What Is a Call Queue?

What a queue is, and why it is called ACD

A call queue is a component of the PBX that holds incoming calls until someone is available to answer them. Without a queue, a caller who arrives when everyone is busy hears a busy tone or reaches voicemail, and usually calls another business. With a queue, the caller hears "All agents are busy, your call will be answered shortly," and waits in order.

In the call center world this is called ACD, Automatic Call Distribution. The name emphasizes the second part of the job: the queue does not just hold calls, it also distributes them. It knows which agents are free, which are on a break, and who has already answered ten calls and who has answered two, and it decides who gets the next call.

A queue is the heart of every call center, but it is not only for call centers. An office with three secretaries, a donations line on campaign night, a collections department — anywhere calls arrive faster than they can be answered.

What happens to a call in a queue — step by step

1. Entry. The call reaches the queue from an IVR menu ("For service, press 2") or straight from the number.
2. Opening message. The caller hears that the call was received and that they are waiting for an agent.
3. Finding an agent. The queue checks who is logged in and available, and rings them according to the distribution method.
4. If no one is available — the caller waits, hears music and announcements, and is told their place in the queue from time to time.
5. Callback offer. After a set wait time, the caller can be offered the option to hang up and get a callback.
6. Answer. An agent becomes free, takes the call, and sees in advance which number it is coming from.
7. Exit without an answer. If the maximum time has passed, or no agent is logged in, the call continues to the fallback destination: a voicemail box, another queue, a message, or an on-call person's mobile.

Distribution methods — and when each one fits

This is the most important decision when setting up a queue. It determines who works hard and who sits on the sidelines.

Example: A nonprofit office has two secretaries and a manager. With "Ring all," the manager would hear a ring on every call — a nuisance. With "By priority," the secretaries are first priority, and the manager gets a call only when both are on calls. Exactly what is needed.

Another example: In a sales call center with eight agents who earn commission, "Fixed order" would cause fights — the first on the list would get the most calls. "Longest idle" distributes fairly.

What the caller hears

The wait is part of the service, not a break from it. A caller who hears silence thinks they were disconnected. A caller who hears the same message every ten seconds gets annoyed. The balance:

• A short opening message — "Hello, all agents are busy at the moment. Your call will be answered in the order it was received."
• Music — calm, and not too loud.
• Interim messages — about once a minute, no more often. You can use them for useful information: "Did you know you can check your order status on the website?"
• Place in the queue — "You are third in line." This is what reassures waiting callers the most: they know there is progress.
• Callback offer — "To leave your number and get a callback, press 1." It saves the wait entirely.

What else you set in a queue

• Maximum wait time — after it, the call moves to a fallback destination. Without a limit, callers may wait twenty minutes.
• What happens when no agent is logged in — for example after work hours, when everyone has left the queue. It is better to send the call straight to voicemail than to leave a caller waiting for no one.
• Ring time per agent — how many seconds the call rings at one agent before moving on to the next.
• Break between calls — a short time after each call during which the agent does not receive a new call, so they can wrap up the previous one.
• Alerts — a message to the manager when a customer waited and hung up, or when a call went unanswered.

Heavy-load scenarios: campaign night and busy season

Campaign night at a nonprofit. Between eight and ten in the evening, hundreds of people call to donate. The regular team is three people. The solution: one large queue, and thirty temporary extensions for volunteers sitting at home, at the call center, or on their mobile. The distribution method is longest idle, so every volunteer gets calls. A short wait message that includes a key for a callback, so no donor hangs up in frustration. The next day, the temporary extensions are deleted.

Busy season in a business. A store that gets three times as many calls in the weeks before the holidays. You can add the warehouse workers to the queue at second priority: they get a call only when the salespeople in the store are busy. When the season ends, remove them from the queue with one click.

Daily peak hour. Most businesses have one hour when the load doubles — in the morning, or in the afternoon. It is worth checking the call log to see when that is, and scheduling agents' breaks outside that hour.

Common mistakes

• A queue with no fallback destination — when there is no agent, the caller waits forever.
• An agent who went home and stayed logged in — the queue keeps ringing them. That is why it is important for agents to leave the queue at the end of the day.
• A message that repeats every ten seconds — turns a two-minute wait into a nightmare.
• A distribution method that does not suit the team — "Ring all" in a team of ten agents creates noise and pressure.
• Not looking at the data — how many hung up while waiting, how long they waited. Without it, you don't know whether you need another agent.

Queue or hunt group?

A hunt group rings several extensions, and if no one answers, the call moves on. A queue holds the caller until someone becomes free, with music, place in line, and callback. For a small team where someone is usually free, a hunt group is enough. When there is heavy load, when you don't want to lose any call, or when you want to see data — a queue.

A campaign night, minute by minute

To really understand a queue, let's follow a fundraising night at a fictional nonprofit. The regular team: three employees. For the night, 25 volunteers were recruited, some at home with an app, some with a SIM extension on their mobile, and five at the office with desk phones. The "donations-campaign" queue was set to "longest idle" distribution, ring time per agent of 15 seconds, a callback offer after 90 seconds, and a fallback destination — a voicemail box that asks for a name and number.

19:30. The volunteers join the queue from the control panel. On the queue screen you see 25 agents "available," zero waiting. The coordinator dials in from his mobile as a test: he hears the campaign message, presses 1, and a volunteer's phone rings within two seconds.

20:00. The message goes out in the groups. Within three minutes, 40 calls. 25 are answered immediately; 15 wait. The first caller in line hears "You are first in line," and after 20 seconds a volunteer becomes free.

20:12. Peak: 70 simultaneous calls, 45 waiting, average wait time 1 minute 40 seconds. Three callers press 1 for a callback and hang up; their places in the queue are kept. On the screen, the coordinator sees two volunteers who did not answer a call that rang at their phone for 15 seconds — one stepped out for a break without pausing himself. The coordinator pauses him from the control panel, and the queue stops wasting 15 seconds on him in every round.

20:25. The load drops to 30 waiting. The system dials the three callers who asked for a callback, each when their turn comes, and connects them to a volunteer.

21:10. 8 waiting, 25 volunteers. From the log: 410 calls came in, 372 were answered, 31 hung up while waiting (most of them between 20:10 and 20:20), 7 left a message. Longest wait: 6 minutes. Conclusion for next year: ten more volunteers, between 20:00 and 20:30 only.

23:00. The volunteers leave the queue. The next day the temporary extensions are deleted, and the number returns to its regular routing.

Distribution methods in numbers: the same team, four results

The table above explains the methods. Here is the same scenario, to see what each method actually does: five agents, one hour, 40 calls of about three minutes each.

Two details that matter no less than the method: ring time per agent (15–20 seconds; after that the queue moves to the next, and an agent who didn't answer is considered unavailable for a few seconds) and the break after a call (10–30 seconds for notes). Without the break, an agent who finishes a hard call gets the next one that same second — and after an hour, they answer poorly.

The agent's side: what they see and do

Whoever talks about queues thinks of the caller. But the queue is also the agent's workplace, and most queue problems start with them. A workday for an agent in a queue looks like this:

• Joining the queue. The agent marks that they are available — from the control panel, or a manager does it for them. Until then their phone does not ring, even if ten callers are waiting.
• A call arrives. The phone screen shows the caller's number and the queue name — "Donations" or "Service." The agent knows which line it is before saying "hello."
• Finishing and logging. The short break after the call is for notes and wrap-up. Someone with a CRM system wraps up there.
• Pausing. Going to the restroom, for coffee, for a long outgoing call — the agent pauses themselves. It isn't "slacking off"; it is what keeps the queue from ringing for 15 seconds at an empty chair.
• Leaving. At the end of the shift, they leave the queue. An agent who stays "logged in" at night makes every call wait until the queue gives up on them.

A tip for managers: don't measure an agent by the number of calls alone. With "longest idle" distribution the numbers are similar anyway; the real difference is how many calls came back from the same caller within a day — a sign that the problem wasn't solved the first time.

The numbers worth knowing — in plain words

In the call center world there are dozens of metrics. For a business, a yeshiva, and a nonprofit, five are enough, and all of them are in the call log:

• Abandonment rate — how many callers hung up before an agent answered. Up to 5% is good; above 10% means you are short of agents, or the wait is boring.
• Average wait time — how long those who were answered waited. Under half a minute in normal times is a reasonable goal for a small business.
• The longest wait of the day — the number that interests the manager most. An average of 25 seconds can hide a customer who waited eight minutes.
• Service level — the percentage of calls answered within a target time, for example 80% within 30 seconds. This is the metric professional call centers live by.
• Calls per agent per hour — not to rank agents, but to know whether the distribution is really equal.

A sixth number that isn't in any report: how many times a callback was offered and how many accepted it. If no one presses 1, the offer comes too late or isn't understood.

When the queue misbehaves — symptom and cause

A queue in a small business: even without a call center

A queue isn't only for call centers. A law office with two secretaries benefits from it just as much as a twenty-seat call center: at nine in the morning, when ten clients call within a quarter of an hour, without a queue eight of them hear a busy signal or voicemail. With a queue they wait a minute, hear "You are second in line," and get answered.

A small office is set up differently from a call center in three ways: a short maximum wait (3–4 minutes, since there is no one else to wait for), a callback offer after just one minute, and a backup destination that is a voicemail box sending a message to both secretaries' email. The distribution method is "ring everyone together," because both sit in the same room and calculated fairness is pointless.

And the fact that justifies it: in the call log of such an office you usually see that most of the missed calls were concentrated in one hour of the day. A queue solves that hour without hiring a third employee.