Key points
- Measure the backlog in crew weeks, not in number of work orders. Forty five-minute jobs and forty overhauls are not the same backlog.
- Watch the trend and the age, not the size alone. A backlog that grows every week, or whose oldest jobs keep getting older, is the warning sign.
- An empty backlog is not the goal. It usually means work is not being written down, or the team has no planned work ready when a slot opens.
- Clean before you cut. Duplicates, jobs already done and requests nobody will ever approve inflate the backlog and hide the real work.
What counts as maintenance backlog
The backlog is work that someone has already decided should be done. That decision is what separates it from a wish list. In practice it includes:
- Open work orders that are approved and waiting to be scheduled or finished.
- Overdue preventive services, because a service past its due point is work the plan says should already be done.
- Repairs on hold, waiting for parts, a permit, a shutdown window or a contractor.
- Corrective work raised from inspections, such as a worn belt found on a walk round that is not yet urgent.
Work requests that nobody has reviewed yet are usually kept out. Until a request is triaged it is not known whether it is real, a duplicate or a job at all, so counting it makes the backlog look bigger than the work. Track unreviewed requests separately, as a queue with its own age. The difference is explained in work request vs work order.
How to measure maintenance backlog
Counting open work orders is the easiest measure and the least useful, because jobs differ so much in size. The standard measure is crew weeks:
Backlog (crew weeks) = estimated hours of approved work waiting ÷ hours your team can spend on that work per week
The denominator is the hours actually available for backlog work, not paid hours. Take out leave, training, meetings and the time the team typically loses to breakdowns that arrive during the week. If you skip this step the backlog looks smaller than it is.
No estimates on your work orders yet? Start with a count and an age split, and add estimates to new work orders as they are raised. Within a few months most of the open work will carry one.
Then measure age. Group open work by how long it has been waiting, for example under one week, one to four weeks, one to three months and over three months, and watch the oldest group. A backlog that is the same size every month but has more work in the oldest group is getting worse.
What a healthy backlog looks like
Planning guides often quote a range of two to four crew weeks as healthy. Treat that as a starting point, not a rule: a site with long shutdown cycles or slow parts supply will carry more, a small team doing mostly quick jobs will carry less. Three signals matter more than any single number:
| Signal | Healthy | Warning |
|---|---|---|
| Trend | Roughly flat month to month | Grows every month |
| Age | Most work is weeks old, very little is months old | The oldest group keeps growing |
| Size | Enough planned work to fill the next few weeks | Near zero, or more than the team can clear in a quarter |
A backlog that is too small is a warning too. If there is nothing planned when a breakdown finishes early, technicians wait or invent work, and it often means faults are being fixed without being recorded.
Why maintenance backlogs grow
- Too much reactive work. Every breakdown takes hours that were meant for planned jobs, so planned jobs slip into the backlog, which leads to more breakdowns.
- Waiting for parts. Jobs that cannot start because a part is not in stock sit in the backlog, often unnoticed.
- Poor job descriptions. A work order that says "check machine" cannot be estimated, scheduled or finished in one visit.
- Work orders never closed. Jobs finished on the shop floor but left open in the system inflate the backlog and make the real figure impossible to read.
- Not enough capacity. Sometimes the honest answer is that the team is too small for the assets it looks after. The backlog in crew weeks is the evidence for that conversation.
Seven ways to reduce a maintenance backlog
- Clean it first. Close jobs that are already done, merge duplicates and cancel requests that will never be approved, with a reason. Many teams find a large share of the backlog disappears here.
- Prioritise by risk. Rank what remains by safety, the cost of the asset being down and how likely the fault is to get worse. Work the top of the list, not the oldest or the loudest request.
- Separate waiting work. Mark jobs that are blocked by parts, permits or shutdowns, and show the reason. Those need a buyer or a planner, not a technician.
- Write better work orders. A clear description, the asset, a photo of the fault and an estimate let a job be scheduled and finished on the first visit. See what is a work order.
- Group jobs by asset and location. When a technician is at a machine for a service, do the small backlog jobs on it in the same visit.
- Protect planned time. Fix the share of each week reserved for planned and backlog work, so breakdowns cannot silently consume all of it.
- Fix the source. If the same assets keep feeding the backlog, look at their preventive plan and repeat failures. Fewer breakdowns is the only lasting way to keep the backlog down. See preventive maintenance intervals.
Reviewing the backlog every week
A short weekly review keeps the backlog honest. Twenty minutes with the planner, the workshop lead and someone from stores is usually enough. Go through four questions in order:
- Is the backlog bigger or smaller than last week, in crew weeks?
- What is in the oldest age group, and why is each job still open?
- Which jobs are blocked, and who will unblock them by when?
- Which jobs go into next week's schedule?
Backlog is one of the ten figures covered in maintenance KPIs, and it reads best beside PM compliance: a growing backlog with falling compliance means preventive work is the first thing being sacrificed.
Maintenance backlog in Solmex
In Solmex the backlog comes from the records the team already keeps. The work orders list shows open work by status and priority and calls out blocked and overdue jobs, each blocker with its reason, such as waiting for parts or a permit. The work calendar lists open work that is past its date or has no date at all, so nothing waits unseen. The preventive maintenance dashboard shows overdue services and how far past due each one is, as a share of its interval, and work requests are kept in their own queue until someone triages them into a work order. Solmex does not estimate hours for you: crew weeks need estimates your planners enter. See work orders and preventive maintenance for details.