Equipment maintenance software: the preventive maintenance guide for mixed fleets.

Every fleet says it does preventive maintenance. Most are actually doing scheduled reaction: a service list written on paper, a mechanic who remembers which machines are due, and a stack of emergency call-outs that quietly absorbs the budget. The difference between the two is not effort — it is whether the machine tells you when it needs attention, or you find out when it stops.

This guide is the programme we implement with mixed fleets: how to set intervals, how to trigger work from engine hours and fault codes, what equipment maintenance software has to do for it to work, and how to put a number on the return.

Step 1 — Build the machine register before anything else.

You cannot plan maintenance on a fleet you cannot list. Before touching software, get one row per machine with: internal ID, make and model, serial number, year, current engine hours, telematics source (OEM API, CAN gateway, or none), owner (owned, leased, rented in), and home site.

This register is the spine of the programme. Nine times out of ten, building it surfaces two or three machines nobody was tracking at all — and at least one where the hour meter has been replaced and nobody wrote it down.

Step 2 — Set intervals from the OEM manual, then simplify.

Start from the manufacturer's service schedule. Then collapse it into a small number of interval buckets your team can actually remember and stock parts for:

• Daily walk-around — operator, before start: leaks, tracks/tyres, greasing points, damage, fluid levels
• 250 hours — engine oil and filter, general inspection, hydraulic hose check
• 500 hours — fuel and air filters, hydraulic filter, undercarriage measurement
• 1,000 hours — hydraulic oil sample, cooling system, brake and steering inspection
• 2,000 hours — major service, gearbox and final drive oils, structural inspection

Add a calendar backstop to each hour interval — typically six or twelve months — so machines with low utilisation still get fluids changed and seals inspected. Fluids age whether or not the machine works.

Step 3 — Trigger on engine hours, not on dates.

This is the single change that separates a working programme from a wish list. A 22-tonne excavator on a motorway job runs 200+ hours a month. The same machine on a municipal contract might run 50. A date-based schedule services one too late and the other too early, and you pay for both mistakes.

The rule we use: raise the work order at 90% of the interval. At 225 hours on a 250-hour service the planner has time to order parts, find a slot and warn the site — instead of pulling a machine off a job the morning it hits 250.

Step 4 — Let fault codes create work, not noise.

Machines already broadcast their own faults. J1939 SPN/FMI codes and OEM DTCs arrive long before an operator notices anything. The problem is volume: an unfiltered fault feed on 40 machines is thousands of events a month, most of them harmless.

Triage every code into three actions and route them differently:

1. Stop now — oil pressure, coolant temperature, hydraulic pressure loss. Call the operator, then a mechanic.
2. Finish the shift — derate conditions, aftertreatment faults, sensor faults on non-critical systems. Work order for the same week.
3. Schedule it — intermittent and informational codes. Attach to the next planned service so the mechanic sees it in context.

The value is not that the codes exist. It is that a stop-class code on a Volvo and a stop-class code on a Komatsu arrive as the same kind of event, with the same urgency label, in the same queue.

Step 5 — Normalise the mixed fleet.

Most fleets are four or five brands deep. If your process depends on five portals, it depends on five people remembering to check five portals. Whatever equipment maintenance software you choose has to read every source — CAT, Volvo, Komatsu, Liebherr, JCB, Hitachi, Doosan, Manitou, ISO 15143-3 feeds on older units, Modbus controllers on generators and power packs, and the CAN bus directly on machines with no usable API — and turn them into one set of fields.

Engine hours are engine hours. A fault is a fault. Once that is true, every rule you write applies to the whole fleet, including the machine you buy next year from a brand you do not own today.

Step 6 — Make the work order the only unit of work.

A maintenance programme lives or dies on whether the work is recorded. Keep it boring:

• One work order per job, tied to the machine and the hour reading at the time
• Parts and labour attached, so cost-per-hour is real rather than estimated
• Photos from the mechanic on anything structural
• Closed only when the hour meter and the notes are filled in

After a year this history is what tells you which model in your fleet actually costs you money, and what a used machine is worth when you sell it.

Step 7 — Measure four numbers, monthly.

• Planned vs unplanned maintenance ratio — target 80/20 or better
• Service adherence — percentage of services completed inside the interval window
• Unplanned downtime hours per machine
• Maintenance cost per engine hour, per model

If the planned/unplanned ratio is not moving, the programme is not working, whatever the software dashboard says.

The ROI of preventing downtime.

The business case is not oil and filters — those costs barely change. It is the failures you do not have. A realistic unplanned failure on a mid-size machine costs:

• Emergency parts at premium pricing and freight: €1,500–€4,000
• Field service call-out and overtime labour: €800–€2,500
• Two to five days of lost production or a rental replacement: €1,500–€5,000
• Knock-on schedule damage on the site: hard to invoice, easy to feel

Call it €6,000 for a single event, conservatively. On a 40-machine fleet running reactively, five to eight such events a year is normal. Removing half of them is €18,000–€24,000 recovered annually — before counting the fuel saved by catching a clogged filter or the resale value protected by a complete service history.

That is the honest frame for evaluating equipment maintenance software: not the licence line on the quote, but how many failures per year it has to prevent to pay for itself. For most fleets the answer is one or two.

A 30-day implementation plan.

1. Week 1 — build the machine register and confirm the data source for every unit
2. Week 2 — load OEM intervals, set the interval buckets and the 90% trigger
3. Week 3 — turn on fault-code triage, tune the stop/finish/schedule split for a week of real traffic
4. Week 4 — run everything through work orders and baseline the four metrics
5. Then — review monthly, adjust intervals against real failure data, not the manual alone

The short version.

1. Register every machine and its data source before you plan anything
2. Trigger service on engine hours at 90% of the interval, with a calendar backstop
3. Triage fault codes into stop, finish-the-shift and schedule — never a raw feed
4. Normalise mixed-brand data or you will maintain five processes instead of one
5. Record everything as work orders so cost-per-hour becomes real
6. Judge the software on failures prevented, not on licence price