Plumbing CSR Booking Rate: Count the Unbooked Calls Before Buying Anything

Why the published benchmarks don’t settle it

The plumbing booking-rate targets we read come from software vendors, AI-receptionist sellers and marketing agencies, and they disagree with one another. Even an honest benchmark would compare numbers built from different definitions: one shop excludes price shoppers and existing-customer calls, another counts every ring.

The call mix differs too. A drain-and-sewer shop that runs emergency work answers a different phone from a shop that does mostly scheduled water-heater and repipe work, and a company with property-manager accounts takes work-order calls that are booked before the phone is picked up. The useful comparison is the shop against itself: the same definition, counted the same way, week after week.

Where unbooked calls hide in a plumbing company

The Bureau of Labor Statistics notes that plumbers work nights and weekends and are often on call to handle emergencies. The phone keeps that schedule too. These are the places a call can leave the booking rate without a trace.

• The 59-second call. A caller asks the service fee, hears it and hangs up. Under a 60-second rule it is excused automatically, and nobody calls back to ask what the drain was doing.
• The convenient call reason. “Out of area,” “price shopper,” “just a question” — each one fair sometimes, and each one a way for an unbooked call to leave the count without a note saying why.
• The answering-service message. Night and weekend calls taken by a service land in its portal or an email, not in the phone system, so they are never matched against the morning’s bookings.
• The cell phone. Repeat customers and property managers call a tech or the owner directly. Those calls are in nobody’s log.
• The promised callback. “Let me check the schedule and call you back” is answered, longer than a minute and not booked — and whether the second call happened is written down nowhere.

A one-week count by hand

No software purchase is needed for a first count. One person in the office, a spreadsheet and one ordinary week are enough.

• Export every inbound call for the week from the phone system: time, caller number, duration, who answered.
• Add the answering-service log for the same week, and ask the techs and the owner for any customer calls that came to their cells.
• Export the jobs booked that week with their created time and the customer’s phone number.
• Match each call to a booking by phone number and time. A call with a job created within the hour is booked.
• Sort the rest by hand into four piles: not a customer (vendors, deliveries, staff); existing job business (reschedules, ETAs, payments); not serviceable, with the stated reason; and serviceable but not booked.
• For the last pile, write one line each: what the caller wanted, who answered, and what happened next.

The last pile is the finding. A blank in “what happened next” does not prove nobody called back; it means any callback happened from memory and left no trace. Doing the same count the following week gives the shop its own trend, which is worth more than any benchmark.

Turning the count into a monthly dollar figure

Illustrative example, not a client: one week’s count finds 12 callers who described a job the shop does, reached a person or the answering service, and ended without a booking or a stated reason. The shop’s own average ticket on those call types is $400. That is $4,800 of work asked for in one week, or about $20,800 over a month at the same pace.

That figure is the ceiling, not the loss. Some callers were never going to book — the price was wrong for them, or another company arrived first. The owner, who knows the market, decides what share was realistically winnable and writes the range down. Even a cautious share of a number nobody had counted before is a better basis for a decision than a vendor’s benchmark. Speed matters as well: how fast the callback happens decides how much of the pile is still winnable by the time someone dials.

What fixes unbooked calls — process before tools

The first fixes cost nothing but a decision about who owns what.

• Review the call reasons. Decide which ones may leave the count, and require a one-line note on any call marked as not a lead.
• Give every serviceable, unbooked call a same-day callback owner by name — the dispatcher, the CSR lead or the office manager — and a time by which it is made.
• Bring the answering-service messages into the same list the morning shift works, before the first truck rolls.
• Ask the techs to forward customer calls to the office line, or to log them, so repeat work enters the record.
• Count again every Friday with the same definitions, and look at the trend.

CSR scripts and phone training are a real lever, and they are a trainer’s job. An answering service or an AI receptionist answers more calls, which helps only if somebody owns what happens after the call. Tools come after the count, and only where the count says a tool would pay.

When not to buy anything for the phones

If the week’s count turns up a handful of serviceable calls, each with a reason and a callback, the phones are working and the money is leaking somewhere else — open estimates, finished jobs waiting to be invoiced, or slow-paying accounts. If every slot on the board was already full that week, the limit is capacity, and more booked calls would only move the wait from the phone to the schedule. And in a one- or two-truck shop where the owner answers every call, the count is the owner’s memory; there is not yet a hand-off to lose.

Our Revenue Review counts this hand-off and every other one from the first call to the deposit, from the shop’s own records: two weeks, read-only, ending in a written report with each leak priced where the records allow. It sells no calls, no answering service and no software, and nobody here plumbs. The free 20-minute call says plainly when the Review is not worth it.