A weekly scorecard for catching appointment-window failures early
Use this weekly owner-GM scorecard to spot missed appointment windows, weak delay notices, and schedule strain before they turn into bad reviews.
The complaint starts before the review
In the weekly owner-GM standup, revenue can look healthy while appointment windows quietly fail underneath it. The board was full, technicians were busy, and jobs closed. That is not the same as customers being kept informed.
Watch the September board, or any period when your own booked hours climb and long installs stack up. Do not assume the season caused the misses. Check the pattern in your own history first.
The risk event is a passed appointment window plus silence. A technician running late is an operating problem. A customer waiting past the promised end time with no useful warning is where that problem becomes a garage door reputation risk. Reviews show up beside a Google Business Profile in Search and Maps, so they are part of the local buying decision before a caller ever reaches your CSR. Recent-review research also found that 74% of consumers look for reviews from the past three months.
This scorecard is not for replying to a bad review faster. It is for seeing the conditions that produce one while there is still time to fix next week’s schedule.
The five numbers to put on the weekly scorecard
Use five measures, each shown as a count and rate. A clean 4% can mean two jobs or 40.
Lock the cohort first. Each week, include every appointment that had an original promised window at the daily schedule lock. Measure arrival against that original end time, even if the job is later moved or its promise changes. Report cancellations, reschedules, and after-lock additions separately as visible exceptions; do not remove them to improve an arrival rate. Same-day emergencies belong in their own cohort.
| Metric | Calculation | Evidence |
|---|---|---|
| Passed-window rate | Locked-cohort jobs arriving after the original window end ÷ locked-cohort appointments | Original promise; arrival time |
| Warning-event coverage | Passed-window jobs with a recorded delay notification event before window end ÷ passed-window jobs | Notification log |
| Median notice lead time | Minutes from first recorded delay event to original window end | Notification log; appointment |
| Schedule pressure | Committed booked hours ÷ net available field hours | Schedule, leave, training |
| Reputation spillover | Passed-window jobs with a tagged complaint or low rating ÷ passed-window jobs | Notes, tags, reviews |
A queued event proves an internal handoff, not that a customer received or read a warning. Track sent, delivered, and read states separately when records support them. A notice after the original window end is not a warning; keep negative lead times visible.
For schedule pressure, exclude proposed and canceled bookings, consistent with Microsoft’s utilization guidance. Review the current week and four-week trend by branch, weekday/window band, technician/route/service area, job type, and after-lock changes. Keep planned service, emergencies, multi-hour installs, cancellations, and reschedules visible as separate cuts.
Read the pattern, not just the average
On-time arrival is arrivals within the booked window divided by the locked cohort, a common field service measure. The cuts show whether the issue is a late route, an overrun install, or an emergency insertion—not merely a weak average.
Diagnose the failure before changing the process
Classify the pattern before acting:
| Pattern | Signal | Inspect / act |
|---|---|---|
| Execution | Misses cluster by route, technician, job type, or day while pressure is normal | Travel, duration, parts, sequence; coach or reroute |
| Capacity | Misses rise across crews with high pressure | Overruns, additions, leave, buffers; reset capacity |
| Communication | Notice events are late or missing despite modest lateness | Trigger, timestamp, owner; repair handoffs |
| Mixed | High pressure and weak notice events | Fix load and notice discipline together |
Bring job-level records: original promise and arrival; notice timestamps, triggers, and states; after-lock changes; planned versus actual long-install duration; customer replies, complaints, reviews; and a reason for each exception. Test one named action for two weeks. One weather day or breakdown is an exception, not a staffing or messaging rule.
Use the notification log as evidence, not decoration
The log should reconstruct the promise, known delay, notification event, and arrival. Appointment-experience research supports timely reminders and delay updates, not a claim that every late notice causes a bad review.
Queue Up can make queued, delay, and on-the-way events visible. It supports the CSR; it does not replace human handling of exceptions.
Set decision rules before the numbers turn red
Assign an owner to each metric and set green, watch, and act ranges from your own baseline.
- Watch: A cut worsens for two weeks: assign a check on routes, duration, or notice timing.
- Coach: Misses stay narrow to one route, technician, or job type: make a job-level planning or coaching change.
- Reset capacity: Passed-window volume rises with pressure and after-lock additions: review limits, buffers, and crew availability.
- Audit messages: Warning-event coverage falls or lead time turns negative: inspect triggers, ownership, and the timeline.
- Escalate: Passed-window volume rises while warning-event coverage falls: set a follow-up date and expected metric movement.
At next week’s standup, review the prior action before new exceptions. End with: What did customers know, and when did they know it?