Fifteen working days a year, on hold.

The wait is not a rumor. Technicians time it.

Trade forums are full of it, and the useful part is that the complaints come with durations attached. A sample of what technicians report about the major manufacturers' support lines:¹

One technician describes calling tech support “daily, sometimes multiple times a day.” Another, asked how often he needs help troubleshooting, says he calls coworkers and friends in the trade constantly — the manufacturer's line is not the first call, because the first call has a queue.¹ That workaround is the actual finding. The question still gets asked. It just gets asked to whoever picks up.

Why it is this hard to look up

A service company sells two or three brands and services whatever is already installed. Equipment runs fifteen to twenty-five years, so the trucks roll up to roughly three decades of model years across every badge on the market — and fifteen or so manufacturers own about twenty-five badges between them. Carrier is also Bryant, Payne and Totaline. Trane is also American Standard, Oxbox and RunTru. Daikin is also Goodman and Amana.

The document count follows from that. Six manufacturers, counted at a single public manual archive:²

That is before York, Bosch, LG, Fujitsu and Nortek, before the controls layer, and before the hydronic side. Each unit carries several documents in its own right: install manual, service manual, parts list, wiring diagram, spec sheet, and a running stream of service bulletins.

No technician is going to find the right page in that pile while a customer watches. So they call. The call is not a training problem or an attitude problem. It is a retrieval problem wearing a phone.

The arithmetic

The median wage for heating, air conditioning and refrigeration mechanics is $29.33 an hour.³ Add the usual 35% labor burden and the technician costs about $39.60 an hour before a truck moves. Service companies bill that hour out somewhere between $70 and $150.⁴ The tables below use $110.

Call frequency is the one input nobody publishes, so this runs three cases rather than pretending to a number: a light week at two calls, a typical week at three, and a heavy week — a newer technician, or a brand with a bad line — at five.

The middle row is the one to sit with. 117 hours a year is about fifteen working days — three working weeks of a technician's year spent listening to hold music. Across ten technicians it is 1,170 hours, which is more than half a full-time position you are already paying for and cannot dispatch.

The bigger number is the second visit

Hold time is the visible cost. The expensive one is the job that does not close because the answer never arrived. Median first-time fix rate across field service is 71.9%,⁵ and when a job is not fixed on the first visit it takes 1.6 more visits on average to finish it.⁶

At eight jobs a week, roughly 117 jobs a year per technician do not close first time, which generates about 187 additional truck rolls. Move first-time fix by four points and the arithmetic moves like this:

Four points is a modest claim, and it is already worth more per technician than the entire hold time. Which is the real argument: the phone call is not the problem. The phone call is what a retrieval problem looks like on a Tuesday afternoon, and the return visit is what it costs.

What we are not claiming

Two things in the arithmetic above are softer than the tables make them look, and you should know which.

Hold time and billable time overlap. The technician who stayed on hold while driving to the supply house was not idle. He was working the queue in the background. So the billable column is the generous reading and the wage column is the honest floor. The truth sits between them, and anyone telling you otherwise is selling harder than we are.

Not every call is a lookup. Warranty authorization, parts availability, a purchase order, a genuinely undocumented failure — no software answers those, and Answerlore does not pretend to. It answers the ones where the manual already holds the answer and nobody can find the page. In our own modeling we assume that is around sixty percent of these calls, and that number is an estimate we would rather show you than bury.

Everything sourced below is published. The call frequency, the sixty percent, the $150 dispatch and the eight jobs a week are ours, and they are the inputs the whole model swings on. Change them and the tables change with them. Why it refuses →