Do proactive ETA updates reduce status calls? A 30-day scorecard
Run a 30-day test to see whether proactive ETA updates cut avoidable status calls without weakening bookings, arrival performance, or customer trust.
The question is not whether texts were sent
September can bring more calls simply because the board is fuller. That is not proof that customer communication got worse. A garage door company with 320 scheduled jobs and 24 status calls has a different problem than one with 160 jobs and 18 calls.
The familiar “Where is my technician?” call is not a moral failure by the CSR. It is a customer trying to close an information gap while a repair, install, or estimate is still open.
Proactive ETA communication may reduce those checks. The reviewed Fieseros and Smart Service software pages say it does. The useful owner-and-GM question is narrower: did proactive ETA updates reduce avoidable status calls without weakening booking work, arrival performance, or customer trust?
Thirty-day test question: Did status calls per 100 scheduled jobs fall after we introduced consistent ETA updates, while our booking and service guardrails held?
Thirty days gives directional operational evidence, not a durable causal verdict. At day 30, make one local decision: keep the workflow, fix it, or stop it; do not treat the result as proof of a lasting causal effect without a concurrent control or holdout.
Set the pass conditions before day one
Do not launch messages and then hunt for a flattering number. Put the customer communication scorecard rules in writing first. No changing call tags, denominator rules, or targets halfway through because Tuesday looked rough.
Use status calls per 100 scheduled jobs as the primary outcome. Set the pass threshold from your own baseline. If your baseline is 8.0 status calls per 100 jobs, you might define success as a sustained drop to 6.5 or lower, with guardrails intact. That number is an example, not a borrowed industry benchmark. No clear, method-sound garage door benchmark should decide this for you.
Your pass/fail checklist should include:
- Primary outcome: status-call rate improves against the comparable baseline.
- Capacity outcome: booking contacts are answered or worked within the service level you chose, and booked-job conversion does not slip.
- Customer guardrails: delayed-job notices are sent when needed, missed arrivals do not rise, cancellations do not rise, and escalations stay within the limit you set.
- Comparison rule: compare matching workdays and note changes in weather, promotions, staffing, job mix, and schedule load.
- Decision rule: define what “keep,” “fix,” and “stop” mean before the first ETA goes out.
Primary formula:
(customer-initiated status calls ÷ scheduled jobs) × 100
Rates matter because raw call totals follow workload. A fuller schedule can create more calls even when each scheduled job creates less service friction. Quality-improvement guidance makes the same point: use time-ordered, normalized measures, and annotate outside influences rather than treating a simple before-and-after total as proof. A systematic review of run-chart use also warns that small samples and unmarked external changes can mislead.
Build a baseline the team can trust
Start with 14 to 30 comparable operating days. Thirty is better if you have clean data. If you only have 14, say so on the scorecard. Unknown is more useful than a made-up estimate.
- Pull daily scheduled jobs. Break them into service, install, and estimate work. Keep emergency calls separate if they behave differently from booked routes.
- Tag inbound contacts. Split calls into new booking, ETA check, reschedule, payment, and other. Decide whether customer SMS replies count as status contacts, then keep that rule fixed.
- Record schedule pressure. Capture same-day and next-day load. A packed board changes both customer expectations and the CSR’s available time.
- Add context notes. Mark weekday, weather, promotions, technician absences, system outages, and unusual demand. A storm week is not a normal comparison week.
- Show totals and rates. Report the daily count, the median daily rate, and weekly rates per 100 scheduled jobs.
A customer who knows a visit is late may still be unhappy, but uncertainty makes waiting worse. In a qualitative study of clinic patients, people described proactive delay information as making waits more tolerable; the study does not establish a call-reduction percentage for garage door service. The research is useful for the mechanism, not the promised result.
Common baseline traps:
- Comparing a high-volume emergency week with a low-volume tune-up week.
- Counting every inbound call as a status call.
- Ignoring work that was never message-eligible.
- Treating a staffing shortage as normal performance.
- Inferring a missing call disposition instead of leaving it visibly unknown or missing.
Bring this baseline to the weekly KPI review. The point is not to make the past look tidy. The point is to give the next 30 days something honest to beat.
Define every metric so the labels stay consistent
One CSR’s “ETA check” cannot be another CSR’s “other.” Use one short disposition guide at every phone station and review a handful of tagged calls during the first three days.
| Metric | Counting rule | Exclusion |
|---|---|---|
| Status call | One customer-initiated inbound contact asking when a technician will arrive, whether they are still coming, where they are, or reporting a missed promised arrival for an open visit | Vendor calls, internal calls, spam, and completed-job follow-ups |
| Repeat status call | A second or later status contact from the same customer and job during the same visit window | A new visit window after a reschedule |
| Booking contact | An eligible new-work phone or web inquiry | Existing-job status, payment, vendor, spam, and internal contacts |
| Booking coverage | Eligible booking contacts answered or worked within the chosen service level | Contacts outside posted coverage or duplicate spam entries |
| Delayed-job notice | A proactive update sent after a known schedule slip | Routine confirmation sent before any known delay |
| Missed arrival | Technician arrives outside the promised window without an accepted update | Customer-requested delay or a window changed and accepted before the miss |
| Escalation | Manager handoff, formal complaint, refund request, or public-review threat tied to an open or recent job | Routine request for an ETA with no escalation |
| ETA message coverage | Eligible jobs with the required ETA message sent | Jobs excluded before launch, such as invalid contact information or documented opt-out |
| Delivery success | Sent ETA messages marked delivered | Messages that failed, were blocked, or have no delivery result |
Keep the job number on the call disposition and the message event. That link lets the GM inspect an exception instead of debating a dashboard tile.
Run the 30-day test without muddying the result
The first three days are for calibration, not for declaring victory.
Days 1–3: check the tags
Have the service manager or lead CSR review a small sample of status-call tags and message records each day. Confirm that job eligibility, call reasons, arrival windows, and opt-outs are being recorded the same way. Fix training problems now, then document the change.
Days 4–30: hold the workflow steady
Set the message timing and wording. Define eligible jobs and exclusions before this stable period begins. At the job level, log whether the ETA message was sent, delivered, failed, delayed, or blocked by an opt-out. Tie each event to the current arrival window or technician status.
Keep messages transactional. Do not tuck an offer into an ETA notice during this test. Preserve consent and opt-out records, and make valid opt-outs visible to dispatch. A valid opt-out prevents further automated ETA messages unless the company has confirmed an applicable legal basis. The FCC recognizes replies including STOP, QUIT, END, REVOKE, OPT OUT, CANCEL, and UNSUBSCRIBE as reasonable ways to revoke consent in covered cases. Its 2024 order also limits an opt-out confirmation to one non-promotional message. Check current federal and state requirements, your message technology, and counsel’s advice before launch.
Sample manual update time instead of asking CSRs for burdensome time sheets. For example, pick two one-hour blocks per week and tally minutes spent giving routine ETA updates. This shows whether time is actually returning to booking talks and real exceptions.
Daily and weekly controls
Run a five-minute daily data check: are jobs counted, calls tagged, and message outcomes present? Then hold one owner-GM review each week.
Do not change phone routing, booking scripts, dispatch policy, or the ETA template during the stable period unless you have to. If you do, annotate the date. Also note storms, technician absences, outages, holidays, promotions, and demand spikes. If possible, compare a similar no-message team, area, or held-out group. If not, be modest about what the result proves.
Use one scorecard for the result and the guardrails
Put the primary rate and the things that can make it meaningless on the same page. Green, watch, and red bands must be set before launch. Do not color a cell green because the meeting needs good news.
| Metric | Baseline | Current week | 30-day result | Target / band | Trend |
|---|---|---|---|---|---|
| Scheduled jobs | [TK] | [TK] | [TK] | Context only | ↑ / ↓ / → |
| Message-eligible jobs | [TK] | [TK] | [TK] | Context only | ↑ / ↓ / → |
| ETA message coverage | [TK]% | [TK]% | [TK]% | Green: [TK]%+ | ↑ / ↓ / → |
| Delivery success | [TK]% | [TK]% | [TK]% | Green: [TK]%+ | ↑ / ↓ / → |
| Status calls per 100 scheduled jobs | [TK] | [TK] | [TK] | Green: ≤ [TK] | ↑ / ↓ / → |
| Repeat status calls per 100 scheduled jobs | [TK] | [TK] | [TK] | Green: ≤ [TK] | ↑ / ↓ / → |
| Booking coverage | [TK]% | [TK]% | [TK]% | Green: ≥ [TK]% | ↑ / ↓ / → |
| Lead-to-booked conversion | [TK]% | [TK]% | [TK]% | Green: ≥ [TK]% | ↑ / ↓ / → |
| Delayed-job notice coverage | [TK]% | [TK]% | [TK]% | Green: ≥ [TK]% | ↑ / ↓ / → |
| Missed-arrival rate | [TK]% | [TK]% | [TK]% | Red: > [TK]% | ↑ / ↓ / → |
| Escalations per 100 scheduled jobs | [TK] | [TK] | [TK] | Red: > [TK] | ↑ / ↓ / → |
| CSR minutes on routine status updates | [TK] | [TK] | [TK] | Green: ≤ [TK] | ↑ / ↓ / → |
Use these formulas beneath the scorecard:
- ETA message coverage:
(eligible jobs with ETA message sent ÷ message-eligible jobs) × 100 - Delivery success:
(delivered ETA messages ÷ sent ETA messages) × 100 - Repeat status-call rate:
(repeat status calls ÷ scheduled jobs) × 100 - Booking coverage:
(eligible booking contacts answered or worked within service level ÷ eligible booking contacts) × 100 - Lead-to-booked conversion:
(booked jobs ÷ eligible booking contacts) × 100 - Delayed-job notice coverage:
(known delayed jobs sent a proactive update ÷ known delayed jobs) × 100 - Missed-arrival rate:
(missed arrivals ÷ scheduled jobs with a promised window) × 100 - Escalation rate:
(escalations ÷ scheduled jobs) × 100
This is a customer communication scorecard, not a text-message report. Coverage tells you whether the workflow happened. The guardrails tell you whether it made the day better.
Read the weekly numbers in the right order
Start with the board, not the headline. Review each week in this order:
- Schedule volume and job mix. Were there more emergency repairs, installs, commercial routes, or same-day calls?
- Message coverage and delivery failures. Low coverage means you did not really test proactive ETA updates.
- Normalized status-call rate. Compare per 100 scheduled jobs, plus the raw count.
- Booking capacity and conversion. A CSR buried in manual updates can miss a new booking call.
- Missed arrivals and escalations. A lower call rate paired with more unhappy customers is not a win.
| Pattern | Likely meaning | Next check |
|---|---|---|
| Raw status calls fall, but scheduled jobs fall faster | False progress may be hiding in lower volume | Compare the rate per 100 scheduled jobs and job mix |
| Total calls rise, but the status-call rate falls | The schedule may be fuller while per-job friction improves | Check booking-contact volume and message coverage |
| Status calls fall, but escalations rise | Customers may be receiving late, wrong, or unhelpful updates | Review message timing, actual arrival times, and escalation notes |
Treat small counts carefully. One angry customer can move a weekly rate on a small board. Read the call notes and job history before you build a big theory. A dated run chart with launch dates and exception notes is more useful than four isolated weekly cells; time-ordered charts help separate ordinary variation from special events when they are properly built and annotated. That is the practical lesson from the review of statistical process control studies.
Make a keep, fix, or stop decision on day 30
At day 30, review a sample of tagged calls, message histories, missed arrivals, and escalations. Test the labels before judging the people who used them.
| Decision | What the scorecard says | What to do next |
|---|---|---|
| Keep | Status-call rate improved, message coverage was adequate, and booking and customer guardrails held | Set the next 30-day target around the biggest remaining verified gap |
| Fix | The rate looks promising, but coverage, delivery, timing, or one repairable guardrail is weak | Repair the specific break, document it, and rerun a stable 30-day test |
| Stop | Adequate coverage produced no useful change, or guardrails show clear harm | Stop the workflow, preserve the findings, and investigate another cause of status contacts |
Document the owner-GM decision, the metric owner, the next review date, and the exact rule you will change, if any. That keeps the test from becoming a dashboard artifact nobody owns.
If manual status work is proven to be the drain, a workflow such as Queue Up notifications can be relevant because it preserves job-level message history for audits and exception review. The point is not to remove the CSR. It is to return routine update time to booking conversations and the customer situations that actually need a person.
The cleanest final question is still the first one: did customers need fewer checks, while the service promise held? Your garage door scheduling metrics should answer that with evidence, not optimism.