Waiting is a design problem
When more people call than there are people to answer, something has to give. Either the extra callers get a busy signal and try again, or they are held in a line until someone is free. Call queueing is the second option. It seems obvious now, but it needed both mathematics and equipment to become practical.
Queues are a general problem. Aircraft waiting to land under air traffic control, and packets waiting in a router under packet switching, are managed with similar thinking.
The mathematics came first
The study of queues began with the telephone. The Danish mathematician and engineer Agner Krarup Erlang, who worked for the Copenhagen Telephone Company, published a paper in 1909 applying probability to telephone traffic. He showed how to estimate how many lines or operators were needed so that the chance of a caller finding everything busy stayed below a chosen level. In later work he treated cases in which callers wait instead of being refused. A unit of telephone traffic is called the erlang in his honor.
This body of work, queueing theory, is now used for many things, from supermarket checkouts to computer networks. Telephone planners used it for decades to decide how many circuits, switches and staff were needed. The conclusions are statistical rather than guarantees: with a given number of agents, long waits become unlikely, but they cannot be ruled out.
Busy signals and the lost call
In the earliest telephone networks, a caller who could not be connected heard a busy tone or was told by an operator that the line or all trunks were busy. The caller then had to try again. Operators sometimes offered to call back when a line became free, a manual form of queueing that depended on memory and notes.
For businesses, a busy signal meant a lost customer. A shop or a clinic with few lines would lose calls at busy times. Using more lines was one solution but cost money, and extra lines were idle in quiet periods.
Automatic call distribution
The equipment for queueing was the automatic call distributor. Placed in front of a group of agents, it answers incoming calls, places them in a line if all agents are busy, and connects each caller, in turn, to the next agent who becomes free. Systems of this kind appeared on private exchanges from the 1960s, and dedicated products became widespread in the 1970s and 1980s.
Once a system was able to hold calls, rules became possible. Calls might be served in the order they arrived, or some callers could be given priority, such as customers with urgent needs. Calls could be sent to agents with particular skills, such as a language or a product specialty. A maximum wait time could trigger an overflow to another group or a message offering a callback.
Recorded announcements and music
Callers who wait need to be told what is happening. The first approach was a recorded announcement played when the call was answered by the system, saying that all agents were busy and asking the caller to hold. Machines played these using loops of magnetic tape or drums and later digital memory.
Music on hold followed. Its invention is commonly attributed in popular accounts to an American, Alfred Levy, in the early 1960s, who is said to have developed a way to play music over a held line, though the details are told in several versions. Whatever the exact origin, companies began playing music and later messages about the business to reassure callers that the line was still connected. Research on waiting suggests that unoccupied time seems longer than occupied time, so providing something to listen to was an attempt to make waits more tolerable. Opinion on the effect varies.
Statements such as “your call is important to us” and information about expected wait times are products of the same effort, aimed at reducing the number of callers who hang up.
Interactive menus and callbacks
From the 1980s, interactive voice response systems added a way to handle callers while they waited, asking them to enter an account number or to select a topic. The answers were used to route the call or to display the customer’s record to the agent. Menus became a feature of customer service lines, and are often criticized when they become long or confusing.
A later development, the virtual queue, lets a caller keep their place in line without staying on the line. The system offers to call back when an agent is free, and the caller hangs up and goes on with the day. Callback features became common in the 2000s as systems became better at predicting wait times, a task for which some vendors now use artificial intelligence.
Beyond the telephone
The structure of the call queue spread to other channels. Web chat, email and messaging systems use similar logic to hold and assign requests to staff. Even outside customer service, the same ideas appear in scheduling and computer systems.
A modest but influential invention
Queueing is rarely noticed unless something goes wrong, yet it shapes the experience of every call to a busy organization. Its history connects the mathematics of a Danish telephone engineer to the music playing in a customer’s ear, and it shows how a technique designed to manage limited telephone circuits became a general tool for managing waiting.