Availability, not headcount
Fourteen unavailable days a year sounds small. Schedule forty per cent of it away and 8.4 working days are left, so each van covers 241.6 of 250 working days — 96.6 per cent availability. Six vans therefore deliver 1,449.6 vehicle-days against the 1,500 the work needs, and the fleet is 50.4 days short over a year without a single thing having gone wrong.
That is the whole insight. The shortfall is not a failure, it is the arithmetic of a fleet with no slack in it, and it appears the moment you write availability down as a number instead of assuming it is one.
A spare is a purchase of van-days
Because a spare is itself only available 241.6 days, it removes 50.4 short days and costs $7,400 to hold. That works out at $146.83 for each short day it covers, which is the number to compare against what a short day actually costs you. The defaults put a short day at $145, so covering the days is cheaper — by $92 a year, on a $7,400 decision.
Ninety-two dollars is not a decision. What that closeness actually says is that the arithmetic has done its job and handed the question back: it now depends on whether a hire vehicle is genuinely available on the morning you need one, in the right size, with your racking and stock in it, which no calculator can answer. A gap that narrow is a signal to go and find that out.
Where the model is optimistic
Downtime is spread evenly here and it never is. It clusters into cold snaps, into batches of tyres that all reach the same age together, into the week two vans go in at once because the same part failed on both. Even spreading also assumes a short day can always be covered by a spare on that exact day, and a spare already out on another job is not a spare.
The practical consequence is to read the short-day count as a floor rather than an estimate. If the table says a fleet is exactly balanced, it will be short some mornings, and how often depends on how correlated your failures are — which your own records will tell you far better than any model.
Related pages
Keep or replace is the other end of the same problem, since an old van is the one generating the downtime days this page is trying to cover. Jobs per day works out what one vehicle can actually deliver in a shift, which is where the number of vehicles you need comes from in the first place. Car ownership cost builds the annual holding figure.
Questions people ask
What downtime figure should I use?
A full year from your own records, not a good quarter. Count every working day a vehicle was not available to send out: in the shop, waiting on a part, being recovered, off for bodywork after a knock, and sitting for an inspection. People routinely undercount this by half because the small days do not feel like events. If you have no records, start keeping them before making a purchase decision on the strength of a guess.
How much downtime can I really schedule away?
Only the part you control, and only if there is somewhere to put it. Routine servicing and inspections can often move to a quiet week or a Saturday if your shop will take them then. Failures cannot move at all, and neither can anything that puts the vehicle off the road for longer than the gap you had available. Set the share honestly — overstating it is the fastest way to make this page recommend a fleet that is short in practice.
Should the driver cost be in the spare vehicle figure?
No, because a spare sitting on the yard has no driver on it. Put in the standing costs only: the payment or depreciation, insurance, registration and anything else that accrues whether it moves or not. On the day the spare goes out it carries the driver who would otherwise have been idle, so the driver cost was already being paid. Adding it to the spare would count the same wage twice.
What if I cover short days by working the other vehicles longer?
Then the cost per short day is the extra hours you pay for, plus whatever it costs to run a vehicle for those hours, and that is what goes in the field. Be careful with this route in the model, though, because it quietly assumes there is spare capacity in the day to absorb the work, and if there were, the fleet would not be short. Whether that pattern is workable at all is a scheduling and an employment question rather than an arithmetic one.
Does the answer change if my downtime clusters?
Yes, and always in the same direction. Clustered downtime means more days where you are short by two vehicles rather than one, and more days where the spare is already committed. The even-spread model cannot see either. Practically, take the short-day figure as the best case, and if your own records show failures arriving in groups, weight the decision further toward holding capacity than the table alone would.