Maintenance Backlog: The Reliability KPI Most Plants Ignore

Key takeaways

• Maintenance backlog = identified work that has not been completed yet, expressed in crew-weeks.
• A healthy plant carries 2 to 4 weeks of backlog. Zero means you are not finding work. Six-plus weeks means you cannot keep up.
• Backlog should be split into ready (parts, plan, permit ready) and not ready. Only ready backlog can be scheduled.
• The ratio of planned-to-reactive work is a separate KPI. A high backlog of reactive work is a different problem than a high backlog of planned PMs.
• Tracking backlog without categorizing it is one of the most common CMMS reporting mistakes.

Short answer: Maintenance backlog is the volume of identified work not yet completed, usually expressed as crew-weeks. A healthy plant carries 2-4 weeks of backlog: enough to keep technicians scheduled but not so much that priority work stalls. Zero backlog usually means the plant is not finding work. Six-plus weeks means the team cannot keep up and reliability is degrading. The KPI only works when backlog is categorized — ready vs not ready, planned vs reactive. See also Condition Monitoring vs Predictive Maintenance.

Why backlog matters

Backlog is a leading indicator of reliability. When backlog grows, planned PMs slip, condition-based work gets deferred, and reactive work eats more of the day. That trend predates breakdowns by weeks or months.

Tracking it weekly gives the maintenance manager an early signal that the team is losing ground before MTBF degrades and OEE drops.

How to measure it

Backlog (crew-weeks) = Estimated labor hours of open work orders / Weekly crew capacity

If you have 480 estimated hours of open work and a crew capacity of 160 hours/week (4 technicians x 40 hours), backlog = 3.0 crew-weeks.

Most CMMS systems calculate this automatically as long as work orders carry an estimated-hours field.

The healthy range

The widely-accepted heuristic:

• Less than 2 weeks: the plant is not identifying enough work. Operators are not reporting, inspections are not happening, condition monitoring is not driving WOs. Reliability is invisible.
• 2 to 4 weeks: healthy. Schedulers have room to plan, priority work clears in days, low-priority work gets done in weeks.
• 4 to 6 weeks: stress signal. Either planning is slipping, parts are unavailable, or the crew is undersized.
• More than 6 weeks: the team cannot keep up. Either capacity increases or work gets deferred indefinitely — and deferred work usually means future breakdowns.

Ready vs not ready

Open work orders are not all schedulable. Split them:

• Ready backlog — parts in stock, plan written, permits approved, equipment available. This is the work the scheduler can pull from for next week.
• Not-ready backlog — waiting for parts, plan, permit, or production access window. Cannot be scheduled until the blocker clears.

A plant with 4 weeks of backlog where only 1 week is ready has a parts or planning problem, not a labor problem. Adding technicians does not help.

Planned vs reactive composition

The second cut:

• Planned backlog — PMs, predictive work, condition-based actions, modifications.
• Reactive backlog — breakdowns, corrective work.

If reactive dominates, the plant is firefighting and the PM program is broken. If planned dominates, the plant is managing reliability. The composition matters more than the total.

Common mistakes

1. Hiding backlog. Closing WOs without doing the work to keep the metric low. Destroys reliability data.

2. Reporting total backlog without categorization. 4 weeks of backlog tells you nothing about whether it is healthy or about to break the plant.

3. Treating zero backlog as a win. It is usually a measurement failure, not a maintenance win.

4. Sizing the crew to current backlog. Crew size should be set by capacity required to keep backlog in the 2-4 week range, not by the size of today's queue.

How a modern CMMS reports it

A modern CMMS shows backlog in crew-weeks, split by ready/not-ready and planned/reactive, with a weekly trend. The trend matters more than the snapshot — three weeks of growing backlog is a leading signal of degraded reliability even if the absolute number still looks healthy.

Fabrico's CMMS reports backlog with both the ready/not-ready split and planned/reactive composition, and surfaces the trend so the early-warning signal is visible before MTBF starts moving.

See how Fabrico captures this automatically — explore OEE for manufacturing or book a demo.

Related reading

• Condition Monitoring vs Predictive Maintenance
• KPI vs OKR in Manufacturing
• Maintenance Cost as Percent of RAV
• The 8 TPM Pillars

Frequently asked questions

Is zero backlog a good thing?

Almost never. Zero usually means the plant is not finding work — operators not reporting, inspections not happening. Reliability is invisible, not perfect.

How is backlog measured in hours vs weeks?

Estimated labor hours of all open WOs, divided by weekly crew capacity. The output is crew-weeks, which is what schedulers use.

Should I include emergency work in backlog?

Yes. Emergency work is part of the workload. Track it as reactive backlog to separately monitor whether reactive work is dominating.

What about deferred PMs?

Deferred PMs are backlog. Hiding them by extending the PM interval distorts both backlog and PM compliance KPIs.

How often should backlog be reviewed?

Weekly. Trends matter more than snapshots. A weekly review catches drift before it becomes a reliability problem.