Operations
Queue Management KPIs: What a Walk-In Desk Should Measure
Four numbers for a walk-in Ticket line: average wait, longest wait, time at the Counter, and missed or cancelled Tickets—and what Esperaly Insights actually shows.
· Esperaly Editorial · 9 min read
A walk-in desk does not need thirty charts. Queue management KPIs are the few numbers that should change a staffing or Service decision next week: how long people waited to be called, how bad the worst wait was, how long the Counter was occupied, and how many Tickets were missed or cancelled. Everything else is optional until one of those four is wrong.
This is a different job from estimating wait times before you buy and from deciding how many Counters you need. Those guides answer “what will the line do?” This one answers “what did the line do, and what single change do we make?”
Esperaly’s model stays Ticket, Counter, and Service. A visitor joins a named Service, gets a Ticket, and staff call the next Ticket at a Counter. The numbers below are measured on that loop. They are not a contact-center scorecard, a clinical outcome, or a ranking of queue products.
The four queue management KPIs worth a weekly review
Review the same four figures, for the same place, on the same weekday range. Mixing Saturday with a quiet Tuesday hides the hour that actually hurts.
- Average wait — from joining the queue to being called. This is the number people quote. It is also the number that flatters you.
- Longest wait — the tail the average hides. If the average is 8 minutes and someone waited 40, the board’s promise was not the experience at the back of the line.
- Time at the Counter — from being called to finishing. Wait gets the complaints. Handling time is usually the lever.
- Missed and cancelled share — called and never served, or left before service, divided by Tickets created. A rising share means the wait or the calling ritual is failing, not that you need a new logo on the Ticket.
In Esperaly, Insights shows those timing cards when the desk actually calls Tickets: average wait, longest wait, and time at the Counter, with a comparison to the previous period. Created, completed, missed, and cancelled counts sit beside them. A check-in list that never calls anyone has no wait to measure, so those cards stay empty on purpose.
Average wait, and why the daily number lies
Average wait is total minutes from join to call, divided by Tickets that were actually called. People who leave before a call never enter that average. If you only stare at it, abandonment looks like a calm room.
A made-up Tuesday, so the arithmetic is obvious: 40 Tickets are called. Thirty waited 4 minutes. Ten waited 26 minutes. Total wait is 380 minutes. Average wait is 9.5 minutes. Half the room would describe the morning as fine. The ten people at the back would not. Publish the average internally if you want, and always write the longest wait next to it.
Split before you react. One Service that runs long will drag a shared daily average even when the other Services were fine. Insights breaks Tickets down by Service. The location home shows arrivals per hour today against yesterday, so you can see whether the pain was a single hour. A high daily average that is really “11:00 to 12:00 on permits” is a roster problem, not a reason to rebuild the lobby.
There is no universal “good” wait. A short retail question and a licensing conversation should not share a target. Set the number you are willing to staff, using the wait-time estimator, and treat published industry charts as someone else’s lobby. Your own busy hour is the baseline.
Longest wait stands in for a service level
A service level is a promise with a shape: “90% of Tickets called within 15 minutes,” not “we averaged 9 minutes.” Averages hide the 10% who had a bad visit. The promise forces you to look at them.
Esperaly does not show a built-in “percent within N minutes” card. Insights shows the longest wait in the range, which is the blunt version of that tail. Use it. If you later want a true percentage, sample called Tickets from the range and count how many beat your N. Do not buy a second analytics product on day one for a percentile you have not defined.
Pick N from your own tolerance, then hold it still for a month:
- Write the target where staff can see it. “Called within 15 minutes on the short Service” is a target. “Be faster” is not.
- Use a longer N for a Service you already know is a long visit. The failure is pretending a 25-minute conversation is a 5-minute one.
- If the longest wait is routinely double your N, the average can stay flattering while the floor is not. That is the week you change capacity or split the Service. It is not the week you redesign the Ticket prefix.
The pre-buy wait guide has example bands for short desks versus long visits. Those are planning examples, not standards Esperaly certifies.
Time at the Counter
Time at the Counter is how long the desk is occupied per Ticket, from call to done. Include the messy minutes: printing, “one more question,” walking someone to another window. A stopwatch on the clean part of the visit understates the line.
Read it by Service before you read it by person. If pickup averages 4 minutes and a full application averages 22, one shared line makes the short visits wait behind the long ones. That is a Service design problem, covered in when to split queue Services. If the same Service varies wildly between Counters, the script or the training is uneven. Call-next training is the fix. A break-room ranking of who closed the most Tickets is not. Insights can break closed Tickets down by staff. Use that to spot a Counter that is stuck, then talk to the person. Do not publish a speed ladder.
Utilisation maths — how close a Counter is to being busy all hour — lives in the Counter staffing guide. This weekly review only needs the direction: handling time up, longest wait up, same arrival rate, means the visit got heavier or the calling got sloppier. It does not, by itself, mean you should hire.
Missed and cancelled Tickets
Missed means staff called the Ticket and nobody was served. Cancelled means the person left before service. They are different failures. A missed Ticket is often a calling problem: the display was late, the name was wrong, or the person stepped out and the Counter moved on. A cancelled Ticket is often a wait problem: they saw the line and quit.
The rate is the count divided by Tickets created in the same range. Keep the denominator. “11 missed” means nothing until you know whether the desk created 40 Tickets or 400.
Ignore percentages on tiny slices. Under about 20 Tickets in a slot, one family that walked out moves the rate by several points. Look at the same weekday over two or three weeks before you declare a trend. The tactics for cutting ghosts — visible waits, predictable calling, a fair way back — are in reducing walk-in queue no-shows. This section is only how you notice the rate moved.
What to watch on the floor versus once a week
The location home is for the next ten minutes, not the archive. Waiting now, the current wait estimate, waiting by Service, and the missed list still open tell a supervisor whether to open another Counter before lunch. Service demand compares today’s arrivals per hour with yesterday. That is a live cue. It is not a quarterly report.
The weekly pass is Insights, with compare turned on:
- Tickets created, against the previous period.
- Average wait and longest wait.
- Time at the Counter.
- Completed, missed, and cancelled, then the same split by Service.
If created jumped and handling time did not, you had more people, not slower visits. If created was flat and longest wait jumped, something on the desk changed: a Counter closed, a Service got harder, or calling slipped. The home screen’s hourly chart is how you find which hour.
Be straight about the limit. Insights is operational visibility for one location: counts, timing, and breakdowns by Service, Counter, and staff. It is not a 90th-percentile engine, a contact-center wallboard, or a BI suite. The laptop full of marketing tiles on a stock photo is the thing you do not need in order to run a fair line.
A 20-minute Monday review
One desk, one range, one decision. Here is a worked sheet you can copy. The figures are an illustration, not a customer result.
- Created: 220 Tickets in seven days
- Average wait: 9 minutes (previous period 8)
- Longest wait: 41 minutes (previous period 22)
- Time at the Counter: 7 minutes (unchanged)
- Missed: 11 (5%). Cancelled: 18 (8%)
Handling time did not move. The tail did. Open the Service breakdown. If one Service owns the 41-minute wait, read when to split Services before you add a person. If every Service blew out in the same hour, the staffing guide and the wait-time estimator tell you whether one more Counter at that hour would have held the tail. Price the payroll against the queue cost calculator. Queue software at $19 per location per month after the trial is rarely the expensive part of that decision.
Rules that keep the review honest:
- Same range every week. Last seven days versus the seven before.
- Change one thing. A new Counter hour, a split Service, or a calling fix. Not all three.
- Write the four numbers down before you change anything, including the week you think you already know the answer.
- Re-read them next Monday. If the longest wait is still double your target, the change did not work. Try the next single change. Do not add a dashboard.
What these KPIs will not tell you
- Whether someone liked the visit. A short wait and a rude Counter can coexist. These figures do not replace talking to people.
- Clinical results, charts, or billing. A clinic can run the same four numbers as a council desk. The review is not a medical record and it does not belong in one. Healthcare and government desks use the same Ticket loop; the Service names differ.
- How full a booking calendar is. Some Services also take booked arrivals, and check-in still becomes a Ticket. This review measures the line after someone is in it. It is not a calendar report.
- Which queue product is “best.” If you are still choosing a system, use the pricing guide and the feature list. Do not shop from a KPI leaderboard.
Measure the desk you already run
Open Insights on the location you actually staff. Turn on compare. Write average wait, longest wait, time at the Counter, and the missed and cancelled shares. Make one change. Read the same four numbers next week.
New workspaces are paused. Register your interest and we will email you when signup opens. Existing accounts can still sign in. The price, when signup is open, is a 14-day trial that requires a card, then $19 per location per month. Counters, staff, kiosks, and displays at that location are included. Cancel before the trial ends and you pay nothing. Details are on pricing.