Queuing Theory and Erlang Distribution Explained from a Call Center Perspective

Queuing Theory and Erlang Distribution Explained from a Call Center Perspective

Queuing is a widely used concept that is an integral part of inbound contact centers. An ACD or Automatic Call Distributor is used by the call centers for distribution of incoming calls to particular agents or resources, ACD can hold the incoming calls in FIFO (Fast In First Out) sequence until an agent is free to attend a call. From the perspective of a caller, if there was no virtual queuing available, they would have only two options available to them; wait till the time an agent is available or hang up/abandon the call and try calling back later. As far as a call center is concerned, a long queue would mean several abandoned calls, customer dissatisfaction and repeat attempts on the part of the customers.

Customers can enjoy the benefit of callbacks due to the sophisticated virtual queuing systems rather than waiting for his turn in a queue. This solution is also applied widely in many famous amusement parks that frequently have long queues for riding the popular rides. There is a computerized solution that enables the visitors to book their positions in a virtual queue instead of waiting in a queue physically.

For a small-sized business, the solutions for virtual queue management are available in two forms:

  • Through notifications sent via SMS texts
  • Apps on tablet and smartphone devices with remote viewing of the queue status and in-app notification.

There are different types of queuing theories and virtual queuing systems that are used in a call center. The standard one is a FIFO that keeps the place of a customer in line to monitor the condition of the queue till the time an EWT or Estimated Wait Time becomes more than a predefined threshold value. If the value of the threshold is exceeded, the system can intercept the incoming calls prior to their entry in to the queue. The system then communicates about the EWT to the concerned customers and also provides a choice of getting a call back within the same time if they were kept waiting on hold.

In case a customer likes to opt for waiting in a queue, his call is directly routed to that queue. On the other hand if a customer chooses a callback, he is asked to key in his contact number and hang up the call. As the ACD queue get worked off, there is a virtual placeholder that can maintain the position of the customers in the queue.  An outbound call is triggered by the system the moment a virtual placeholder is about to reach the queue’s top.

When the customer answers the callback, a confirmation is done by the system to assess whether the correct person is being called or not and prepared to interact with a representative. After getting the confirmation, the call is routed by the system to the next free agent who takes it like a regular inbound call.

Impact of Queuing System

A virtual queuing system has a big impact on the performance metrics of a call center in several ways.  ASA or Average Speed-to-Answer is the metric used for measuring Queue time.  Whenever the callers are given an option for receiving a FIFO callback, the acceptance rates of the callers are between 45 percent and 55 percent.

Thus about 50 percent of the calls that could have been in a queue for five to ten minutes will now have an ASA of about 10 seconds. Moreover, since calls cannot be abandoned while in a virtual queue, the total number of abandoned calls also comes down considerably.  Though there is positive customer satisfaction due to this, it is difficult to measure them in an objective manner.

Erlang Distribution with respect to Call Centers

This model was introduced by A.K. Erlang for examining the total number of calls that can be made simultaneously to the operators located in the switching stations. Though the theory started to be applied on telephone calls for traffic engineering, its scope has now been broadened to regard the waiting times in general for any type of queuing system.

Today, Erlang calculators are being used globally for implementing different types of statistical calculations that are related to call centers and telecommunications systems.  There are different types of Erlang calculators for being applied for different purposes.

For instance, an Erlang C calculator is used for monitoring the performance of queuing systems instead of a caller hanging up after he gets a busy tone.  The application of queuing includes helpdesks, call center representatives and switchboard operators.

The assumptions made by the Erlang C model are mentioned below:-

  • The size of the queue is unlimited.
  • Customers will be routed to the first available representative.
  • Users are attended in the sequence of arrival
  • Time of service is exponential.
  • Callers wait when they find that the system is busy but the impact of abandoned calls is not taken into account.
  • There is random offering of calls in the queue.
  • The call volume for the period of calculation does not exhibit drastic changes in call volumes.