What a missed arrival and a 'when will my technician arrive?' call really cost

Short answer

The cheapest text is not the lowest-cost customer update. A text fee is one visible line. A silent missed arrival costs recovery labor, extra field time, a concession and some chance of a lost review or referral. Price one incident from your own records, cost the arrival-status calls that eat office hours, and compare both with the monthly cost of the fix. The honest result is a break-even count and a payback range, not a promise of savings.

All dollar figures below are worked examples built from round numbers. They are not benchmarks. Replace every input with your own.

Price one silent missed arrival

A homeowner takes an afternoon off for a repair. The quoted window is 1:00 to 3:00. At 3:15 an earlier job is still running and no update has gone out. The customer wonders whether anyone is coming, the CSR takes the call instead of booking the next lead, and the owner may end up in an escalation or a review response.

Some delays carry real pressure. A broken spring, worn cable or balance problem can make a garage door unsafe, and those repairs belong with trained professionals (DASMA).

Message cost and incident cost are different numbers:

Message cost	Missed-arrival cost
Per-message charge	CSR and owner recovery time
Consent and opt-out administration	Extra route or revisit time, when the delay caused it
Setup and monitoring time	Discount, waived fee or refund
	Lost job contribution or reputation exposure

Blank cells are yours to fill in

Price one incident with five inputs:

Recovery labor. Added CSR and owner minutes times each person's loaded hourly cost.
Extra field cost. Paid travel, idle time, a return visit, mileage or route disruption, only when this incident caused it.
Lost job contribution. Gross profit, not the invoice, and only the part not recovered by a reschedule or backfill.
Service recovery. The actual credit, discount, waived trip fee or refund, plus incremental admin time.
Reputation and referral exposure. A probability of a negative review or lost referral times the contribution impact you believe is at risk.

Hypothetical example. A silent missed arrival creates $30 in CSR and manager time, $45 in extra field cost and a $40 concession. The job is rescheduled, so no job contribution is lost. The owner assigns a 10% chance of a $300 reputation or referral impact, which is $30 of expected exposure. Expected incident cost = unrecovered job contribution (if any) + direct recovery cost + probability-weighted downside = $30 + $45 + $40 + $30 = $145.

Do not count the same loss twice. If the original job's profit is truly lost, do not also claim the same open capacity as a separate displaced booking.

The review term is an assumption, but it belongs in the conversation. Consumers say review sentiment, recency, ratings and owner responses affect local-business choices (BrightLocal 2026 survey). Research on service failures has linked faster recovery and compensation to loyalty outcomes, though not for garage door work specifically, so do not turn it into a forecast (PMC).

Set the break-even before you buy

Add the monthly notification cost, including message fees and staff time.
Divide by your estimated avoidable cost per silent missed arrival.
The result is how many incidents a month must be prevented or softened to break even.

If alerts cost $120 a month and a timely alert is expected to prevent or soften $60 of an incident's cost, break-even is two incidents a month. That $60 is not the full $145. It is the portion the alert can reasonably affect.

An alert buys time. It does not fix the day board. It can flag a technician as ahead or behind, send a delay notice, give the CSR time to call before the window expires and record that an update went out. A person still decides what promise is realistic, handles complaints and vulnerable customers, resolves access issues and checks schedule accuracy. Separate the cause: an alert cannot repair a capacity problem, a bad estimate or a dispatch mistake.

Score missed arrivals by cause, not by count

A callback total cannot explain itself. A missed arrival is an incident type, not a root cause. A route change, no access and a safety stop need different fixes. Assign one primary cause, and a contributing factor if needed, from the evidence.

Cause bucket	Qualifying evidence	Common examples
Dispatch control	Promised window, route changes, contact record and update log show an internal miss	Wrong window, late route change, no customer update, bad number left uncorrected
Customer readiness	Job notes and contact attempts show the site or person was unavailable	No access, site not ready, gate locked, homeowner absent
Documented exception	An outside trigger is recorded with discovery time and a timely update	Traffic closure, weather, safety stop, prior overrun with prompt notice

Keep cause separate from outcome. "Rebooked," "canceled," "discounted" and "refunded" say what happened afterward, not why. Use unknown when evidence is missing, because forced coding creates fake precision and points at the wrong fix (ASQ root-cause guidance).

Keep one record per incident: job, booking date, promised window, status changes with owners, the final no-arrival time and outcome; the notification channel, send time, delivery state and reply; access and prep notes; the exception trigger and the time the customer was told; and an evidence-quality tag of complete, partial or missing. Keep text consent and privacy fields separate from performance scoring.

Price incremental make-good work, not the original job twice:

Try it with your numbers

return laboradded truck costincremental parts and materialsthird-party chargesone customer concessionother documented make-good spend

Direct make-good cost
Enter your numbers
Direct make-good cost = return labor + added truck cost + incremental parts and materials + third-party charges + one customer concession + other documented make-good spend

Guardrails: use one loaded hourly rate and do not add burden twice (BLS on employer compensation); use a mileage proxy such as the IRS rate or actual vehicle costs, never both for the same miles (IRS); exclude parts already in the original job; count one concession once; and do not include a warranty reserve and the payment for the same incident in the same period. Show lost capacity separately, and tag each figure actual, estimated or unavailable (ASQ cost of quality).

Cost the arrival-status calls behind the labor variance

If recurring revenue stayed flat and office labor rose, look at the call that barely registers: "When will my technician arrive?" The CSR checks the schedule, sees a window that may be stale, pings the technician, calls the customer back and finds their place again. The phone system shows three minutes. The work took longer.

Track the task chain, not only the inbound call: answer and identify the account, check the route and window, contact the technician when the schedule cannot answer, return the call or send the update, add a note and resume the parked work.

Define the call before counting

Working definition. A technician arrival call is an inbound customer contact whose main purpose is to learn when the technician will arrive or whether the stated window changed. Count the inbound contact once. Track the callback and technician ping as time attached to that contact, not as new calls.

Exclude or tag separately: reschedule requests where timing is not the issue, complaints about a late or missed visit, billing, treatment or account questions that mention the route, and technician-initiated updates with no inbound request. Sample two to four weeks across busy and slow days, routes, branches and shifts, and record the source, exclusions and unusual conditions. Season matters. Pest activity peaks at different times, for example late summer for stinging insects and October for spiders (NPMA), so sample by month and market rather than annualizing a peak week.

The monthly cost model

Input	What to use	Unit
Average daily status calls	Qualifying calls in the sample divided by sampled office days	calls/day
Office days per month	Scheduled office days for the forecast month	days/month
Direct handling time	Talk time plus wrap-up per call	minutes/call
Follow-up time	Schedule check, technician ping, callback and notes	minutes/call
Loaded CSR hourly cost	Wage plus consistent employer add-ons	dollars/hour

Monthly call volume = average daily status calls x office days per month
Minutes per call = direct handling + follow-up
Base monthly cost = monthly call volume x minutes per call / 60 x loaded CSR hourly cost

Use payroll, taxes and benefits from your own books. For context only, BLS lists a May 2025 national median CSR wage of $21.53 an hour (BLS occupation profile) and private-industry office and administrative support compensation of $36.73 an hour in June 2026 (BLS). Neither replaces your loaded CSR cost.

Worked example (round numbers, not an industry average): 12 calls a day, 22 office days, 4 direct plus 3 follow-up minutes, $32 loaded cost. Volume is 12 x 22 = 264 calls. Base cost is 264 x 7 / 60 x $32 = $985.60 a month.

Case
Low
Calls/day
8
Minutes/call
5
Monthly cost at $32/hour, 22 days
$469.33
Case
Base
Calls/day
12
Minutes/call
7
Monthly cost at $32/hour, 22 days
$985.60
Case
High
Calls/day
16
Minutes/call
10
Monthly cost at $32/hour, 22 days
$1,877.33

Case	Calls/day	Minutes/call	Monthly cost at $32/hour, 22 days
Low	8	5	$469.33
Base	12	7	$985.60
High	16	10	$1,877.33

Leave interruption cost out of the base case unless you measure it locally. Research suggests interruptions can slow resumption and affect accuracy, but results vary by task (Frontiers in Psychology 2025, Applied Ergonomics review). If finance tests it, use a local figure such as "add our measured 30 seconds per call."

Forecast the returned capacity without assuming every call goes away

No workflow removes every arrival call. Late technicians, changed routes and customers who prefer a person will keep some alive.

Returned hours = baseline monthly calls x minutes per call x tested reduction rate / 60
Capacity value = baseline monthly cost x tested reduction rate

Case
Low
Reduction rate
25%
Hours returned a month
7.7
Capacity value a month
$246.40
Claim type
Conservative
Case
Expected
Reduction rate
45%
Hours returned a month
13.9
Capacity value a month
$443.52
Claim type
Forecast to test
Case
High
Reduction rate
60%
Hours returned a month
18.5
Capacity value a month
$591.36
Claim type
Only if observed results support it

Case	Reduction rate	Hours returned a month	Capacity value a month	Claim type
Low	25%	7.7	$246.40	Conservative
Expected	45%	13.9	$443.52	Forecast to test
High	60%	18.5	$591.36	Only if observed results support it

Capacity value is not automatic payroll savings. If the CSR stays on payroll, the result is time returned to booking, account help and collections. Claim payroll savings only when a documented payroll expense falls.

Payback range

Net monthly value = capacity value - recurring workflow cost - monthly internal support cost
Payback months = one-time setup cost / net monthly value

Illustrative inputs only, not any product's price: $600 one-time setup, $150 a month workflow, $50 a month internal support.

Capacity value
Low
$246.40
Expected
$443.52
High
$591.36
Workflow cost
Low
$150.00
Expected
$150.00
High
$150.00
Internal support
Low
$50.00
Expected
$50.00
High
$50.00
Net monthly value
Low
$46.40
Expected
$243.52
High
$391.36
Payback
Low
12.9 months
Expected
2.5 months
High
1.5 months

	Low	Expected	High
Capacity value	$246.40	$443.52	$591.36
Workflow cost	$150.00	$150.00	$150.00
Internal support	$50.00	$50.00	$50.00
Net monthly value	$46.40	$243.52	$391.36
Payback	12.9 months	2.5 months	1.5 months

Under those assumptions the range is 1.5 to 12.9 months. If net monthly value is zero or negative, payback is not supported.

Score a vendor against the labor you measured

Before any demo, take two to four weeks of comparable route days, tag only arrival-status, ETA, delay, reschedule and missed-appointment calls, and keep booking, billing, treatment and complaint calls separate. Basic appointment reminders are table stakes. Arrival coverage starts after the booked time stops being true.

Score each row 0 to 3 (0 absent, 1 manual, 2 partly automated, 3 automatic with reviewable proof), multiply by the weight and total it.

Criterion	Strong answer	Weight
ETA updates	Refreshes a customer-facing window as route conditions change	2
Schedule-change alerts	Sends the right update automatically after a delay, reorder or reassignment, with an office review path	3
On-the-way texts	A clear departure trigger plus duplicate and too-early-send controls	2
Message records	Job-level send time, delivery state, message type and office visibility	3
Office-labor measurement	Reporting for coverage, exceptions, call volume and manual touches	3

Logs show recorded send or delivery status, not whether a customer read the message or whether the ETA was right. Define retention, time zones, access and protection (NIST SP 800-92). Test on a packed route that actually slips. Pass only if a delay, reorder or reassignment sends the affected customer a revised window, an on-the-way text sends once at the departure trigger, a CSR can see type, send time and delivery state when a customer calls, and missing numbers, opt-outs, failures, replies and dispatcher overrides are visible, assigned and auditable.

Score only what you see demonstrated in your own workflow. Verify consent handling too: record consent, provide opt-out and make opt-outs visible to the office (FCC, CTIA).

For Queue Up, the product page lists a place-in-line text and estimated window at the start of the day, updates as the schedule shifts, an alert when a technician runs well ahead or behind, an on-the-way text and a log the CSR can review (supabase:tgl-cms.ref_platform.products, slug queue-up). Treat its quote and documented behavior as inputs to your pilot, not as proof of payback.

The 30-day pilot decision

Pick pilot routes and dates before launch and run a 30-day review against the baseline, not a memory of a busy month. Track status-call volume and handling minutes before and after, notification coverage, failures, opt-outs, replies and manual interventions, labor effect separately from software cost, and service-quality notes.

Try it with your numbers

released status-call hours x loaded hourly costmonthly software cost

Monthly pilot value
Enter your numbers
Monthly pilot value = released status-call hours x loaded hourly cost - monthly software cost

Agree the decision rule with the GM before results arrive. For example: continue only if the pilot lowers handling minutes, keeps coverage up and produces exceptions the office can work from. Review at 30, 60 and 90 days with the same call definition, or the comparison is broken before it starts. The first forecast is supposed to be imperfect. Its job is to make the hidden cost visible, and the second forecast should be better because it uses real data.

The weekly scorecard

Show volume and dollars together, with thresholds from your own baseline.

Metric	Formula	Owner
Missed-arrival incident rate	Incidents / scheduled jobs x 100	GM
Preventable missed-arrival rate	Approved preventable incidents / scheduled jobs x 100	Service manager
Average direct make-good cost	Direct make-good cost / preventable incidents	GM
Complete notification-history rate	Incidents with complete logs / all incidents x 100	CSR manager
Rolling cause share	Incidents in a cause bucket over four weeks / coded incidents over four weeks	Service manager

Review the largest expense, the fastest-rising cause and the weakest evidence. Cleaner KPIs make the next miss harder to repeat.