Short answer. A queue management system does four things. A visible order: each person takes a ticket, per service, and calling is done by number, on screen and by voice. An announced wait: the ticket gives an estimate, and the phone gives notice when the turn is near. A queue you can leave: people wait somewhere other than the room. Figures: arrivals per hour, service time, wait, to decide the number of counters. In this page's example, at 28 arrivals an hour, going from three to four counters brings the average wait down from 26 minutes to 2.
This page is for those who run a place that receives the public: branch, practice, clinic, laboratory, after-sales service, company or administration counter, one site or several. The definitions come from public sources: queueing theory, queue area, virtual queue, first come first served, the Erlang C formula, the multi-server queue, Little's law, the Poisson distribution, service level. The branch, its arrivals and its counters are a constructed example, and the waiting times in it are calculated, not measured. The page covers neither regulatory obligations on reception nor tax, and quotes no hardware price.
What is inside a queue management system
The figure stacks five layers; the last is the one that justifies the project.
Entry. A queue area is a place where people queue, first come, first served, for goods or services. Physical queueing is sometimes replaced there by virtual queueing: the person uses their phone, the internet, a kiosk or another method to join, or takes a numbered ticket from a machine. Services. Not everyone comes for the same thing: collecting a document takes two minutes, opening a file twenty. One queue per reason, an average duration per reason: that is what makes the estimate possible. Counters. The agent has a simple screen: “next”, absent, transfer to another service, break.
Calling. A display in the room and a voice announcement, in Arabic and in French, see the application in Arabic and in French; and a message on the phone, through the channel that actually arrives, see notifications. Measurement. That is the accented layer, detailed below. Without a connection. Entry and counters keep working on the local network when the internet drops, see the offline application. The phone queues too. Incoming calls have their own queue, covered in the telephone switchboard; queueing theory in fact comes from there, from the models Agner Krarup Erlang created to describe the incoming calls of a Copenhagen telephone exchange.
The life of a ticket
The figure follows a ticket in six steps.
First come, first served. The rule is that people are served one at a time, and that whoever has been waiting longest goes first. The ticket makes that rule visible: nobody has to watch who was there before. The estimated wait. The ticket shows it: number of people ahead, likely duration. People experience “occupied” time as shorter than “unoccupied” time, and generally overestimate how long they waited by around 36%: announcing a duration, and letting people do something else, shortens the wait as it is felt without touching the real one. The call. Number and counter, on screen and by voice; and a message when only three people are left.
The no-show. Called, not there: the agent marks it, the ticket is called once more later, then closed. Those who leave. The theory distinguishes two behaviours: deciding not to join the queue because it is too long, and leaving it after waiting too long. The system counts the second, tickets taken and never served, and it is one of the most useful figures: each departure is a lost visit. Four timestamps. Ticket, call, start, end: from them come the wait, the service time and the occupancy of each counter.
Five ways of organising the wait
The figure sets side by side five ways of organising the wait.
The virtual queue. The idea is used in call centres and in other businesses to improve waiting: the person no longer has to stay on the line, and is free to do something else. Applied to a room: you scan a QR code at the entrance, or send a message, and wait at the café opposite, in your car or at your desk. The room empties without a single counter having been added. The appointment. For a service that is long or has to be prepared, the slot remains the right answer; its mechanisms are described in online booking. Most places mix the two: some counters for appointments, others for walk-ins.
Priority. In a priority queue, high-priority people are served first. An emergency in a clinic, a fasting sample in a laboratory, an account customer in a branch: the rule is decided by you, written down, and displayed, so that nobody sees it as a favour. The kiosk. A tablet on a stand is often enough; the principle is the same as for a restaurant's ordering kiosk: few choices, large buttons, two languages. Several sites. The same services, the same figures, compared from one branch to the next.
When the queue forms
The figure follows a day's arrivals, hour by hour.
Capacity. Six minutes per person: one counter serves ten people an hour, three counters thirty. Arrivals. In the example, 192 people in the day, but 38 between 10 and 11 am and 34 in the following hour. During those two hours more people arrive than the counters serve: 8 people are waiting at 11 am, 12 at noon, and the last to arrive waits 24 minutes. The queue only clears at lunchtime, when arrivals fall back to 16. The fourth counter. Open from 9 am to 12 pm, it raises capacity to 40 an hour: in this calculation, the queue no longer forms.
What that requires. Three hours of one person, not a whole post; you still have to know which hours. Without measurement, you see a full room and overwhelmed agents; with tickets, you read a curve per hour, per day of the week and per period of the month, on a real-time dashboard. Spreading. The other lever is on the arrivals side: offering appointments in the afternoon, announcing the quiet hours. This calculation is optimistic. It assumes regular arrivals. In reality they come in bunches, and that is the subject of the next figure.
«The day's total says nothing; the hour says it all.»
One more counter
The figure takes an hour with 28 arrivals, with three, four or five counters.
The model. Queueing theory is the mathematical study of queues. Its multi-server model assumes random arrivals, at a constant mean rate and independent of one another, which is what the Poisson distribution describes. It requires the counters' utilisation, arrivals divided by capacity, to stay below 1 for the queue to be stable. The Erlang C formula then gives the probability that an arriving person will need to queue. The calculation. At 28 arrivals an hour for a capacity of 30, three counters are 93% busy: 88% of people wait, 26 minutes on average. With four counters, 70% busy: 2 minutes. With five, 31 seconds.
Why such a gap. Close to saturation, the slightest bunch of arrivals creates a queue that the counters no longer have the margin to catch up. Little's law links the quantities: the average number of people present equals the arrival rate multiplied by the average time spent. 28 people an hour waiting 26 minutes means 12 people waiting on average; at 2 minutes, just one. A benchmark. With four counters and 34 arrivals an hour, that is 85% busy, the same calculation gives 7 minutes; at 38 arrivals, 95%, it gives 27. Beyond 85%, the wait takes off. The limits. The formula assumes that nobody leaves the queue and that the pace lasts: it is an order of magnitude, which your own figures replace from the first month. The target. It is set as a service level: the share of people who wait less than a given time.
«A counter that is 93% busy looks well used; it is the regime in which the room overflows.»
Eight weeks for one site
Eight weeks in the example, for one site.
- What you receive: entry by kiosk, QR code or message; queues per service with priorities; the agents' screen; calling in the room, on screen and by voice, in Arabic and in French; messages before the call; a dashboard per hour, per service and per counter.
- What the place avoids: arguments about the order of arrival, the question “how long is left?”, a room full of people who could wait elsewhere, people leaving without being seen, the counter opened at the wrong hour.
- What is measured: arrivals per hour and per service; average wait and share of people served in under ten minutes; service time per reason; no-shows at the call; tickets taken and never served; occupancy of each counter.
- What Symloop does: scoping the services, entry and tickets, the agents' screen, calling, the virtual queue and its messages, the dashboard, the link with your appointments and your business software, for one site or several.
Scoping. One week: the list of reasons for visiting, a few services timed, the number of counters, the hours everyone knows are busy. Equipment. A tablet or a kiosk at the entrance, a display in the room, the agents' workstations: the software adapts to what you already have, and a printed QR code is enough to start the virtual queue. The trial. A quiet day first, with the agents, before the first peak day.
Ready-made or custom. Ticket systems exist, sold with their kiosk and display, and suit a simple room with a single queue. Custom work is justified when you need several services and several sites, a mix of appointments and walk-ins, notice by message, or a link between the queue and the person's file in your business software. What does not change. A queue system does not create capacity: it shows where it is missing, and at what time. The decision to open a counter remains yours; it becomes a costed one. The durations are those of a constructed example.
- What is a queue management system?
- Software that gives each person a number, per service, calls them to a counter by screen, voice or message, and records the time of each step. It brings order to the room and produces the figures of the reception.
- How do you manage a queue?
- By making the order visible with a ticket, announcing a duration, letting people wait elsewhere, and opening counters according to measured arrivals, hour by hour.
- How do you reduce waiting time?
- With three levers: opening one more counter in the hours when arrivals exceed capacity, spreading arrivals through appointments in quiet hours, and shortening the longest services by preparing them.
- How many counters should be open?
- Enough for utilisation to stay clearly below 100%. In the example, at 28 arrivals an hour and 6 minutes per person, three counters give an average wait of 26 minutes, four give 2.
- Is a ticket kiosk needed?
- Not necessarily. A tablet at the entrance or a printed QR code is enough to begin. The kiosk becomes useful when the flow is heavy or the public does not always have a phone.
- What is a virtual queue?
- A queue in which you keep your place without staying in place: you sign in by QR code or by message, follow your position on your phone, and come back when your turn is near.
- How is the waiting time shown on the ticket estimated?
- From the number of people ahead, per service, the average duration measured for that service and the number of counters open. The estimate sharpens with each week's figures.
- Can appointments and walk-ins be mixed?
- Yes, and it is the most common case: some counters are reserved for appointments in certain time bands, the others serve the queue, and the rule is displayed.
- Does the system work without the internet?
- Entry, calling and counters have to carry on over the local network; messages and the remote dashboard resume when the connection returns.
- What does a queue system measure?
- Arrivals per hour, the wait, service time, no-shows at the call, people who left before being served and the occupancy of each counter, per service and per day.