When Service-Business Demand Outruns Dispatch Capacity: A Practical Capacity Model

Headcount is a poor unit of capacity. Ten technicians on the schedule do not equal ten technicians' worth of billable work. Start with paid hours and subtract everything that is not time on a job: drive time between calls, warehouse and supply-house runs, meetings, training, paperwork, paid time off, and unplanned absences.

What remains is productive (on-site) hours. Many operators who calculate this for the first time find it is meaningfully lower than they assumed. That is not a failure; it is the baseline every scheduling decision should start from.

Demand is usually tracked as calls or booked jobs, but jobs are not interchangeable. A maintenance visit, a diagnostic call, a repair, and a full system replacement consume very different amounts of time and very different skill levels. Group work into a handful of job types and assign each an average duration based on completed-job records, not estimates.

Multiply expected volume by average duration for each type. The total is demand expressed in the same unit as capacity. Only then does the comparison mean anything.

Hypothetical example, for illustration only: a company has 8 technicians paid 45 hours per week, or 360 paid hours. Travel averages 40 minutes per job at roughly 3.5 jobs per day, about 11.7 hours per technician per week. Non-job time is 4 hours per technician, and absences and callbacks are budgeted at 3 hours. Productive capacity is about 45 − 11.7 − 4 − 3 = 26.3 hours per technician, or roughly 210 hours for the team.

In a peak week, the company books 60 diagnostic and repair calls averaging 2 hours, 50 maintenance visits averaging 1.25 hours, and 6 replacements averaging 12 crew-hours. Demand is 120 + 62.5 + 72 = 254.5 hours. The gap is about 44 hours, which is roughly one and a half to two additional technicians' worth of productive time — or an equivalent amount of time recovered elsewhere.

The numbers are invented. The method is the point: once both sides are in hours, the size of the gap becomes visible, and so do the levers.

Hiring is slow, expensive, and hard to reverse, so it is worth checking the other levers first. Each one either adds productive hours or reduces the hours demand consumes.

Capacity is only as large as its tightest constraint. A company can have enough technicians and still be constrained by one licensed installer, by a single dispatcher, by equipment lead times, or by the office's ability to confirm appointments and process invoices. Adding technicians in front of a different bottleneck increases cost without increasing completed work.

A useful test is to ask where jobs wait. If they wait for a technician, field capacity is the constraint. If they wait for parts, approvals, permits, or a callback from the office, the constraint is elsewhere.

Rising lead time from booking to arrival, more overtime, more rescheduled appointments, a higher callback rate, and declining customer-review sentiment often move together. Watching them as a set, week over week, is more reliable than reacting to a single busy day.

Averages hide variation. A day with three unexpected long repairs can break a schedule that looks fine on a weekly basis, so capacity plans need a buffer. Durations also drift as equipment, staff experience, and job mix change, so the inputs should be refreshed from completed-job data periodically. Finally, the model says nothing about whether additional demand is profitable; that requires a separate look at margin by job type and lead source.