Your techs call the office all day. Which of those calls should exist?

In most service companies the phone traffic between the field and the office is constant, and it is treated as simply the cost of doing business. It is worth taking apart, because the calls are not all the same kind of thing and they do not all deserve the same response.

Sort the calls into three buckets

Have dispatch tally inbound tech calls for two weeks with a rough category against each. Nothing elaborate — a tick sheet is enough. Nearly all of them land in three groups.

Lookups

What does this code mean. What is the charge for this model. What size breaker. What is the part number. These are questions with one correct answer that exists in a document, and the technician is calling because finding it themselves is slower than asking.

History

What did we do here last time. Is this under warranty. Has this failed before. Who was here in March. Again a factual answer, held in your own records rather than a manual — and again, they are calling because retrieving it is awkward from a truck.

Judgment

Should I replace or repair. The customer is refusing the recommendation. This is out of my depth. These require a person, and they are the calls you want more of, not fewer.

Count the real cost of a lookup call

A five-minute call is not five minutes. The technician waits, and often waits twice because the first person did not have the answer. Whoever answers loses their thread on their own work. If it goes to a senior tech on a job, it costs their billable time too.

Run your own numbers rather than trusting a benchmark: tally the calls for two weeks, estimate the true cost of each including both sides, and multiply. For most shops the annual figure is larger than expected and it is entirely invisible, because it never appears as a line item anywhere.

Then look at the calls that never happened

This is the part that matters more than the interruption cost. When asking is slow or embarrassing, technicians stop asking — and they do not stop needing the answer. They guess.

A guess that turns out wrong is a callback, and it arrives weeks later, attributed to skill rather than to the fact that the tech could not get an answer at the moment they needed one. It is worth asking your team directly how often they have proceeded without checking something they wanted to check. The answers are usually illuminating and rarely flattering to the current setup.

Remove the lookups, protect the judgment calls

The goal is not fewer calls. It is a different mix. Every lookup and history question a technician can answer alone is an interruption removed and, more importantly, a guess prevented.

Meanwhile the judgment calls should be actively encouraged. If your tally shows those are rare, that is not efficiency — it is a sign that people are making calls alone that they should not be making alone. A useful signal after any change here is that total call volume drops while the share of judgment calls rises.

What to fix first

1. Whichever lookup question is asked most. It is usually a small number of repeated questions, and there is often one document that answers most of them.
2. History retrieval from the field. If a tech cannot see a site's past work from a truck, every visit re-derives it.
3. The escalation path. Make it explicit who to call for judgment questions, and make it visibly safe to use.