The vehicle built for the hard job is often idle when the easy jobs need it

A vehicle equipped for two-person handling can also do a one-person job, obviously. It's strictly more capable, not a different kind of vehicle. Yet in a lot of constraint setups, that vehicle only ever gets matched to two-person jobs, and sits idle the rest of the time, while a separate vehicle gets deployed for the ordinary job sitting one block away.

The vehicle wasn't the constraint. The category it got filed under was.

Capability is a range, not a category

Most fleets are mixed, some vehicles built for standard delivery, others equipped, with crew or equipment, for special handling. A vehicle set up for two-person work can still do everything a one-person vehicle does, that's what being more capable means. But when a constraint model treats two-person handling as its own bucket, separate from one-person work, that vehicle only ever gets offered two-person jobs. It never gets offered the ordinary jobs it's equally able to do. Modeled as a range instead of a category, the same vehicle becomes available for both, and covering the same volume of work takes fewer vehicles overall, not more capable ones sitting idle between the jobs that need them.

This is an easy mistake to make, because it looks like precision

Defining a vehicle strictly by what it's for feels careful, this vehicle is for two-person jobs sounds controlled and specific. But precision about what a vehicle is uniquely suited for is a different thing from precision about everything it can also do. The first is a label, useful for routing the hard job to the right vehicle. The second is a range, and skipping it is exactly what leaves capable vehicles idle.

One vehicle, two jobs, one constraint model

Picture a single shift for a vehicle equipped with a two-person crew.

• The special job, a two-person crew completes an installation that needed the extra handling. The vehicle is now free for the rest of the shift.
• The nearby job, a standard, one-person delivery sits three streets away, well within what the same vehicle and crew could handle without any trouble.
• The assignment, because the constraint model only offers this vehicle jobs tagged for two-person handling, the nearby standard job goes to a second vehicle instead, and the first one waits for the next special job to come along.

The route optimizer did its job correctly. It matched every job to a vehicle that could do it. It never found out the first vehicle could do the second job too.

Every new constraint is a fresh chance to make the same mistake

The range-versus-category mistake doesn't happen once and get fixed. It happens again every time a new constraint type gets added, because the list is never finished. Fuel type, range, weight, volume, stop count, duration, crew size, equipment, skills, geography, none of this is fixed. Fleets keep adding vehicle types, operations keep adding requirements, and each new one is a fresh axis where a vehicle's full range of capability can quietly get filed as a category instead. A mature setup needs to configure any conceivable vehicle, not just a limited internal catalog, and that includes the real capacity of external carriers and contracted providers a mixed fleet depends on, since the same discipline has to hold there too.

Getting a vehicle's type configured completely, not just labeled for what it's for, also matters beyond capacity planning. For EV fleets specifically, the system pulls real range and consumption characteristics directly from that vehicle-type definition, which is its own deep topic worth covering separately. A vehicle defined completely is what the rest of the system can use. A vehicle only labeled is not.

This shows up in any fleet with more than one tier of capability

Big and bulky and installation-heavy operations feel this most directly, since two-person and specialized-equipment vehicles are exactly the ones prone to being modeled as exclusive categories. Mixed fleets combining owned vehicles with contracted carriers feel it doubly, since external capacity needs the same range-based modeling as internal fleet, or the same idle-capability problem shows up across company lines instead of within one fleet. Any operation running more than one tier of vehicle capability is exposed to this in some form, whether or not it's visible yet in the utilization numbers.

The takeaway

Check whether your most capable vehicles are modeled as able to do standard jobs too, or only the special ones they were configured for. If it's the second, you're likely running more fleet than the work requires.