Key Takeaways
- The department you walked into is roughly where the industry is, and treating it as a personal emergency is what burns the first ninety days.
- Order beats effort. Diagnose what you actually have, stabilize it with wins people can see, then build the record that holds the gains.
- Every phase you skip gets paid for later, usually by you, in the quarter you needed most.
Why did you inherit a maintenance department in this state?
Because nothing above the department ever forced it to change. An operation that is coping, even badly, doesn't generate the meeting that fixes it. It generates invoices that get paid and breakdowns that get absorbed. The condition holds until somebody new arrives, or until something breaks badly enough to be noticed upstairs.
One head of engineering at a manufacturer running fifteen sites put this to us plainly. Their sites were comfortable with a few hundred thousand a year in untracked maintenance costs, for the better part of a decade. The push to change, when it finally came, arrived from a global function rather than from any site deciding to modernize itself. Sites don't modernize themselves. Somebody has to arrive and decide.
There's a second reason, and it's the one that never makes it into a handover document. The person who had this job before you took the department's working memory out of the building with them. Which machine makes a particular noise before it fails. Which supplier actually delivers. Which repair held, and which one was a bodge that has quietly been holding since 2019. None of it was written down, because writing it down was never anybody's job. What landed on your desk was the paperwork. The knowledge left through the car park.
That inertia is ordinary, and so is the state it leaves behind. Plant Engineering and ATS found in their 2022 maintenance study that the average facility still manages about 35% of its assets reactively. The department you inherited is close to the middle of its own industry, which is a more useful thing to know than it sounds. A mess you take personally gets attacked from every side at once. A mess you recognize as ordinary gets sequenced.
What should a new maintenance manager do first?
Find out what you actually have, before fixing any of it. Three weeks spent building an accurate picture pays for itself across the ten weeks that follow, because it tells you where to point. Four questions carry most of the value, and none of them need software to answer.
- What exists. Every asset, where it sits, and which ones the operation stops without.
- What's open. The work already promised, however it was promised, including the requests sitting in somebody's phone.
- What repeats. The failures the team stopped reporting because everyone has accepted them.
- Who knows what. The person whose memory the operation quietly runs on.
None of this is glamorous, and it's the only work that makes the next eight weeks land where you aim them. Three weeks of looking is cheap. Ten weeks of fixing the wrong thing is not.
Why does fixing things first cost you the quarter?
Because the loudest failure is seldom the most expensive one, and in week one you have no way to tell them apart. Fixing on instinct spends your attention on whatever shouted most recently. Diagnosis is what tells you which of the fires actually deserves you, and which one is just close to the office.
There's a real price on being good at unplanned work. Klaus Blache at the University of Tennessee's Reliability and Maintainability Center puts emergency repairs at three to more than ten times the cost of the same job done as planned work. Same task, same part, often the same technician. The multiplier comes from doing it at the worst possible moment, with a line down and nobody's calendar respected.
A department that runs on emergencies gets very good at emergencies. The team gets fast, somebody gets thanked for the save, and the schedule underneath keeps slipping. Arriving new and joining that rhythm is the most natural mistake available, because it's the one that earns immediate credit. Six weeks in, you're the best firefighter the plant has had in years and nothing structural has moved.
Stabilizing means taking the top of that list off the board permanently, so the operation stops eating the hours you need for everything else.
How do you earn credibility before you have results?
With a small number of visible wins you chose on purpose. Nobody upstairs is waiting ninety days for a transformation. They're watching to see whether the new manager finishes things. A fix people can point at does political work that a folder of planned improvements never will, and that political work is what buys permission for the bigger change later.
Arriving new is an advantage here, and it's worth naming, because it rarely feels like one. That same head of engineering, and separately a maintenance director with thirty years on the floor, landed on the same observation. New, motivated maintenance managers are the ones who drive change through. Long-tenured managers running a system that already works tend to resist it. The thing you're short of, which is history with this plant, is the same thing that lets you see it clearly.
What stops most managers at this point is a belief about difficulty. The most common conviction blocking a maintenance manager from putting a record in place is that it will be harder than it turns out to be. They've lived through an enterprise rollout, or heard enough about one, and somebody three weeks into a new job has no appetite for a project that could swallow the quarter their credibility depends on.
The way through is to choose the wins from the diagnosis rather than from the noise. The test is whether a fix is small enough to finish inside a fortnight and visible enough that the team notices it was finished. Which fixes clear that bar depends entirely on what your first three weeks turned up. Picking them off a generic list is how new managers end up solving somebody else's problem impressively.
When does a new maintenance manager put a system in?
After the operation stops burning, and before the gains fade. Stabilization buys you a quieter fortnight, and that fortnight is the whole window. Spend it building the record and month six looks like month two. Spend it enjoying the quiet and the improvement lasts exactly as long as you keep standing over it.
Almost nobody goes looking for a maintenance system out of curiosity. The search starts from one of two events. Either somebody new arrived and wants to prove they can fix what they were handed, or something broke so badly that leadership started asking questions nobody could answer. If you're reading this in your first quarter, you're the first kind.
That's worth sitting with, because it means the role found you rather than the other way around. The people who make these systems work are usually pushed into it by an incident, an audit, or a meeting they couldn't answer. Forced champions still succeed, and what decides it is whether they take ownership of a job they didn't ask for.
The horizon is shorter than it looks from week one. Broderick's Handmade had reactive maintenance down 60% four months after going live, and that came from the team using the record daily rather than from any single repair.
Starting clean is easier, and it's worth being honest that it's a different job. Arcadia Cold Storage's director of maintenance startup got to institutionalize best-in-class maintenance from day one, with no inherited backlog and no undocumented history to reconstruct. You're aiming at the same place with somebody else's decade in the way.
What does day ninety actually look like?
Less finished than you'd like. Ninety days doesn't fix a maintenance department. It buys the right to keep going, with a boss who has seen movement and a team that has watched a manager finish something they started. That's a smaller claim than most first-quarter advice makes, and it's the honest one.
Plenty is still broken at day ninety.
- The backlog is smaller and still real.
- The asset list is accurate for the equipment that matters and thin everywhere else.
- The team is split, half using the record properly and half quietly testing whether you'll keep asking.
- The knowledge the longest-serving technician carries is still in their head, and you've barely started on it.
The instinct to start fixing in week one is a good one. You were hired because something needed fixing, and you want to fix it. Ninety days is enough time to do that badly by starting everywhere at once, or properly by working in order.
When a vendor walks you through their system, ask them the question your situation actually raises. Ask what the first month looks like for a manager with no accurate asset list, no reliable history worth importing, and a team that has watched two managers come and go. Their answer tells you whether they've done this with somebody in your position, or only with somebody starting clean.
The department took years to get like this. The work of getting hold of it is shorter and more specific than it looks from week one, and every new maintenance manager who comes through the first ninety days does it the same way. They find out exactly what they've got before they decide what to do about it.
