If you could only track one number in the front office of an HVAC, plumbing, or electrical company, call booking rate would be a strong candidate. It tells you how many of the people who called wanting service actually ended up with a job on the calendar.
It is also one of the most misreported numbers we see.
ServiceTitan has published data on this. In one of its reports on average call booking rates, the typical shop on its platform converted around 42% of calls into booked jobs, far below what most owners assume. Their analysis also found booking rates varied by trade and season, and dropped in the busiest months, exactly when the calls are worth the most.
Whatever your number is, the first question is whether it is measured correctly.
The formula is simple. The inputs are not.
Booking rate is:
Booked calls ÷ calls that were real booking opportunities
The denominator is where things go wrong. If you divide by every inbound call, vendors, existing job updates, and wrong numbers drag the number down and hide what your CSRs are actually doing. If you let CSRs decide which calls count, the number tends to look great, because calls that did not book quietly get labeled "not an opportunity."
Both mistakes are common. Both make the number useless for coaching.
How to measure it honestly
Define what an opportunity is, in writing
A caller who wants service, maintenance, or an estimate, and is inside your service area, is an opportunity. A caller asking about an existing appointment is not. A price shopper who hangs up after hearing the diagnostic fee usually is, because that is a call your script should be handling.
Write the definition down and use the same one for everyone.
Classify calls consistently
This is the part that used to require a manager listening to recordings. It is also where AI is genuinely useful today. A model can listen to every call, classify it against your definition (new service request, existing job, vendor, spam), and flag the opportunities that did not book. Your manager reviews the flagged calls, not all of them.
Split residential and commercial
A homeowner with no heat and a facility manager requesting a quote for a new site behave very differently on the phone. Their booking rates should be tracked separately, and so should the scripts.
Look at it by hour and by CSR
An overall rate hides the pattern. The useful views are by hour of day, day of week, CSR, and call source. That is where you find that the evening shift books far less than the morning, or that one marketing channel brings in calls that rarely convert.
What usually moves the number
Once the measurement is honest, the improvements are rarely mysterious:
- Answer more calls. Abandoned and voicemail calls are opportunities you never had a chance to book. We wrote about measuring the cost of missed calls separately.
- Show real availability. CSRs who cannot see the board offer callbacks instead of appointments.
- Handle the price question. A clear, confident answer on diagnostic fees and price ranges books more calls than avoiding the question.
- Coach from real calls. Ten minutes a week reviewing the flagged non-booked calls with each CSR beats any training deck.
- Cover the gaps. After hours, lunch, and peak-season overflow are where an AI voice agent can book calls that would otherwise go unanswered, as long as it is scoped correctly.
A number everyone trusts
The goal is a booking rate that the owner, the call center manager, and the CSRs all believe. When people trust the number, they use it to improve. When they do not, it becomes a monthly argument.
Getting there is mostly about definitions, consistent classification, and connecting your phone system to your field service software so booked calls are matched to actual jobs. That is a small project with an outsized effect on everything that follows.