The arrival window you promise is a traffic bet, and it decides the booking
Booking rate rises when the window is realistic and falls hard when it is not honored.
A caller in Marietta at 2pm and a caller in Decatur at 2pm are different scheduling problems, even if both are the same distance from your shop. The difference is what sits between the truck and the address at 4pm.
Give the person taking calls a simple zone and time matrix rather than judgment. Which zones can still be promised same day at this hour, which are tomorrow morning, which need a firm two hour window instead of a vague afternoon. A dispatcher who knows the answer books with confidence, and confidence closes.
Track late arrivals as a conversion metric, not a service metric. A window missed by ninety minutes costs the review, the repeat visit and often the replacement job that would have followed.
Where you cannot promise, offer something specific instead: a call thirty minutes ahead, a confirmed morning slot tomorrow, a text when the truck is assigned. Specific beats optimistic.
Route by dispatch zone before you route by who is free
Most booking failures are routing failures that happened before anyone looked at the schedule.
Capture the service address at the start of the call, before the symptom conversation. The address determines the zone, and the zone determines what can honestly be promised. Companies that ask for it last end up promising first and discovering the problem afterward.
Zones should reflect drive time and the counties you actually work. Inside the Perimeter, the northern arc, the Cobb side, the Gwinnett corridor. Cross-metro routing that looks efficient on a mileage basis often is not once the day gets going.
Give the call handler explicit authority to say no to a zone that is full, and a script for doing it well. A clean decline that offers tomorrow morning keeps the customer. A booking you cannot run loses them permanently.
Log declined jobs with the reason and the address. It is the cheapest possible market research and almost nobody does it.
The first hard freeze breaks your phone system before it breaks your trucks
Demand in a two-season market arrives in sudden bursts, and intake capacity is what fails first.
The first sustained cold night and the first sustained hot stretch both produce call volume several times a normal day. If the phone plan, the after hours answering arrangement and the staffing are sized for an average day, a large share of the year's best inquiries hit a voicemail.
Pull the call records from last winter's first freeze and last summer's first heat wave. Count missed and abandoned calls by hour. The number is usually worse than anyone in the office believes.
Fix it with a plan rather than heroics: an overflow answering service briefed with your zones, a callback queue, a text-back for missed calls, and a rule about who picks up the second line. Decide it in the shoulder season, not during the event.
An unanswered call is a click you already paid for and a customer who is now calling the next result. Measure it as lost revenue, not as a busy day.
Replacement quotes die in the follow up, not on the kitchen table
System replacement is a considered purchase, and most contractors quote once and go quiet.
The homeowner comparing a heat pump against dual fuel is usually getting two or three quotes, talking to a spouse, and deciding over a week or more. A proposal sent by email with no follow up sequence is a proposal that competes only if the customer remembers to open it.
Build a simple, human sequence: a call the next day to confirm the proposal arrived and answer questions, a follow up several days later with the specific point they hesitated on, and a final check before the season turns. Written by a person, not by a template that reads like one.
Make the proposal itself do work. What was found, what is being proposed and why, what the options cost, what the warranty covers, and what happens next. Ambiguity is what stalls a decision.
Track the interval from quote to decision and the close rate by tech. Both usually vary more than owners expect, and both respond to coaching.
Log the reason you said no, then read the list every month
The jobs you decline describe your business more accurately than the jobs you take.
Every decline has a cause: outside the zone, no capacity that day, equipment you do not service, price sensitivity, warranty work you do not take. Recorded consistently, that list becomes a monthly decision document.
Patterns show up quickly. A cluster of declines from one suburb might mean a second crew or a satellite start point is worth pricing. A cluster of equipment you do not service might mean a training decision or a referral partner.
It also corrects the map. Companies routinely advertise coverage they no longer honor, and the decline log is the fastest way to see the gap between the service area on the website and the one dispatch actually runs.
Read it with the marketing spend next to it. Paying for clicks in a zone you decline is a leak with a number attached.
With one metro of volume, test the season, not the button
A single market rarely produces enough traffic to detect small differences, so change things large enough to show up in revenue.
Split tests on headline wording or button color need volume most single-market contractors do not have. Running one anyway produces a result that looks conclusive and is noise.
Test process changes instead, and test them sequentially over a period long enough to matter. An after hours answering arrangement against no arrangement. A two visit maintenance plan against a one visit plan. Requiring the address first against asking for it last. Those move numbers far enough to see.
Hold the season constant when comparing. Comparing January to April in a two-peak market compares weather, not process. Compare against the same period last year, or run the change for a full season and watch the booked job rate rather than the click rate.
Write down what you expected before you start. A change that everyone remembers as an improvement afterward is not evidence.
Questions we actually get
- Where do most HVAC inquiries actually leak?
- In our experience the usual order is unanswered calls during peak hours, arrival windows promised that cannot be honored, and replacement quotes that get no follow up. All three are process problems rather than marketing problems, and all three can be measured with records you already have: the phone log, the dispatch board and the proposal list.
- How do we know whether the ads or the intake are the problem?
- Reconcile lead source against the dispatch board for a month. If plenty of calls arrive and few become run jobs, the problem is intake or capacity. If few calls arrive but most book, the problem is upstream. Making that one comparison prevents a lot of money being spent solving the wrong thing.
- Is call recording worth setting up?
- Yes, and reading a batch of them is usually the highest-value hour an owner spends all quarter. You will hear what is actually promised, where callers hesitate and which zones cause trouble. Recording rules and disclosure requirements vary, so it is worth confirming your setup with your own counsel before turning it on.
- Should we answer calls after hours ourselves or use a service?
- Either can work, and the decision usually turns on volume during the peak weeks rather than on average nights. What matters more is the briefing: whoever answers needs your zones, your realistic same day rules and the authority to book or to decline cleanly. An answering service reading a generic script will book jobs you cannot run.
- How long should we run a change before judging it?
- Long enough to cover a comparable stretch of weather, which in a two-season market usually means most of a season or the same period as last year. Judging a process change across a few weeks that happened to include a cold snap tells you about the cold snap. Decide the measurement window and the metric before you start, and use booked jobs rather than form fills.
