Building a replacement scoring model
How to rank vehicles for replacement using a point model — what to score, how to set thresholds, and why the model's real job is defending the list rather than producing it.
Every fleet manager knows which trucks are the problem. The difficulty is proving it to someone who controls the budget and has heard “this one is really bad” before.
A replacement scoring model turns that judgement into a number produced by criteria you published in advance. It is not more accurate than your instinct — it is more defensible, and in a public fleet that is worth more.
The six factors
The structure below is the one most public fleets converge on, in some variation. Six factors, each contributing points, higher total meaning stronger replacement case.
Age — 1 point per year
Simple and uncontroversial. Age drives parts availability, emissions compliance, and the technology gap between the unit and what your technicians are trained on.
Usage — 1 point per interval
Miles for on-road units, engine hours for equipment. The interval defines the class: 1 point per 10,000 miles is a reasonable default for light duty, per 5,000 for severe-duty or heavy units, per 1,000 hours for equipment.
Use hours rather than miles wherever the vehicle does significant stationary work. A bucket truck that drives 6,000 miles a year and runs the boom for 900 hours is not a low-use asset, and scoring it on odometer alone will keep it in service years past the point it should have gone.
Service severity — 1 to 5 points
The same model of truck wears very differently plowing salted roads than shuttling between municipal buildings. Score the duty cycle: load, terrain, stop-start frequency, corrosive exposure, idle time.
This is the factor most often left out, and leaving it out systematically penalises administrative vehicles while protecting the units that are actually being destroyed.
Reliability — 1 to 5 points
Unscheduled in-shop events over the trailing 12 months. Count events, not dollars — the cost shows up in the next factor, and what you are measuring here is disruption. A unit that fails five times cheaply is often a worse operational problem than one expensive repair.
Weight road failures heavily. A breakdown in service costs recovery, driver time, and sometimes a missed service commitment that never appears in the maintenance record.
Repair cost — 1 to 5 points
Lifetime maintenance and repair as a percentage of current replacement cost. Under 20% scores 1, 80%+ scores 5.
Exclude accident damage — that reflects an event, not the asset’s condition. Whether to exclude scheduled PM consumables is a judgement call; excluding them makes the factor a cleaner signal of deterioration, and most agencies do.
Condition — 1 to 5 points
A physical inspection by a technician: body and frame corrosion, drivetrain condition, interior, safety-relevant defects. This is subjective, which is fine, as long as the same person or the same rubric does it every year.
Have a technician do it, not the assigned driver. Drivers are systematically optimistic about vehicles they like and pessimistic about ones they don’t.
Setting thresholds
A common banding, on a scale that tops out near 28:
| Score | Band | What it means |
|---|---|---|
| Under 18 | Below threshold | Keep. Re-score next year. |
| 18–22 | Qualifies | Add to the multi-year plan. Start specification now — procurement lead times are long. |
| 23–27 | Needs replacement | Priority for the current budget request. |
| 28+ | Immediate | Replace as funding allows; consider pulling from front-line service. |
Where a score is not enough
Assets with no replacement market. Specialised equipment sometimes has to be rebuilt rather than replaced. Score it, then handle it as an exception with a written rationale.
Safety-critical defects. A unit with a structural corrosion problem comes out of service regardless of its total score. Build an override into the policy so you are not arguing about it under pressure.
Fleet-wide standardisation moves. Sometimes the right decision is to replace a set of units together to consolidate on one platform, even though some score below threshold. That is a legitimate call — make it explicitly, not by quietly adjusting scores.
Running it annually
The model is only useful if it runs on a schedule and feeds the budget calendar. In practice:
- Score every unit once a year, ideally at the same point each year
- Rank, and draw the funding line
- Publish the full ranked list, not just the funded portion — departments seeing where their unit sits is what makes the process credible
- Keep prior years, so you can show a unit’s trajectory
That last point does more work than people expect. “This truck scored 19, then 24, then 29” is a far stronger argument than a single year’s number.