Curtis Boyle looks at fifty-three unanswered calls and thinks about another monthly bill. “We can't afford an answering service. Voicemail costs nothing.” For a four-truck shop watching expenses, that is a reasonable place to start.
Voicemail is not free.
Curtis is our illustrative service-heavy owner in Toledo, not a client case study. His wife, Marcy, handles phones, dispatch, bookkeeping and parts. The question is not whether she works hard enough. It is what happens when the second line rings while she is already helping someone.
His thirty-day printout contains 206 inbound calls. Fifty-three went unanswered. One is highlighted at 2:14 on a Tuesday. Before Curtis buys coverage, he needs to know what that coverage must actually do.
Which hours need a backup, and for what kinds of calls?
Start with your log, not a vendor's package. Sort calls by the time they arrive, whether someone answers and whether you reach the caller afterward. Separate office-hours overflow from evenings, overnight periods and weekends.
Curtis's arithmetic is 53 ÷ 206 = 25.7% unanswered. The remaining 153 were answered. Those are call counts, not counts of qualified leads, lost customers or completed jobs. His log does not establish that all fifty-three calls arrived after closing.
Deduplicate repeated attempts. Separate suppliers, appointment checks, unsupported services and callers outside your territory from eligible new-job inquiries. Then record which eligible inquiries were reached later and which booked. You cannot value every missed ring as a lost replacement.
Mark the coverage window that contains a useful, recurring gap. That may be lunch and late afternoon before it is midnight. Work through the cost of missed HVAC calls if you need to build that first view.
At 2:14, Marcy was handling a supplier problem. All four technicians were out. The highlighted caller was a longtime customer who later replaced a system elsewhere. A nighttime answering package alone would not have covered that Tuesday. Curtis needs a second-line plan as well as a closed-office plan.
Which coverage model can actually finish the handoff?
Compare the task each option can perform and who takes over afterward. “Someone answers” is too vague to buy against.
| Option | A useful fit | What still needs an owner |
|---|---|---|
| Trained internal cover | Predictable busy blocks; someone needs your local knowledge and booking authority | Shift coverage, absences and a backup when both people are occupied |
| Shared answering or reception service | Variable overflow or closed-office intake without staffing each hour internally | Accurate instructions, appointment permissions and escalation when the agent cannot resolve the request |
| Internal on-call rotation | A defined set of calls needs a person with operational authority outside office hours | Primary and backup availability, compensation, fatigue and next-day workload |
| Automated text-back | A permitted mobile contact can receive an acknowledgement when nobody answers | A real person must read, call back, qualify and offer an available appointment |
A mobile app can deliver the call or message to your team. It does not create a fifth person. A shared agent may collect details well but still be unable to reserve a visit unless you grant appropriate calendar access and rules.
Use the call handling checklist to specify what a completed conversation requires. Choose a hybrid if the gaps differ: external overflow while the office is open, for example, and a limited internal escalation rota after closing.
Marcy knows the customers and their equipment. Covering the moments she cannot answer should preserve that knowledge in the handoff, rather than requiring her to remain available every hour the phone can ring.
What can the caller safely be promised after closing?
Write the promise before the script. Decide whether the business can offer live intake, a callback during a stated window, an appointment from a controlled set of slots, or actual technician dispatch. Those are different services.
The BLS description of HVAC work schedules notes evening and weekend work, on-call emergencies and irregular peak-season hours. Your rotation needs to account for the crew doing that work, including the following day's commitments. Forwarding every call to a technician is not a capacity plan.
Give the handler a written route: permitted job types and territory, details to collect, slots they may book, the primary escalation contact, a backup, and what happens if neither accepts. The handler should use your approved urgency instructions and escalate concerns, rather than diagnose equipment or invent a response time.
If you only provide next-business-day callbacks, the acknowledgement should say that. For example:
Thanks for contacting [company]. Our office is closed. We received your request; our team will call during [actual staffed window]. This message does not confirm an appointment or technician dispatch.
Only use a channel the customer has permitted and a window somebody owns. Do not advertise 24/7 emergency service because an automated message can run overnight.
If you choose a sixty-second acknowledgement target, configure and test the trigger on an eligible missed call. That is your operating target, not a benchmark or guarantee. Deduplicate repeat calls against an open inquiry, verify delivery, and route the record to the shared queue with a named person responsible for the callback.
In Curtis's illustrative history, a January Saturday no-heat call waited in voicemail until Monday at eight. The homeowner had already bought a $9,400 replacement elsewhere. The missing element was a timely conversation with an honest next step. A faster automated acknowledgement alone would not have supplied it.
What does each quote need to include?
Ask for the actual coverage windows, included work, billing unit, minimum charge, overages, booking access and escalation charges. Add the internal time required to monitor and complete the handoff.
These are invented budgeting inputs, not current vendor quotes, market rates or Curtis's bills:
| Modeled option | Assumed monthly calculation | Cost and scope to compare |
|---|---|---|
| Internal overflow shifts | 12 paid coverage hours × $28 loaded hourly cost | $336 for those twelve hours, not every evening and weekend |
| Shared reception | $180 base + 100 billed minutes × $2 | $380 before your internal review and follow-up time; verify the assumed contract's windows |
| On-call rotation | $200 availability allowance + 6 callback hours × $30 | $380 for the assumed availability and callback work; additional field labor is separate |
| Text-back plus callbacks | $60 tool fee + 4 callback hours × $28 | $172 for acknowledgement and callback work, not continuous live answering |
The cheapest row may perform a smaller job. Twelve scheduled hours and all-week intake are not equivalent coverage. Replace every input with a written quote or your own loaded labor cost, then compare the same required windows and tasks.
Check how the provider rounds minutes, handles spam and duplicate calls, and charges for transfers or follow-ups. Ask where the call record lands and who corrects an unavailable appointment.
Curtis is right to question a fee. The correction is to price the whole assignment, including the evenings he or Marcy would otherwise spend returning calls. Their time belongs in the comparison too.
How many additional completed jobs would pay for the coverage?
Use contribution after direct job-delivery costs, not the invoice total. The service and install distinction matters because the work, timing and direct costs differ.
Continue the assumed reception example: $380 in service charges plus $40 of additional internal review time equals $420 in incremental monthly coverage cost. Suppose each genuinely additional completed job contributes $140 after direct delivery costs, before this coverage expense.
The break-even calculation is $420 ÷ $140 = three additional completed jobs.
- Two additional completions contribute $280, leaving a $140 shortfall.
- Three contribute $420, covering the modeled expense.
- Six contribute $840, leaving $420 after the coverage expense.
Those are assumptions to test, not predicted results. A booked appointment that cancels does not contribute $140. A caller who would already have booked after your normal callback is not an additional job created by the new coverage.
Compare completed work with the existing recovery process over reasonably comparable periods. Note weather, advertising and availability changes. If more work would require additional staffing or overtime, include those direct costs in the contribution assumption.
The $9,400 in Curtis's Saturday story is a replacement invoice, not a profit figure and not a value you can assign to every unanswered call. His fifty-three misses identify a coverage problem. They do not tell him which option will earn its fee. The model gives him a question he can measure instead.
How do you test the system before a real caller needs it?
Choose one coverage gap and a short pilot with a spending limit and review date. Give both sides the same written instructions. Your team's true weekly capacity should determine which appointments the handler may offer.
Run test calls for six situations: the office's second line is busy; the office is closed; the next service slot is full; the address is outside your territory; the caller already has an open inquiry; and the primary escalation person does not answer.
For each test, follow the record to its end. Was the caller told the truth? Did one responsible person receive the request? Did the backup activate? Did someone explicitly accept the callback? Could two people accidentally book or contact the same household?
Use one shared queue with coverage status, callback owner, promised response window, next action and booking status. Check it at shift changes. An email alert sent is not the same as a handoff accepted.
Track live answers, missed calls recovered, still-unreached eligible inquiries, bookings and completed work separately. Review the workload with the people doing it. The capacity planning checklist should include the office's ability to receive demand, not just technician hours.
For Curtis, Marcy belongs in that test. She knows which details an outside handler must pass along and which promises the office will have to repair Monday morning. The process should make her work possible.
What changes before the next 2:14 call?
You do not have to buy full-time reception to learn where coverage is missing. Start with one measured window and an assignment that reaches an honest conclusion: an available booking, an accepted callback, or a clear explanation of what your company can offer.
Put the primary and backup on a real schedule. Test the routing. Review the expense against additional completed work and the burden shifted onto your own team. Expand only when those results support it.
Picture Curtis taking the printout out of the drawer and sitting beside Marcy. The yellow line still says 2:14. He cannot change the call that happened, or the Saturday that waited until Monday.
He can stop treating her availability as the backup plan. Before the next second line rings, they can point to the person who receives it, the promise that person may make, and what happens if they cannot answer either.