Maintenance Frequency Against Maintenance Cost

Why this matters

Neither of these numbers means anything alone, which is why they belong in one article rather than two. A high event count on its own reads as a vehicle constantly in the shop; a high total on its own reads as an expensive vehicle. Together they tell you which programme a vehicle is on, because many events at a low average cost and few events at a high average cost are opposite operating philosophies that can post the identical total for the year.

The reason the pair matters more than the total is that the expensive half of the second pattern is not in the maintenance log at all. It is the day the truck is not on the road, and the jobs that were on it. A shop that reads only the spend total concludes that deferring maintenance was free, because the evidence that it was not free is filed somewhere else.

What the two numbers are, and what the log has to say for them to mean anything

The count is maintenance events recorded in the window. The total is the spend on those same events, from the same log, over the same window. Divide one by the other and you get the average cost per event, which is the figure that actually carries the signal.

Three things about the log decide whether any of it is readable, and none of them are visible in the output:

• What counts as one event. A single shop visit that changed the oil, rotated the tyres and replaced a bulb is one visit and three line items. A log that counts line items reports three times the frequency and a third of the average cost against a log that counts visits, on identical work. Pick one and hold it. For this diagnostic the visit is the right unit, because a visit is what takes the vehicle off the road.
• Events that never reached the log. A tyre plugged at a depot, a bulb changed off truck stock, a hose clamped by the tech on the way home. Each one is a real event with little or no invoice behind it, and every one that goes unrecorded drags a vehicle toward the low-frequency, high-average shape, which is exactly the shape this article says to worry about.
• Events with no cost at all. If the log only holds costed lines, warranty and goodwill work disappears and a vehicle under warranty reads as low-frequency and high-average when it is the opposite. A one-year-old van with 9 events, 6 of them warranty at no charge and 3 paid totalling 3.0 units, reads as 3 events at a 1.0 unit average with the free ones dropped and 9 events at 0.33 units with them kept: three times the frequency, a third of the average, same year's work.

Throughout, index spend in units where 1.0 unit is what one standard oil-and-filter service costs on this fleet's light van.

Two vehicles, identical totals, opposite programmes

Two vans, same model, both four years old, both covering 20,000 miles in the same twelve-month window.

Blue's 14 items: four oil-and-filter services at 1.0 each, two tyre rotations at 0.3, a full tyre set at 3.0, pads and rotors at 2.0, a battery at 1.0, a coolant and belt service at 0.8, and four small items (wipers, a bulb, a cabin filter, a top-up) at 0.3 each. That sums to 12.6 units and it rides 6 visits, all booked into a first slot or a Saturday, which is why a fourteen-item year cost half a working day.

Red's 4 items, each its own visit: a planned oil and filter at 1.0; an alternator that failed on the road, tow included, at 3.2; a seized caliper caught on a pre-trip check, pads and rotors at 3.4; tyres and a front end worn out by the alignment nobody had checked, at 5.0, booked in after the caliper visit flagged it.

Both totals are 12.6 units. A per-vehicle spend figure calls these two vans equally expensive and stops there.

Which unit you count decides the headline

Average cost per event, computed both ways on the same records:

Same vans, same window, same spend, and the gap moves from three and a half times to one and a half times on nothing but how the log breaks up a visit. Blue is the van that moves, because a preventive programme batches: its 14 items ride 6 visits, while Red's 4 items are 4 separate trips. Counting line items therefore flatters the comparison in the direction you already believed. Whenever you quote an average cost per event, say which unit it is counted on.

The split that actually decides it: planned against unplanned

The pair is a screen, not a diagnosis. What it is screening for is one specific split, and most logs can produce it from a date and a note: was this visit scheduled before the vehicle needed it, or did the vehicle choose the date?

Hold the two percentages apart: 50 percent is a share of Red's four visits, and 52 percent is a share of Red's own 12.6 units of spend. They are close by coincidence, and on most vehicles they are not, because an unplanned event costs more than a planned one by the nature of the thing.

A workable starting point, tune it to your own trailing years: once unplanned visits pass about a third of a vehicle's visits in a twelve-month window, that vehicle is being run to failure whatever its total says. Red is at 2 of 4, or 50 percent, and is flagged. Blue is at 0 of 6 and is not. There is no credible published figure for what this share should be across small field-service fleets, so do not go looking for one to beat. Benchmark each vehicle against its own prior year and read the direction of travel.

The cost that is not in the log: the day the truck is off the road

Red cost 2.0 more working days off the road than Blue. To add that to a spend comparison you have to convert it, and the conversion has one trap in it: a lost day is not worth a day of booked revenue, because you do not spend the parts, fuel and technician hours on a day you did not work. The honest quantity is the contribution the truck makes in a day, and then only the part of it that never comes back.

Three steps, each one measurable from records a shop already has:

1. Booked revenue per working day for that vehicle. Trailing twelve months of work attributed to the truck, divided by the working days it was available. On this shop's numbers that lands at about 6.0 units per working day, meaning a day books roughly six times what one standard oil-and-filter service costs.
2. Take out the cost of delivering it. Technician time, parts and fuel run about two thirds of the ticket here, so the contribution is about one third: 2.0 units per working day.
3. Take out the work that simply moves. A displaced job is not automatically a lost job. Count how many of the jobs bumped by your last breakdown were actually completed inside the following two weeks. Until you have measured it, start at half of a lost day's work never comes back in your busy season, and most of it comes back out of season, then correct it with your own figure.

Lost contribution equals off-road days, times the share that never comes back, times 2.0 units per day.

Putting the days back in, and what deferral actually saved

Everything below uses the same 12.6 units of logged spend for each van and the same off-road days, 0.5 for Blue and 2.5 for Red. The only thing that moves is the share of displaced work that never comes back.

Work the middle row: Red is 2.5 days times 0.5 times 2.0 units, or 2.5 units of lost contribution, on top of 12.6, giving 15.1. Blue is 0.5 days times 0.5 times 2.0 units, or 0.5 units, giving 13.1. 15.1 over 13.1 is 1.15, so Red cost about 15 percent more than Blue for a year in which the two spend totals were identical to the decimal.

That is the answer to "did deferring maintenance save anything." On the logged spend it saved exactly nothing, because the totals match. Once the off-road days are priced at all, it cost between 15 and 29 percent more, and the only shop for which it broke even is one that had a spare vehicle sitting idle the whole year. Notice what the sensitivity is telling you: the maintenance programme did not set the size of the penalty. Your slack did.

Why there is no minimum-visit filter here

Read this pair on a full twelve months and not less, the same window the per-vehicle spend figure needs, and for the same reason: maintenance is lumpy and seasonal, and one clutch inside a quarter swamps everything around it.

It is tempting to add a second gate, something like "at least six visits before a vehicle is worth reading." Do not. A low visit count is the signal this article exists to find, so a minimum-visit filter deletes exactly the population it was added to clean up. Red has 4 visits and Red is the finding.

What a low count does break is ranking. With four visits, one of them a tow, Red's 3.15 average is one event away from being a different number, so under this article's own rule Red's 3.15 cannot be ranked against Blue's 2.10 as a finding, and the 1.5x in the table above is a screen result, not a verdict. Below about six visits in the window, drop the average and read the two figures that stay honest at low counts: the unplanned share and the off-road days. Both of those called Red correctly without needing an average at all.

Two cases that collapse the whole comparison

The shop with genuine slack. A spare vehicle, a schedule that is not full, and every displaced job resequenced inside the week. The share that never comes back is near zero, both vans land at 12.6 units, and deferral genuinely cost nothing this year. The pair is still worth reading, as a leading indicator rather than a bill: the run-to-failure vehicle will produce the same 2.5 days next year, and one of those days will land in the week the spare is already out.

The single-truck shop. Nothing reschedules, because there is nowhere to put it, so the share that never comes back is 1.0 by construction. Red's 2.5 off-road days are 5.0 units of lost contribution against 12.6 units of logged spend, so the downtime is roughly 40 percent again of everything the maintenance log shows. At that end, the frequency-and-cost pair stops being a fleet metric and becomes a continuity question.

The pair tells you which vehicles will generate off-road days. What those days cost is a fact about your spare capacity, not about the vehicles, and it is worth pricing before you call a low-frequency vehicle a cheap one.

References

• See related: Maintenance Cost per Vehicle and the Three Things It Conflates, for the spend total this pair sits beside and the twelve-month window both share
• See related: Preventive vs Reactive Maintenance: A Decision Matrix, for choosing the programme once the reading says which one you are on
• See related: The Real Cost of Downtime on a Single-Truck Fleet, for the degenerate case where nothing reschedules
• See related: Condition-Based vs Calendar-Based Maintenance: A Decision Guide, for what sets the visit count in the first place
• See related: Zero-Revenue Jobs and Which Numbers They Touch, for the same zero-value-row effect on the revenue side of the records