The phones roll to voicemail at five o’clock.
A made-up residential HVAC, plumbing and electrical company with about sixty technicians, five dispatchers and four comfort advisors, serving one metro area. Dispatchers answer the phone from 7 a.m. to 5 p.m., and after that every call rolls to voicemail.
The owner brought one question to the intro call: which workflow costs us the most? Everyone in the office named missed calls first. Nobody had counted them.
The baseline puts a month of calls on the record.
Before anyone talked about a fix, the office pulled four weeks of records from the phone system and the CRM and scaled them to a month. Each figure carries how sure the team was of it.
| What was counted | How sure | Baseline |
|---|---|---|
| Calls from new customers | Measured | 900 |
| Calls that reached voicemail | Measured | 162 |
| Of those, after hours or at the weekend | Measured | 110 |
| Median time to call a voicemail back, business hours | Measured | 47 min |
| Median time to call a voicemail back, after hours | Measured | 13 hrs |
| Web inquiries | Measured | 140 |
| Median first reply to a web inquiry | Measured | 3 hrs 10 min |
| Callers answered live who booked a job | Measured | 58% |
| Voicemail callers who booked a job | Estimated | 31% |
| Web inquiries that booked a job | Measured | 30% |
| Average ticket on a booked job | Measured | $1,150 |
| Gross margin on a booked job | Estimated | $520 |
| Voicemail callers who rang a competitor next | Guess | about half |
Measured means read from a system’s own records. Estimated means counted by hand or taken from last year’s books. A guess is written down so nobody forgets it, and it stays out of the math.
The voicemail booking rate is an estimate because nobody had tied callbacks to jobs. The dispatch lead matched them by hand for two weeks. The margin is the owner’s figure from last year’s job costing.
The readout sized the value at stake as a range.
The owner and the service manager answered about twenty minutes of questions on a call. The owner then connected the phone system’s call log and the CRM, read-only, after seeing exactly what would be read. StrikeArc Scout scored the plays in the library for fit against those answers, and three rose to the top: P07 Respond to new leads faster, P08 Get quotes and contracts out faster, and P12 Find the answers your team needs.
Before the 45-minute readout, one of our people sized the value at stake for the top play and signed every figure. The readout showed it as a low and a high figure, because two of the inputs were estimates.
The high case stops short of the 58% that live calls book at, because a text back is a weaker first touch than a person picking up. The guess about competitors stayed out. So did the web inquiries: the readout timed them and put no dollar figure on them.
The readout’s do-not-build list had three lines, each with its reason.
- A voice system that answers every call live. The owner wanted a person on every emergency, a gas smell or no heat with a child at home, and a text back closes most of the gap with less to go wrong. It reopens if after-hours voicemail callers still book below 40% after a quarter of text backs.
- Dispatch scheduling. Crews and routes change by the hour, and nothing in our library fits. It reopens on the custom path if dispatch becomes the workflow that costs the most.
- Estimates for replacements, for now. P08 scored second. One play goes in at a time, measured against its baseline, so it waits for the second quarter.
The owner decided on the call to start with P07.
The install ran gate by gate.
The play went into the phone system and the CRM the office already used. The dispatchers had nothing new to learn, and nothing went live until the last gate.
IT, here the office manager, gave read access to the call log and the CRM. Write access covered two things only: the CRM’s lead records and texts from the office number. The dispatch lead was named the one operator.
The service manager wrote down what counts as urgent, which ZIP codes the company serves and who is on call each night. The owner approved the after-hours policy.
The service manager approved, word for word, the three texts the play may send: a missed call in business hours, a missed call after hours with the on-call options, and a web inquiry. Anything else goes to a person.
Last month’s missed calls and web forms ran through the play before any real caller was touched. Every test lead was answered, scored and routed once. The dispatch lead checked the scores and corrected two urgency rules.
The owner approved the launch with the baseline attached, so the first quarter would be measured against the table above.
The first missed call after launch got its text at 6:12 p.m. and reached the on-call technician with a summary. Every lead, score and touch now sits in the dispatch lead’s queue.
The dispatch lead’s time across the install came to about two and a half hours, most of it on the test leads. IT’s part was granting the access and checking it.
The first quarter shows up in the ledger.
The play reports answer time, booked jobs and revenue per lead. Here is the first quarter against the baseline.
| What the play reports | Baseline | Quarter |
|---|---|---|
| Median first reply to a missed call, business hours | 47 min | under 1 min |
| Median first reply to a missed call, after hours | 13 hrs | under 1 min |
| Median first reply to a web inquiry | 3 hrs 10 min | 4 min |
| Voicemail callers who booked a job | 31% | 46% |
| Revenue per new lead | $575 | $602 |
The cost to run comes from your own invoices and your team’s time records, so the example leaves those two lines blank.
The ledger counts margin, the figure the readout sized. The same 73 jobs brought in $83,950 in revenue. A quarter at this rate runs to about $152,000 a year, inside the readout’s range.
A person approved every step that mattered.
Every gate is a named human. Agents draft everything; they approve nothing. In this example the names are roles.
- The owner approved the after-hours policy, the launch and the quarter’s ledger.
- The service manager approved every word the play may send, and the rules for what is urgent.
- The dispatch lead works the queue. Urgent leads after hours reach the on-call technician, and any lead the play cannot score goes to a person with the voicemail attached.
- The Value & Deployment Specialist and the owner signed each figure in the ledger before it was reported.
Some things did not change, for good reasons.
- Web inquiries booked at the same rate. Replies went from three hours to four minutes, and bookings stayed near 30%. Most web inquiries asked for a price on a replacement, and a fast reply without a price left that question open. That is the estimate problem, which is why P08 is next.
- Calls answered live booked as before. The play never touches a call a dispatcher picks up, so the 58% stayed where it was.
- Dispatch still runs on the board. It was on the do-not-build list, and it stayed there.
- The guess stayed a guess. Nobody can count the callers who rang a competitor, so the ledger counts only jobs that were booked.
Run the same method on your own numbers.
The baseline above took four weeks of records and two weeks of matching by hand. You can start the same count today.
- The lead response audit times a week of your inquiries: the median reply, the slowest, after hours and by channel.
- The Workflow Brief writes the workflow down on one page: who owns it, the steps, the systems and today’s numbers, each marked measured, estimated or a guess.
Bring both to an intro call, and the readout starts from your numbers. The guide to speed to lead covers what slows the first reply and the ways to close the gap.
This example runs on one play.
The play in this example is P07 Respond to new leads faster. It reports answer time, booked jobs and revenue per lead, the same lines as the ledger above.