Maintenance cost per unit produced: the number that ends the budget argument

The spreadsheet is already up on the screen when you walk in, and Finance has colour-coded it. One cell is red, and it is yours: maintenance spend, up again on last year. The director taps it once and lets the silence do the talking.

Here is the thing, though. You know exactly why that cell is red, and it has nothing to do with your team getting careless. The plant ran harder this year. More product out the door means more wear, more parts, more hours on the tools. But "we made more, so we spent more" sounds like an excuse even as you say it, and you can feel it landing that way around the table.

What makes it so frustrating is that nobody in the room is wrong. Finance is right that the bill grew. You are right that the work grew alongside it. And still the meeting stalls, because nobody has done the one small piece of arithmetic that would let everyone off the hook.

Divide the spend by what the plant actually made, and the argument simply dissolves. The question stops being "did we spend more?" and becomes "did we spend more for every kilogram we shipped?" The first question starts a fight. The second one has an answer, and an answer is something you can work with.

What it actually measures

That answer is maintenance cost per unit produced, and it is just what it sounds like: your total maintenance cost for the period, set against the standard units your plant turned out in the same window. Dollars per kilogram, per litre, per tonne, per carton, even per equivalent standard hour of running time, whatever unit your industry already lives by. Different sectors have their own version of the same idea, from cost per tonne in process plants to labour hours per vehicle on an assembly line, or even cost per horsepower installed.

"Total cost" is generous here, and deliberately so. It takes in everything it costs to keep the assets alive: labour, including the maintenance your operators do themselves, materials, contractors and outside services, the cost of outages and turnarounds, and the capital you spend replacing worn-out machinery at the end of its life. The one thing it leaves out is the money you spend growing the plant or adding new capacity, because that is building the future, not looking after the present. The units are counted gross, before any quality losses, in whatever standard measure your industry already trusts.

How to work it out

The calculation is a single division, which is a good part of why it is so hard to argue with.

Maintenance unit cost = Total maintenance cost ($) / Standard units produced

Put a real year through it. Say the site spent $2,585,000 keeping the place running and shipped 12,227,500 kg of product over the same twelve months.

Maintenance unit cost = $2,585,000 / 12,227,500 kg = $0.21 per kg

Twenty-one cents of maintenance in every kilogram you ship. Hold on to that figure, because on its own it cannot tell you very much yet. Its real value shows up over time, and side by side.

There is no magic number, but there is a powerful comparison

If you are hoping for an industry figure to measure yourself against, here is the honest answer: there is not one worth chasing. By its nature this is an internal benchmark. The units themselves are specific to your trade, so a figure that signals triumph in one sector is meaningless in another, and even within an industry, products and processes vary too much for a borrowed target to hold.

What it is superb at, though, is comparing like with like inside your own organisation. One distribution business that measured maintenance cost per carton across four near-identical sites found the figure varied widely from one to the next, with no difference in what they made; that spread became one of the best improvement tools they had, because the cheaper sites were simply a phone call away and had nothing to hide. Track the number month after month and site against site, and it tells the truth a raw spend figure cannot: if output climbs next year but your cost per kilogram holds, you are not overspending, you are scaling, and you can prove it. If it drifts upward, you have an early, specific warning long before it surfaces as a budget overrun.

A low number is not always a good one

It is tempting to read a falling cost per unit as an unbroken victory, but the figure can flatter you without anyone meaning to cheat. A plant can buy cheap uptime in the short term, leaning on breakdown repairs and a shelf full of in-line spares, and post a handsome number for a while. The cost has not vanished; it has simply moved downstream, into shorter asset life and larger failures waiting their turn. A reactive repair typically runs two to three times the cost of the same job done on a plan, so a plant that looks cheap per unit today is often quietly buying an expensive tomorrow.

There is a subtler trap, too: spending more does not automatically buy a better number, and spending badly actively harms it. In one body of power-plant data, more than half of unplanned outages were traced to errors made during a recent maintenance job; at another site, half of all failures struck within two weeks of installation. Sloppy maintenance creates the very failures it then has to fix. So the honest way to bring the number down is rarely a smaller budget; it is better planning and sharper execution. When work is planned, parts staged and jobs scheduled rather than scrambled, the same crew accomplishes far more for every dollar, and a plant that plans the bulk of its work gets several times the proactive value from it. The lever is preparation and precision, not the budget axe.

Where it can mislead you

Like any good number, it rewards a little care. A few things to keep in mind before you put it on a dashboard:

• Give it a full year, or weight shorter periods so they carry their fair share of planned outages and turnarounds. A quiet month with no shutdown will flatter you; the month you take a major turnaround will look like a disaster. Neither picture is real.
• Keep the unit of measure consistent. Mixing products with different production costs, even when they share the same units, will quietly bend the comparison out of shape.
• Watch the denominator. When production goals are under threat, the count of units or downtime can quietly be massaged, and a curtailment ordered for business reasons rather than maintenance will inflate the figure even when your team has done everything right. Note the context whenever you report it.
• Mind the overheads on a single asset. If you push the metric down to one machine, give it a fair share of the site-wide costs it genuinely leans on, or you will understate what it really takes to keep that machine alive.

So the next time the spreadsheet comes up and the silence settles, you will have something better than a defence. Spend, on its own, is a big figure that invites a fight. Spend per unit is a figure that ends one. It turns the most defensive meeting on the maintenance calendar into a calm conversation about which way the trend is heading, which, when it comes down to it, is the only conversation worth having.