How should booking rate be calculated using ServiceTitan data?
Bookings divided by bookable calls, not by all calls. Wrong numbers, vendors, employees, spam, and customers checking on an existing appointment are not opportunities, and leaving them in the denominator makes a good intake team look mediocre and hides where the real misses are. Separating bookable from non-bookable requires classifying what the call was about, which the phone record alone cannot tell you.
The denominator is the whole argument
Total inbound calls is an easy denominator and a bad one. A meaningful share of inbound volume is not a service opportunity at all, and the share varies by source, by day and by season. Comparing two CSRs on total calls compares the queues they happened to be in.
A bookable call is one where a person wanted service you offer, in an area you serve, that you were not already scheduled to do. Anything else belongs outside the denominator with a documented reason.
The word documented is doing real work in that sentence. Exclusions that live in someone's head become a way to make the number say whatever is convenient. Written exclusion rules, applied by the same logic every week, are what make the metric worth putting in front of a team.
What to exclude, and why each one matters
- Existing-appointment calls. Confirmations, reschedules and where-is-my-tech. High volume, zero booking potential, and heavily concentrated on days when the schedule slipped.
- Vendors, suppliers and job applicants. Pure noise that rises whenever hiring is active.
- Out-of-area and out-of-trade. A real miss for marketing, not for the CSR. Track it separately — it is often the most actionable thing in the whole dataset.
- Spam and hang-ups under a few seconds. Exclude by rule, and publish the rule.
- Price shoppers you refuse to book. A policy outcome, not a handling failure, and it should be measured deliberately rather than blamed on intake.
Booked is not the same as kept
Booking rate measures the conversation. Kept rate measures whether the job actually happened. A team can book everything and lose a third of it to cancellations caused by a long wait or a poor confirmation process.
Report both. The gap between them belongs to dispatch and scheduling, not to intake, and treating it as an intake problem sends coaching in the wrong direction. Cross-referencing that gap with job and appointment data is how you find out which of the two owns the loss.
Classification is the hard part
Nobody is going to hand-tag thousands of calls. Duration heuristics are unreliable — a long call can be a complaint and a short one can be a booking. Classifying reliably means working from what was said, which is what transcript-level call analysis provides.
Once calls are classified, booking rate becomes stable enough to compare across people, sources and weeks. Before that, it mostly measures who received the strange calls, and the number is not worth arguing over.
Topics: booking rate · call handling · CSR · metrics
Have a version of this question about your own business?
The useful answer usually depends on which systems you run and how they're connected. That's a conversation, not a blog post.