Monthly Reject-Pile Review Agenda

Thirty minutes. Once a month. Operations and maintenance together. The meeting that keeps identification installed.

Companion to: Fix the System, Chapter 5. Who attends: maintenance manager, maintenance planner/coordinator, operations lead (or operations manager, if single-area review). Cadence: monthly, same day each month (e.g. first Thursday). Duration: 30 minutes maximum. Over time may become 20 if the triage rubric is working. Pre-work: the triager prepares the reject pile for the month — the full list of rejected notifications with their written reasons.

Why this meeting exists

An identification system without a monthly reject-pile review decays within six months. Here's why:

• Triagers drift. What looked like a legitimate rejection in month one might be a drift away from the original intent by month four. The review catches drift.
• Operations and maintenance see rejections differently. Without a regular venue, the two organizations form separate narratives about what counts as a valid notification. The narrative divergence eventually becomes trust erosion.
• Operators stop writing when they see a pattern of rejections without explanation, or when they disagree with rejections and have no channel to raise it.

The review fixes all three in thirty minutes a month.

Agenda

1. Metrics review (5 minutes)

• Total notifications received this month: ___
• Routed to immediate action: ___
• Routed to planning queue: ___
• Rejected: ___
• Average time-to-triage (should be <24h): ___
• Reject-with-reason rate (should be >80%): ___%
• Notifications per operator per month (should be 2-6): ___

If any metric is red, flag for discussion in the closing section.

2. Walk the reject pile (15 minutes)

Spend about one minute per rejection. For each:

• Read the notification aloud (or displayed on screen).
• Read the rejection reason.
• Ask: "Does everyone agree with this rejection?"

Rate each rejection in one of three categories:

• Clear reject (the reason is solid) — move on. No discussion needed.
• Borderline — short discussion. Was the operator right to write it? Should the rejection reason be rewritten? Should the triage rubric change? Decide.
• Wrong reject — the notification should have been accepted. Reopen it as a WO. Follow up with the operator. Note the pattern for triager coaching.

3. Pattern detection (5 minutes)

Look across the month's rejects for patterns:

• Is one area of the plant getting over-rejected? (Possible bias or knowledge gap)
• Is one operator getting their notifications rejected disproportionately? (Either a coaching opportunity for the operator or a coaching opportunity for the triager — figure out which)
• Are the same types of observations being rejected repeatedly? (Signal that the triage rubric or the notification form needs adjustment)
• Are operators writing about issues that should actually be caught by PMs or inspections? (Signal that PM frequency may be wrong)

Note any patterns for action.

4. Rubric adjustments (3 minutes)

Based on the walk and pattern detection, adjust the triage rubric if needed:

• New rejection reason language to add to the standard set?
• New acceptance criteria that were previously unclear?
• New form-field changes or clarifications needed?

If changes are made, communicate them to the triager and to operators within 48 hours.

5. Operator feedback loop (2 minutes)

• Is there any operator who needs direct feedback on their notifications this month? (Either praise or coaching)
• Is there any shift or area that should hear a summary of what's working?
• Are there any "near-miss wins" — notifications that became planned jobs that prevented a failure — that deserve recognition?

Recognition is the cultural fuel of an identification program. If operators see that their notifications produced wins, they write more. If they don't, they stop.

What NOT to do in this meeting

• Do not argue about individual operator competence. This meeting is about the system, not the people.
• Do not debate technical root cause of failures. Those belong in RCA meetings (Chapter 12), not here.
• Do not extend past 30 minutes. If the reject pile requires more than 30 minutes of review, your triage is producing too many rejections or not writing clear enough reasons. That is a triager coaching problem, not a meeting-length problem.
• Do not skip this meeting. If you skip two months in a row, your identification stage will start drifting and you'll notice the notification count dropping within the third month.

Running the meeting the first few times

The first three months of a new identification program often produce a reject pile that's 30-40% of total notifications. That's normal. Operators are calibrating. The triage rubric is calibrating. The reject pile is doing what it's supposed to do.

By month four, well-run programs see the reject rate settle between 10% and 20%. The operators know what to write. The triager knows what to route vs. reject. The system is producing signal.

If your reject rate is still above 30% at month six, the form is probably still too open, or the triage rubric is still too narrow. Walk through it in the review and tighten.

If your reject rate is below 5% — congratulations, your operators are producing well-calibrated notifications. Make sure the triager isn't being too permissive (accepting things that should be rejected to avoid confrontation).

Tracking across months

From Fix the System by Ivan Getov. Companion to Chapter 5: Identify. getov.xyz/toolkit/fix-the-system/ch05/reject-review-agenda. Free. No email required.