Point Calibration Workflow

Calibration sessions allow managers to collaboratively determine fair point values for strategic initiatives through weighted voting.

flowchart TD
    A["Admin creates calibration session"] --> B["DRAFT
Admin sets Anchor (2x weight),
determining the range"]
    B --> |"Admin opens voting"| C["OPEN
Managers vote
(Owner vote = 3x weight)"]
    C --> |"Admin closes voting"| D["VOTING CLOSED
Review votes, flags,
set final points"]
    D --> |"Admin finalizes session"| E["FINALIZED"]
    E --> F["Points applied to initiatives"]

    B --> |"Admin cancels"| G["CANCELLED"]
    C --> |"Admin cancels"| G
    D --> |"Admin cancels"| G

    style B fill:#f3f4f6,stroke:#6b7280,color:#1f2937
    style C fill:#d1fae5,stroke:#059669,color:#064e3b
    style D fill:#fef3c7,stroke:#d97706,color:#78350f
    style E fill:#e0e7ff,stroke:#4f46e5,color:#1e1b4b
    style F fill:#bbf7d0,stroke:#16a34a,color:#14532d
    style G fill:#fee2e2,stroke:#dc2626,color:#7f1d1d

    click A call noop() "Select which initiatives to include in this calibration session"
    click B call noop() "Anchor is Admin's best estimate (default 2x weight). Range: +100% / −75% of anchor (min 5 pts). Both configurable in Settings."
    click C call noop() "Soft enforcement: out-of-range votes allowed with warning flags (yellow/red). Owner's vote carries 3x weight in the weighted average."
    click D call noop() "Admin reviews all votes and flags. Sets final points — within range needs no justification, outside range requires written justification."
    click E call noop() "Session complete. Point values locked in."
    click F call noop() "Final calibrated point values written to each initiative. Dispensing can now begin."
    click G call noop() "Session terminated. No points applied. A new session can be created if needed."
  

How Voting Works

Admin Range

Before opening a session for voting, the Admin sets an anchor — their best estimate of each initiative's point value. The system uses the anchor to calculate a voting range that guides managers toward reasonable estimates.

• Upper bound = Anchor + 100% (e.g. anchor of 50 → upper bound of 100)
• Lower bound = Anchor − 75% (e.g. anchor of 50 → lower bound of 12.5), with a hard minimum of 5 points

The range percentages and hard floor are configurable in Calibration Settings.

Soft Enforcement

Managers can vote outside the range, but will see a warning before submitting. Out-of-range votes are flagged for Admin review:

• Yellow flag — vote is slightly outside the range
• Red flag — vote is significantly outside the range

Finalizing Points

After voting closes, any Admin (not just the session creator) can set the final point value for each initiative. Two paths:

1. Within range — no justification required
2. Outside range — the Admin must provide a written justification

Admin Anchor

The Admin's anchor serves two purposes: it establishes the voting range and is factored into the weighted average when calculating final bounds. The anchor:

• Is hidden from managers during voting to preserve independent judgment
• Is revealed when voting closes alongside all manager votes
• Carries a configurable weight (default 2x when required, 1.5x when optional)

Calibration States

Not all initiatives require a formal calibration session. When creating a calibration session, the initiative picker shows three states:

Frequently Asked Questions

Q: What if I set the wrong anchor before opening voting?

You can change the anchor freely while the session is in Draft status. Once you open voting, the anchor and range are locked — managers vote against the range you set.

Q: What if I need to initiate a re-vote?

Create a new calibration session with the same initiatives. Set your updated anchor values, then open the new session for voting. The old session can be cancelled. When you finalize the new session, the updated point values will overwrite the previous calibration.

Q: Can I recalibrate an initiative that has already been calibrated?

Only if no points have been dispensed yet. You can add an undispensed initiative to a new calibration session, and the new finalized points will replace the old values. However, once checkpoint points have been dispensed on an initiative, it cannot be recalibrated — the system will block it. For in-progress initiatives that need a point adjustment, the Manager should request a scope adjustment instead, or the initiative can be cancelled and recreated.

Q: Why can’t I change the anchor after voting opens?

Managers cast their votes based on the range you set. Changing the anchor after votes are in would invalidate those votes without giving managers a chance to re-vote. To maintain fairness, anchors are locked once voting begins.

Q: Can a manager benefit from inflating their own initiative’s point value during calibration?

No. The system has three overlapping controls that make honest calibration the best strategy for managers.

Control 1: Vote Capping During Calibration

Votes that fall outside the Admin’s configured range are flagged and have their weight reduced. An Owner vote that normally carries 3x weight is progressively reduced to 2x or 1x the further it strays from the range. Extreme outlier votes are pulled back toward the consensus, limiting their influence on the final weighted average.

Control 2: Efficiency-Adjusted Dispensing Score

A manager’s dispensing score uses the formula: score = dispensed × (dispensed / calibrated). The ratio dispensed / calibrated is your efficiency — how much of the calibrated budget you actually awarded to contributors. If you inflate the calibrated value but dispense the same real amount of work, your efficiency drops and your dispensing score decreases. In practice, most managers will dispense their full budget, so this control is more of a mathematical backstop than a day-to-day deterrent.

Control 3: Admin-Gated Completion Tiers (Strongest)

Completion tiers — Full (3.0x), Partial (1.5x), and Incomplete (0.5x) — account for roughly 75% of a manager’s total Manager Score. The Admin assigns the tier, not the manager. The Admin reviews whether the delivered work justifies the calibrated point value, so inflated calibration makes it harder to earn Full tier because the bar is now higher than the actual scope of work.

Worked Example

In this scenario, the manager who inflated their calibration lost 25 points because the Admin judged the delivered work didn’t justify the higher value and assigned Partial instead of Full tier. The outcome is worse still if the Admin assigns Incomplete (0.5x), which would yield a total of only 225 points. Honest calibration consistently produces the best results.