Calibration sessions: the facilitator's guide
Calibration is the highest-leverage 3 hours in your performance cycle. Run it badly and you institutionalise unfairness.
- Pre-work matters more than the meeting. No surprises in the room.
- Anchor on the rubric, not the manager's narrative.
- Facilitator's job: enforce the standard, name the patterns, protect quiet voices.
- Document the deltas — what changed and why. That's the artefact, not the ratings.
Calibration goes wrong in predictable ways: loud managers win, quiet teams get penalised, the standard drifts by 'who's in the room'. The facilitator's job is to make those failure modes visible and prevent them in flight.
Pre-work (5 days before)
- Distribute the rubric + level definitions. Same version everyone uses.
- Each manager submits proposed ratings + 2-sentence rationale per person.
- HRBP pre-reads, flags: distribution skew, missing evidence, calibration outliers.
- Send pre-read to all attendees 48h before. No surprises in the room.
- Time-box: people-per-minute math; if the agenda doesn't fit, run two sessions.
The 3-hour agenda
| Time | Activity |
|---|---|
| 0:00-0:10 | Facilitator frames: rubric, distribution, the standard we're holding |
| 0:10-2:20 | Person-by-person review — manager 90 seconds, room 90 seconds, decision |
| 2:20-2:45 | Distribution check across teams; force conversation on skewed groups |
| 2:45-3:00 | Confirm decisions, capture deltas, owner for each follow-up |
Facilitator moves + scripts
'I'm noticing 80% of this team rated at the top of the rubric. Let's read the rubric for the next level aloud and check the evidence against it for two of them.'
'I hear the narrative. Help me see the evidence against the rubric dimensions — scope, complexity, impact, influence, autonomy. Which dimensions are at the proposed level?'
'Before we move on — Priya, you haven't spoken about your team yet. What's the picture from your side?'
After the meeting
- Document every delta from proposed → final, with the rubric reason.
- Notify managers within 24h of any changes; they own the conversation with the employee.
- Pay-equity sanity check on the final dataset before letters go out.
- Post-cycle retro: what dragged, what worked, what to change next time.
Read next
All playbooksPromotions get political when the process is informal. Here's the end-to-end design — nomination, packet, committee, decision, communication — that holds up…
Leveling rubrics rot at scale. Here's the design pattern — common dimensions, function-specific examples, and the calibration discipline — that keeps the bar…
The system around the review matters more than the review itself. A modern approach to goals, feedback, calibration, and the conversation.