Every vehicle has an economic replacement point: the age or mileage at which total cost per mile stops falling and starts rising. Finding it is arithmetic. Acting on it is a capital allocation decision that most fleets postpone, usually for reasons that are understandable and expensive.
The cost curve
Two forces work in opposite directions:
- Ownership cost per mile falls with age. Depreciation is heaviest in the first years; spreading a fixed purchase price over more miles reduces the per-mile figure.
- Operating cost per mile rises with age. Maintenance increases, reliability falls, fuel efficiency degrades slightly, and downtime becomes more frequent and more expensive.
Total cost per mile is the sum. It falls, flattens, and then rises. The bottom of that curve is the economic replacement point.
In practice the curve is usually flat across a range rather than sharply pointed, which is good news: the decision does not need to be precise, and there is often a two-year window where cost is effectively equal.
Building the analysis
1. Assemble the data per vehicle, by year of life:
- Acquisition cost and funding cost
- Depreciation (or estimated market value by year)
- Maintenance and repair cost
- Tyres
- Fuel or energy
- Insurance, tax and licensing
- Downtime days
- Distance travelled
2. Compute cumulative cost per mile at each year:
`` Cumulative cost per mile (year n) = (acquisition − residual at year n + operating costs to year n) ÷ cumulative distance to year n ``
3. Plot it by vehicle class. Not fleet-wide — a 3.5-tonne van and a 26-tonne rigid have entirely different curves.
4. Identify the minimum and the range within which cost is within a few percent of it. That range is your replacement window.
Why fleets keep vehicles too long
- Capital constraint. Replacement requires cash or credit; keeping the old vehicle does not, at least not visibly.
- Depreciation looks like a sunk cost. It is not — the residual value being lost each year is a real cost of keeping the asset.
- Maintenance spend is normalised. A vehicle costing a lot to maintain is "just an old one", not a candidate for disposal.
- Downtime is not costed. The operational cost of failures lands in a different budget from the maintenance saving.
- No trigger. Without a defined review point, nothing happens until a vehicle fails catastrophically.
The last is the most fixable. A rule that any repair above a defined proportion of the vehicle's residual value triggers a formal repair-versus-replace review catches most of the value with no analysis overhead.
Complications worth modelling
Residual value volatility. Used vehicle markets move substantially, and the timing of disposal can affect returns more than a year of extra maintenance. Where the market is strong, replacement is cheaper than the model suggests.
Emission zone compliance. A vehicle that cannot enter urban zones has a restricted useful life regardless of its condition, and its residual value reflects that.
Electrification. Replacing with an electric vehicle changes the cost structure entirely — higher purchase, lower running cost, uncertain residuals, plus infrastructure. It is a different analysis, not a variant of the same one. See EV telematics.
Lease versus purchase. Contract hire fixes replacement timing by contract, converting the decision into a procurement one. Whether that is good depends on whether the contract term matches your economic point.
Warranty expiry. Costs frequently step up at the end of manufacturer warranty. Some fleets replace at warranty expiry as a policy, which is a defensible simplification if it approximates the economic point.
From analysis to policy
Convert the curve into rules people can apply:
| Vehicle class | Target replacement | Review trigger |
|---|---|---|
| Light van, urban multi-drop | Defined age/mileage from your curve | Repair above X% of residual |
| Light van, motorway | Longer, from your curve | Same |
| Heavy rigid | From your curve, often hours-based | Major component failure |
| Specialist equipment | Case by case | Annual review |
Publish the policy, apply it consistently, and review the underlying analysis annually as costs and residuals change.
Frequently asked questions
How long should we keep a delivery van?
There is no universal figure — it depends on annual mileage, duty cycle, maintenance cost and residual values in your market. Build the curve from your own data; fleets with similar vehicles in different applications routinely reach replacement points years apart.
Is it cheaper to run vehicles until they fail?
Almost never, once downtime, roadside recovery, service failures and the risk of a catastrophic failure are included. Running to failure looks cheap in the maintenance budget and expensive everywhere else.
Should we replace on age or on mileage?
On whichever drives cost in your operation, usually mileage for high-use vehicles and age for low-use ones. Most policies should specify both, with replacement at whichever comes first.
What about vehicles that are barely used?
Question whether they should exist. A vehicle used a fraction of the time incurs depreciation, insurance, tax and compliance cost regardless. Hiring for the peak is frequently cheaper — see telematics data you already pay for for how to identify them.
How does leasing change the analysis?
It fixes the term and converts capital cost into a predictable monthly charge, which simplifies budgeting but removes the flexibility to replace at your own economic point. Compare the contract term against your calculated replacement window before signing, not after.
