A calibration call starts with a sensible plan: five managers will review ratings before the appraisal deadline. Thirty minutes later, two managers have retold their entire appraisals, the most disputed case has not been discussed, and the host is asking for agreement with three minutes left.
That ending is risky because calibration is a comparison process, not a sequence of presentations.
Put Evidence Comparison at the Center
UC Davis describes calibration as a process after managers draft appraisals and before they share them with employees. The purpose is to compare proposed ratings against common criteria and adjust where the evidence warrants it. That makes the live meeting valuable when managers need to test different interpretations of the same standards.
Use live calibration time for exceptions, borderline cases, and conflicting evidence, not for narrating material everyone could read before the call.
Equal airtime does not create a fair process. A manager with a straightforward, well-supported rating may need one minute. A disputed case may need careful questioning, a look at documented behavior, and a route for obtaining missing information. The agenda should make room for the latter.
Build an Evidence Queue Before the Meeting
Ask managers to provide proposed ratings and concise evidence in advance. The UC Davis manager checklist calls for concrete behavioral examples and data-based evidence, and asks presenters to keep their contribution short so discussion has room.
Create three queues:
- Confirm: cases with aligned rating and clear evidence. Review asynchronously or in a fast consent pass.
- Compare: cases with a borderline rating, conflicting examples, or disagreement about the standard. These get live discussion.
- Complete: cases missing evidence or a required reviewer. These get an owner and a follow-up date, not a rushed rating.
The queues stop a host from discovering the meeting's hard cases at the end. They also make the deferral route explicit. Deferral is not avoidance when it records what evidence is missing, who will obtain it, and when the group will reconvene.
For example, if one manager describes sustained delivery and another describes repeated missed commitments, the facilitator should ask for the review-period evidence behind both accounts. If the record is incomplete, mark the case complete only after the missing material is obtained. A hurried vote would hide the inconsistency calibration is meant to surface.
Use a 60-Minute Agenda That Protects the Close
| Time | Segment | Output |
|---|---|---|
| 0-8 min | Reconfirm criteria, confidentiality, and evidence standard | Shared basis for comparison |
| 8-15 min | Review distribution and identify exception queue | Cases to discuss live |
| 15-45 min | Compare disputed or borderline cases | Evidence-based alignment or documented gap |
| 45-50 min | Bias and consistency check | Questions that need another look |
| 50-60 min | Record changes, unresolved cases, owners, and dates | A usable follow-up record |
The final ten minutes should stay protected. If a new complicated case appears at minute 52, put it into the complete queue. The alternative is a verbal consensus that no one can later explain. Engagedly's sample calibration agenda similarly recommends evidence submitted ahead of time and a timed structure for the session.
Use a Neutral Deferral Phrase
When managers disagree, the facilitator does not need to decide the case from the clock. At minute 47, the facilitator might say, “We have two accounts of the evidence and no shared record to compare. We will capture the question, assign the missing evidence, and return to it with the same criteria.”
That phrasing protects the process without treating delay as a judgment on the employee or the manager. Calibration should never use time pressure to make a consequential rating feel settled.
Make the Transition Visible in Google Meet
Minute Minder starts from a connected Google Calendar event and shows the host time remaining inside Google Meet. It can help the facilitator preserve the evidence window without repeatedly checking a separate clock.
For the 60-minute agenda above, configure Minute Minder:
- Minute 45: “15 min left: stop new cases and protect the decision record.”
- Minute 50: “10 min left: capture unresolved evidence, owner, and review date.”
- Sound: off for a sensitive discussion unless the facilitator needs a stronger private cue.
Minute Minder can also show color cues, play an optional sound, and draft a timed agenda from the invite. It does not record or listen to the meeting. It should be used to protect fair discussion time, never to force a decision.
The reminder text is intentionally about the record, not “wrap up.” In a sensitive people process, the facilitator needs to know whether the team has an agreed rating, a documented evidence gap, or an assigned follow-up. Those are different outcomes and should not be blurred together.
For remote calls, send the exception queue before the meeting and keep it visible to the facilitator. That prevents a shared screen full of appraisal prose from becoming the agenda. It also gives managers notice of the cases where their evidence may be questioned.
When a Longer Review Is the Right Choice
A timer cannot make an underprepared or high-stakes calibration safe. If many cases are contested, the rubric has changed, legal or HR guidance is needed, or managers lack the evidence to compare ratings, extend the process or schedule a separate review. The meeting's job is to make a defensible comparison, not to complete a list at any cost.
The practical improvement is modest: move summary out of the live call, put contested evidence in a visible queue, and leave enough time to write down what remains unresolved. That lets the host end on time without asking people to agree before the facts are ready.
Sources
- UC Davis: Calibration 101
- UC Davis Manager Checklist for Calibration
- Performance Calibration Meetings Guide
FAQ
What should be included in a performance calibration meeting agenda?
Include a brief criteria review, an exception queue, evidence comparison for disputed cases, a bias and consistency check, and time to document decisions and follow-ups.
How long should a calibration meeting be?
It depends on the number and complexity of cases. A 60-minute call can work for a prepared group with a limited exception queue; complex or contested cases may need a separate review.
Should all employees be discussed in a calibration meeting?
Follow your organization's process. A practical approach is to handle aligned, well-supported cases through a fast consent pass and reserve live discussion for exceptions or evidence that needs comparison.