HVAC call management for small teams

This is the long version, written for a shop with two to six trucks, no dedicated call center, and an owner who still turns wrenches. If you have a dispatch department, most of this will be familiar and the sequencing will not apply to you.

Everything below assumes one constraint that never goes away: the people best qualified to answer your phone are the people physically inside somebody's furnace. No amount of process solves that, so the aim is not to answer more calls yourself. It is to make fewer decisions depend on you being free.

Start by measuring, not by buying

You cannot sequence any of this without knowing where the leak is, and guessing is consistently wrong in the same direction. Owners underestimate ring-outs because the only misses they remember are the ones somebody told them about.

Pull a month from your carrier's call log, ideally a busy one. Count calls with zero or near-zero duration, and anything that went to voicemail. Compare that against jobs booked in the same period. Then check your evenings and weekends separately, because a monthly total averages the damage away.

The full method, including the four other places your number is already recorded, is in how many HVAC calls are you missing.

While you are there, do the eleven-second test: ring your own line from an unfamiliar number at nine in the evening and see what a customer gets. A surprising number of shops discover a forwarding rule stopped working sometime last year.

Understand what you are actually competing on

Worth internalizing before you spend anything, because it reorders your priorities.

A homeowner with a dead system does not compare companies. They work down a list until somebody picks up, and the first company that answers and can come out gets the job. The third and fourth names never learn they were candidates. Your reviews and your certifications get you onto the list; they do very little after that.

Which produces the awkward dynamic at the center of this whole subject: being good at the work makes you slower on the phone, because everyone is busy doing the work earlier calls produced. The market routes emergency work toward whoever is least busy, and that is frequently whoever is least established. The full argument, and why that costs established shops work they should win.

Choose a coverage model deliberately

Four exist and most shops are using one by accident rather than by choice.

Voicemail catches almost nothing, because the caller motivated enough to ring is the caller least willing to leave a message. It is not a coverage model. What it costs.

An answering service takes a message and hands you a callback list. The flaw is structural: the callback is the part you had no time for, and by the time you make it the job is usually gone. The comparison.

An in-house CSR is the best experience of the four and the wrong shape for this trade's demand curve. You need them most in the two weeks the season turns, and they get sick, take holidays, and go home at five. When to hire one anyway.

An agent that texts the caller back within seconds has no queue and no busy signal, which addresses the specific way small shops lose calls. Weaker than a good human on anything unusual. How the two products in this category differ.

Compare them on cost per booked job rather than monthly fee, since a cheap option that recovers two jobs and an expensive one that recovers nine are not competing on price. How to run that.

Write your triage rules down in April

Under pressure, whoever is answering will default to first-come-first-served, which is the worst possible ordering because it lets a filter change consume the slot a changeout needed.

Decide the order while it is quiet: safety first, then no-conditioning in an occupied home with a vulnerable resident, then no-conditioning generally, then partial failure, then replacement assessments, then maintenance. Put it on a card at the desk.

Two rules to add. Replacement inquiries must not sit behind small repairs, because those callers are shopping several companies. And anything with a safety flavour leaves the queue entirely.

The rush playbook covers what else to decide before the weather turns, including moving every movable plan visit out of your peak weeks.

Sort your questions into three buckets

A large share of your inbound calls are not bookings, they are lookups with fixed answers: coverage, hours, brands, whether you take card. Those should stop reaching a human at all.

A second group needs information gathered and a person deciding later: what is wrong, what will it cost, repair or replace. Collect and book, never diagnose or quote.

A third group escalates on sight: safety, complaints, billing disputes, commercial accounts with a process, anyone asking for a person.

The full taxonomy, and the limits worth knowing before you automate any of it.

Decide what happens overnight

Separate two things that get conflated: answering a call at two in the morning, and sending a van at two in the morning. Most shops need the first and not the second, and conflating them is why so many choose voicemail.

Most overnight callers do not need somebody tonight. They need to know somebody is coming, and a firm first-thing appointment holds them. What loses them is silence.

The exception is real and it needs a written protocol: gas smell, carbon monoxide alarm, anyone feeling unwell, no heat in a freeze with a vulnerable occupant. Those wake a person, and the design rule is that ambiguity must break toward waking somebody rather than toward a polite Tuesday booking. The protocol, and whether you need overnight coverage at all.

Fix the follow-up, which is where the money is

Inbound gets all the attention and outbound holds more recoverable revenue in most shops.

Estimates sent and never mentioned again are the biggest single pot. Three messages recover most of what would otherwise go quiet, and here is the cadence with the actual wording.

Customers due for a seasonal service, contacted early enough to fill your mild weeks rather than competing with your own emergency calls. The message craft.

Dormant customers, sorted by equipment age rather than by how long since you saw them. Why that ordering matters.

Plan renewals and warranty dates, both entirely date-driven and both routinely missed. Plans and warranty touchpoints.

The order to fix things in

Measure first. Then take the lookups off your people, because it buys the desk room to breathe. Then catch the ring-outs, which is where the largest inbound loss sits. Then cover the evenings. Then work the estimate backlog. Then the reminder and reactivation lists, in batches sized to your actual calendar gaps.

Hire when the work left over is complex rather than merely constant. When that is the right call.

Judge it on booked jobs

Take a baseline before you change anything, then track the chain: ring-outs, replies sent, customer responses, jobs booked, jobs serviced. Cost per booked job is the figure that settles it. Ignore response-time and message-volume metrics, which prove a system is running rather than working. The measurement guide.

Where we fit into this

We build Orelle for exactly this size of shop: she sits behind the number already on your trucks, replies to ring-outs within seconds in your voice, handles the bucket-one questions, books onto the calendar your dispatcher already reads, works the follow-up lists, and escalates by rules you wrote. Your techs change nothing, which is the only way any of this survives a July.

If you would rather have the short version aimed at doing something this week, start here. If you want the trade-specific overview, what we build for HVAC. And if you want somebody to read your call log with you, that is a conversation.