Book companion resource · Picker
Playbook Picker
Letter & A4. Print-ready. Free, no email required.
← Back to toolkit · Browse by chapterPick your worst KPI. Start there. Not the easiest one. Not the one your boss asked about. The one that is costing the most.
Companion to: Fix the System, Chapter 14. Who uses it: the maintenance manager or reliability engineer deciding which of the ten Part IV playbooks to run next. When: at the start of each quarter, or any time a KPI drifts out of range. Why this exists: most plants pick the most tractable playbook instead of the most consequential one. That produces motion without result. The picker forces you to rank by downstream cost, not by ease of execution.
Step 1 — Read your numbers
Pull the current reading for all eight KPIs. If you do not have a reading, mark it as unknown and do not skip it. An unknown is information.
| # | KPI | Current reading | Benchmark (target) | Gap | Downstream cost if left alone |
|---|---|---|---|---|---|
| 1 | PM Compliance % | ≥90% | Corrective work rises 2–3% for every 1% drop below benchmark | ||
| 2 | Reactive Work % | ≤20% | Labor cost runs 2–5× planned-work rate; schedule chaos cascades | ||
| 3 | Schedule Compliance % | ≥85% | Wrench time drops; planner's work gets disrespected; re-planning loop | ||
| 4 | Wrench Time % | ≥55% | Paying for labor you do not receive; capacity gap widens | ||
| 5 | Repeat Failure % | ≤10% | Same asset eats labor 2–5× more than a permanent fix; trust erodes | ||
| 6 | Backlog (ready weeks) | 2–4 | Below 2: you are not finding work. Above 4: you are losing control | ||
| 7 | Inspection Compliance % | ≥95% | Failure modes develop silently; CBM data gaps | ||
| 8 | Parts Fill Rate % | ≥95% | PM delays, hidden truck time, stockpile growth in techs' toolboxes |
Step 2 — Circle the worst
Circle the KPI with the biggest gap weighted by downstream cost, not the one with the biggest percentage gap. A KPI that is 5 points below benchmark but drives everything downstream is worse than one that is 15 points below benchmark but affects only itself.
The three most common "worst" KPIs, and their matching playbooks:
- PM Compliance below 85% → Playbook 15: PM Compliance Recovery (Chapter 15)
- Reactive Work above 40% → Playbook 18: Reactive Work Reduction (Chapter 18)
- Repeat Failure above 15% → Playbook 16: Repeat Failure Investigation (Chapter 16)
If your worst KPI is not one of these three, use the matching table below.
| If your worst KPI is… | Run this playbook | Chapter |
|---|---|---|
| PM Compliance | PM Compliance Recovery | 15 |
| Repeat Failures | Repeat Failure Investigation | 16 |
| Backlog (too large or too small) | Backlog Growth Control | 17 |
| Reactive Work % | Reactive Work Reduction | 18 |
| Schedule Compliance | Schedule Compliance Improvement | 19 |
| Inspection Compliance | Inspection Compliance Recovery | 20 |
| Parts Fill Rate | Spare Parts Management | 21 |
| Declining MTBF on critical assets with no repeat-failure signal | Condition-Based Monitoring | 22 |
| Downtime hours | Downtime Reduction | 23 |
| MTTR | MTTR Reduction | 24 |
Step 3 — Pick a minor (optional)
One major + one minor is the limit. A minor should feed the major, not compete with it.
| Major playbook | Most supportive minor | Why |
|---|---|---|
| 15 PM Compliance Recovery | 21 Spare Parts Management | Parts unavailability is the #1 cause of PM misses |
| 16 Repeat Failure Investigation | 22 Condition-Based Monitoring | CBM catches the failure modes RCA surfaces |
| 17 Backlog Growth Control | 19 Schedule Compliance | Compliance discipline drains backlog |
| 18 Reactive Work Reduction | 20 Inspection Compliance | Better inspections surface work before it becomes reactive |
| 19 Schedule Compliance | 18 Reactive Work Reduction | Reactive work is the single largest threat to the schedule |
| 20 Inspection Compliance | 22 Condition-Based Monitoring | Inspection data feeds CBM thresholds |
| 21 Spare Parts Management | 15 PM Compliance Recovery | Parts only matter if PMs are trying to consume them |
| 22 Condition-Based Monitoring | 16 Repeat Failure Investigation | CBM without RCA is data with no decision |
| 23 Downtime Reduction | 18 Reactive Work Reduction | Most downtime is a reactive-work symptom |
| 24 MTTR Reduction | 15 PM Compliance Recovery | Most MTTR problems are missing procedures, which PMs expose |
Skip this step if you do not have the staff to run two tracks.
Step 4 — Confirm the trigger (week 1 only)
Before launching the playbook, spend one week confirming the KPI is drifting for the reason you think.
- Pull twelve months of data for the KPI and the assets behind it.
- Interview the three people closest to the work (one planner, one supervisor, one tech).
- Sit in on two executions of the work that produces the KPI.
- Write a one-paragraph statement of the actual root cause before launching.
If the actual root cause does not match what the playbook is designed to address, re-pick. You have one week to make this decision without cost. Launching into the wrong playbook burns the whole 90 days.
Step 5 — Read your playbook chapter
Flip to the chapter that matches your circled KPI. Read it end to end before starting week 2. The roadmap is paced. The sequence is the sequence.
If you finish the chapter and the trigger does not fit your situation, come back to Step 2. Do not start the install.
When to run two majors
Almost never. The exception is an existential event: regulatory action, customer loss, safety incident, corporate watch list.
If you must run two majors:
1. Add two senior staff before starting. Not one. 2. Pick playbooks that use different data and different meetings. 15 + 16 is possible. 15 + 18 is not. 3. Plan for 120–180 days, not 90.
If you cannot do 1, 2, and 3, you cannot run two majors. Pick one.
What to avoid
- Picking the most tractable playbook instead of the most consequential. That is the #1 failure mode of this section. A visible quick win on a KPI that is already in range is worse than a slow grind on the one that matters.
- Running three or more playbooks at once. Attention collapses. Every program gets undercooked. You get motion across the plant and results nowhere.
- Launching in week 1. The first week is always diagnosis. Always.
- Stopping at day 60. The stabilization phase (days 60–90) is where the gain becomes permanent. Most regressions happen because the program was declared done at day 45.
- Skipping the benchmark comparison. A KPI with no benchmark is a KPI you cannot pick with. Chapter 4 has the numbers. Use them.
End of Playbook Picker. See Chapters 15–24 for the ten playbooks.