Maintenance teams get asked for KPIs constantly — by plant managers, operations directors, and finance teams trying to quantify whether the maintenance budget is producing results. The frustrating reality is that most facilities track 2–3 metrics that are easy to pull from their CMMS and ignore the 7–8 that would actually tell them whether their maintenance program is heading in the right direction.
This article covers the 10 KPIs we consider essential for a reliability-focused maintenance team, with exact calculation formulas and the specific data sources in IBM Maximo, SAP PM, and Fiix where each metric lives. Not all 10 are equally important for every facility — but you should at least know how to calculate them before deciding which ones to prioritize.
The Foundational Three: MTBF, MTTR, and OEE
These are the KPIs most facilities already track. We'll keep the definitions crisp and focus on the calculation errors we see most often.
1. Mean Time Between Failures (MTBF)
Formula: Total uptime ÷ Number of failures in the period
Data source: Work orders with failure-type classification (not just all corrective work orders — PM completions are not failures).
Common error: Including planned maintenance downtime in the "failure" count, which artificially deflates MTBF and makes the asset look less reliable than it is. In IBM Maximo, filter work orders to Activity Type = "EM" (emergency/breakdown) only. In SAP PM, use Order Type = PM01 (corrective) and exclude PM02 (preventive).
2. Mean Time To Repair (MTTR)
Formula: Total repair time ÷ Number of repair events in the period
Data source: Work order actual finish time minus actual start time. Not report date minus creation date — that includes the waiting-for-parts period, which is a procurement KPI, not a repair execution KPI.
Common error: Using the full work order duration instead of the wrench-turn time. A work order that sits open for 14 hours waiting on a bearing has an MTTR of 14 hours on paper but an actual repair time of 45 minutes. Separate "waiting time" and "active repair time" if your CMMS supports it.
3. Overall Equipment Effectiveness (OEE)
Formula: Availability × Performance × Quality
Where Availability = (Planned production time − Downtime) ÷ Planned production time
OEE is the most widely misquoted KPI in manufacturing. An 85% OEE target is often cited as the aspirational benchmark — but that figure applies to discrete manufacturing with medium complexity and moderate changeover frequency. For high-mix, low-volume operations with frequent changeovers, 65–70% OEE is the realistic median. Comparing your OEE to an abstract benchmark without accounting for product mix and changeover frequency is misleading.
The Maintenance Mix KPIs: Tracking Where Your Labor Goes
4. Corrective-to-Preventive Maintenance Ratio (CM:PM Ratio)
Formula: Corrective work order hours ÷ Preventive work order hours in the period
Target range: Most reliability frameworks target a CM:PM ratio below 1.0 (more planned work than reactive work). Facilities in early-stage reliability programs often run 2:1 or higher.
Why it matters: This ratio is a proxy for how reactive your maintenance culture is. High CM:PM ratios usually indicate either underfunded PM programs, PM tasks that aren't calibrated to actual failure modes, or assets that need condition-based maintenance but are being maintained on fixed intervals instead.
5. PM Compliance Rate
Formula: PM work orders completed on schedule ÷ PM work orders due in the period × 100
Target: 85–95% depending on facility complexity. Below 70% is a program health warning.
Data source: In Fiix, this is the "Compliance Rate" dashboard widget. In SAP PM, run the MCI4 (maintenance compliance) report against the relevant work center and period. Don't confuse completion rate with on-time completion rate — a PM completed 3 weeks late still counts as complete in some CMMS reporting modes.
Asset Health and Failure Intelligence
6. Mean Time to Detect (MTTD)
Formula: Time of failure event − Time of earliest recorded anomaly or alarm
This KPI requires your PdM or condition monitoring system to log anomaly timestamps. Without a condition monitoring system, MTTD is effectively zero for most facilities — you find out about failures when they happen, not before. Tracking MTTD is one of the most direct ways to measure whether a PdM program is actually advancing your detection capability. A well-functioning edge AI system should show MTTD values of 24–72 hours on bearing and gearbox failures.
7. Planned Maintenance Percentage (PMP)
Formula: Planned maintenance hours ÷ Total maintenance hours × 100
Closely related to CM:PM ratio but measured in total labor hours rather than work order count. Target: 70–85% planned for a mature reliability program. PMP below 50% typically indicates a reactive maintenance culture with insufficient PM planning bandwidth — technicians are always firefighting, which is also why PM compliance suffers.
8. Repeat Failure Rate
Formula: Number of assets with 2+ corrective work orders in 12 months for the same failure mode ÷ Total monitored asset count × 100
This KPI is underused and highly valuable. Repeat failures on the same asset for the same failure mode are almost always a diagnostic problem — either the root cause wasn't identified the first time, or the repair addressed the symptom without fixing the underlying degradation mechanism. A repeat failure rate above 10% on critical assets is a signal to review your root cause analysis process, not just your PM schedule.
Financial and Resource Efficiency KPIs
9. Reactive Parts Spend Percentage
Formula: Emergency/unplanned parts spend ÷ Total maintenance parts spend × 100
Industry data puts the reactive parts cost premium at approximately 40% above planned procurement rates — emergency sourcing, expedite fees, and premium freight add up quickly. Facilities running mature PdM programs typically see reactive parts spend fall to 10–15% of total. Facilities in reactive maintenance mode often run 35–50% reactive, sometimes higher.
This is one of the clearest financial signals available to justify a PdM investment: track your reactive parts spend for one quarter, then calculate what a 50% reduction in that number would mean to your maintenance budget. In our experience, the number is usually large enough to make the business case on its own.
10. Maintenance Cost as a Percentage of Asset Replacement Value (RAV)
Formula: Annual maintenance cost ÷ Estimated replacement value of all maintained assets × 100
Benchmarks: Well-performing manufacturing facilities typically run maintenance cost at 1–3% of RAV annually. Facilities with high reactive maintenance and aging equipment often run 4–6%. Above 6% consistently is a strong indicator of deferred maintenance and asset deterioration that is compounding over time.
RAV is harder to calculate than the other inputs, but most facilities have replacement cost estimates in their fixed asset register or insurance valuation. Don't use book value — use current replacement cost.
Building a Dashboard That Actually Gets Used
The practical challenge with a 10-KPI framework is that maintenance teams don't have time to manually compile these metrics every week. The solution isn't to track fewer KPIs — it's to automate the data pulls from your CMMS into a dashboard that refreshes automatically.
Most CMMS platforms expose report APIs or export functions that make this feasible without IT involvement. In IBM Maximo, the BIRT report module can be scheduled to email CSV exports. In SAP PM, transaction code MCIT (equipment analysis) covers most of the failure history metrics. In Fiix, the analytics module includes most of these KPIs out of the box.
Start with three: MTBF, CM:PM ratio, and PM compliance. Once those are running automatically, add MTTD when your condition monitoring system is in place, and reactive parts spend when you have visibility into your procurement data. The goal is a weekly 5-minute review that tells your team whether the program is improving — not a quarterly slide deck that no one reads until budget season.
