managers are comparing performance against what the role requires rather than against an informal personal benchmark.
Ask the group to flag roles with materially different scope, recent changes in responsibilities, or missing objectives. These do not automatically invalidate a review, but they should be visible before ratings are compared.
3. Review the calibration data set - 10 minutes
Present the employee list, proposed ratings, role levels, tenure, and any relevant performance evidence in a consistent format. Avoid opening with a forced distribution chart. A distribution can reveal patterns, but it should not dictate the result.
Look for signals that require discussion: one manager rating nearly every team member at the top level, similar roles receiving very different ratings, or a sharp gap between self-assessment and manager assessment. These are prompts for investigation, not proof that a rating is wrong.
Blind submission controls strengthen this stage. When employee and manager assessments remain locked until both are submitted, neither party can shape their view around the other’s answer. CALIBRATE uses this side-by-side comparison model to surface genuine gaps without allowing premature influence to distort the record.
4. Discuss material outliers and rating gaps - 25 to 40 minutes
Work through the cases that need calibration, starting with the largest inconsistencies. The manager should give a concise evidence-based rationale, followed by questions from peers and the moderator. Keep each discussion anchored to the role, the review period and the published rating definitions.
Use questions such as: What specific outcomes support this rating? Which responsibilities were performed beyond, at, or below the expected level? Is the result sustained across the review period? Would we apply the same standard to a comparable employee in another team?
This is not an invitation to collect every anecdote. One isolated incident should not outweigh a year of delivery, and popularity should not substitute for evidence. Conversely, qualitative feedback can be relevant when it identifies a repeated pattern with clear impact on customers, colleagues or execution.
If there is a disagreement, identify its source. It may be a difference in available evidence, a misunderstanding of role scope, or an inconsistent interpretation of the rating scale. Each requires a different response. More evidence may resolve the first; clarification of expectations may resolve the second; and a rating decision may be needed for the third.
5. Confirm rating decisions and development actions - 10 to 15 minutes
For every case discussed, record one of three outcomes: rating confirmed, rating adjusted, or decision deferred pending evidence. Document the rationale in plain language. “Adjusted after cross-team comparison” is not sufficient. Record which evidence or role expectation changed the group’s view.
Separate the final rating from the development conversation. An employee can be meeting expectations while still needing to build a skill that will matter in the next role. Likewise, an employee may have delivered strong results while requiring a specific behavioural improvement. Combining these points into a single vague statement reduces the value of both.
For deferred decisions, set an owner, required evidence and a deadline. Open decisions create inconsistency when they are quietly resolved later without the same scrutiny.
6. Check for process bias and close - 5 minutes
Before ending, review the session itself. Ask whether certain managers had stronger representation, whether comparable roles were considered, and whether the group relied on evidence consistently. This quick check helps HR identify where manager training, clearer role definitions or stronger review controls are required before the next cycle.
Close by confirming how decisions will be communicated, when records will be finalised, and what should remain confidential. Calibration outcomes should never be discussed casually outside the authorised review process.
How to prepare managers before the meeting
The quality of the meeting depends on work completed beforehand. Send managers a short preparation brief that asks them to review role responsibilities, proposed ratings, objectives, feedback and relevant delivery data. Require them to prepare a rationale for any rating at either end of the scale.
It also helps to set a rule that managers cannot introduce new claims in the meeting without supporting evidence. This does not prevent useful context from emerging, but it limits last-minute arguments based on recency bias or informal reputation.
HR should prepare a pre-read that identifies likely outliers while withholding unnecessary personal detail. Include the applicable rating definitions, the decision rules, a list of attendees and the process for escalating concerns. The aim is not to pre-decide outcomes. It is to ensure the meeting time is spent on judgement, not administration.
Common agenda mistakes to avoid
The most common mistake is treating calibration as a ratings correction exercise after all meaningful review work has finished. When criteria are unclear from the start, a final meeting cannot fully repair the problem. Calibration is strongest when role expectations, employee self-assessments and manager assessments all use the same structured foundation.
Another mistake is overcorrecting for rating distribution. A team may genuinely have several strong performers, particularly after a period of hiring, restructuring or major delivery. Challenge the evidence, not the appearance of the curve.
Finally, do not let the most senior voice become the default answer. Senior judgement is valuable, but it should be applied through the same evidence standard expected of every manager. A meeting that appears consistent but cannot explain its decisions will be difficult to defend with employees or leadership.
A calibration agenda is a control point, not a calendar placeholder. When every rating is tied to actual role requirements, blind assessment inputs and recorded reasoning, managers leave with more than a final number. They leave with a clearer standard for the next performance conversation.