Define “bad” before ranking
Frequency alone promotes nuisance failures; cost alone promotes one-off catastrophes; downtime alone misses safety and quality. Define a site-specific loss model that combines consequence dimensions without hiding them inside one unexplained score.
Use a rolling period long enough to reveal recurrence, then preserve separate views for frequency, downtime, maintenance cost, production loss and risk. The top-ten list should be explainable to the people who work on the assets.
Validate the evidence before the workshop
| Data symptom | Risk | Check |
|---|---|---|
| Many “other” failure codes | False pattern | Read work-order text and interview maintainers |
| Duplicate asset tags | Loss split across hierarchy | Reconcile functional location and equipment |
| Downtime equals repair duration | Production loss overstated | Confirm operating state and redundancy |
| Cost only on parent asset | Wrong component targeted | Trace notifications, parts and labor |
| No exposure denominator | High-use assets penalized | Normalize when duty varies materially |
Use an analysis funnel
- Screen: confirm the loss and asset boundary.
- Pattern: separate chronic recurrence from one major event.
- Mechanism: review failure modes, timeline, changed conditions and physical evidence.
- Method: use a focused defect review for known conditions, RCA for significant unexplained events, or RCM/FMECA when the strategy itself is weak.
- Act: assign an owner, due date, verification measure and exposure period.
Exit criterion
An asset leaves the list only when the agreed loss metric stays below threshold for the defined exposure period — not when an action is marked complete.
Run a short, disciplined cadence
Review the top losses monthly, but manage actions weekly. Limit active deep dives so engineering capacity is not diluted across twenty investigations. Publish movements: new entrants, sustained exits, data-quality holds and actions overdue.
The common failure is a permanent top-ten slideshow with no owners and no verified exits. Five stars because the method converts scarce reliability effort into visible business outcomes — when the evidence and governance are real.
References & further reading
This page is original explanatory writing. Follow the sources for the complete material and context.
- Reliabilityweb. A Risk Threshold Investigation Is Born — includes bad actor system selection. Open source
- Reliabilityweb. Defect Elimination From a CMMS Perspective. Open source
- Rob Reliability. Bad Actor Diagnostic service. Open source
Disclaimer. This independent educational summary is written entirely in Rob Reliability's own words. It is not affiliated with or endorsed by the authors or organizations cited. No source figures or substantial source text are reproduced. This page is a field guide, not a substitute for the originals.
Move from a red list to verified exits
Our diagnostic cleans the evidence, ranks loss and builds an action-ready bad actor dossier.