Online Erlang Calculator Manual

Introduction

This manual gives guidance about using the free online version of CCmath's Erlang calculators to make capacity planning calculations and familiarize yourself with contact center workforce management. You will find the definition of some parameters required and understand the capabilities of Erlang calculators. Do you prefer making calculations in Excel? Please explore our new Erlang Add-in for Office 365. Start your free trial and find out the convenience of this one-of-a-kind tool!

Parameters

The required parameters for online Erlang calculators are briefly explained below. If you want to explore more workforce management terminology, visit our WFM definitions page. Please ensure that you use the correct time units while making your calculations. For any questions, see our FAQ page.

Parameters Definition and Requirements
Abandonments The percentage of arbitrary customer that will abandon the queue. (0 < Abandonments < 100)
Acceptable Waiting Time (AWT) The maximum allowed waiting time of customers to be considered to have a good service. The service level is defined as the percentage of customers that are served within this duration. (AWT = 20 seconds by default)
Average Handling Time (AHT) The average time an agent spends on a call. (AHT > 0)
Average Patience The average time a customer is willing to wait in the queue. A simple estimator for the patience is calculated by dividing the total waiting time (including the waiting times of the abandoned customers) by the number of abandonments. It is important to filter out extreme values in advance. (Patience ≥ 0)
Average Speed of Answer (ASA) The average waiting time that an arbitrary customer with infinite patience in the queue. (ASA > 0)
Forecast The average number of arrivals per unit of time. (Forecast ≥ 0)
Lines The total number of trunk lines, consisting of: lines in use by customers in service, and lines in use by customers that wait. (Lines ≥ Agents, integer)
Number of Agents The number of agents available for taking calls. (Number of agents > 0)
Occupancy The percentage of time that agents are handling calls (0 < Occupancy ≤ 100)
Outbound Calls The average number of outbound calls per unit of time. (Outbound calls > 0)
Redials The percentage of customers that once abandon the queue but reconnect than. (0 ≤ Redials ≤ 100)
Service Level (SL) The percentage of customers that waits less than acceptable waiting time. (0 < SL < 100)
Service Level Definition A flag to switch between different modes of calculating the service level in Erlang X: On offered calls, answered calls, or virtual test customer
Threshold The number of agents that are kept idle when there are no inbound calls before taking outbound calls into service. (0 < Threshold ≤ Agents)

Calculators

Erlang C

The Erlang C model is a queuing system where:

  • Customers arrive according to a Poisson process.
  • Customers are served by a fixed number of single-skilled servers.
  • All calls that find all servers busy wait in the queue until they get served.
  • The calls are answered in order of arrival, thus longest-waiting call is served first.

The Erlang C model is often used for single-skill call centers with only inbound calls. This model uses the following input parameters from the historical data based on three different scenarios:

Scenario Example
Inputs Outputs

1) Calculation of ASA, service level, and agent occupancy based on given number of agents

See the example.
  • Forecast: 4 calls/minute
  • AHT: 3 minutes
  • AWT: 20 seconds
  • Number of agents: 14
  • ASA: 43.35 seconds
  • SL: 61.43%
  • Occupancy: 85.71%

2) Calculation of the minimum number of agents required to answer calls within the given ASA

See the example.
  • Forecast: 4 calls/minute
  • AHT: 3 minutes
  • AWT: 20 seconds
  • ASA: 20 seconds
  • Number of agents: 15
  • SL: 77.13%
  • Occupancy: 80%

3) Calculation of the minimum number of agents required to satisfy the given SL

See the example.
  • Forecast: 4 calls/minute
  • AHT: 3 minutes
  • AWT: 20 seconds
  • SL: 80%
  • Number of agents: 16
  • ASA: 9.21 seconds
  • Occupancy: 75%
Erlang X

The Erlang X model is a queueing system with the following features:

  • Queued customers are impatient and abandon the queue after an exponentially distributed amount of time.
  • Calls are not queued but disconnected due to a finite number of lines. Blocking avoids long waiting time and therefore bad service by limiting the number of customers allowed to be in the system at any time.
  • Abandoned customers may redial with a certain probability.

The Erlang X model gives a better estimate than the Erlang C model by involving customer behavior in the queuing system, considering patience, abandonments, and redials.

In the free online version, you may use the following input parameters from the historical data based on four different scenarios:

Scenario Example
Inputs Outputs

1) Calculation of ASA, SL, customer abandonment, and agent occupancy based on given number of agents

See the example.
  • Forecast: 4 calls/minute
  • AHT: 3 minutes
  • AWT: 20 seconds
  • Average patience: 1 minute
  • Redials: 50%
  • Number of agents: 14
  • ASA: 7.34 seconds
  • SL: 83.23% on offered calls
  • Abandonments: 9%
  • Occupancy: 81.68%

2) Calculation of the minimum number of agents required to answer calls within ASA

See the example.
  • Forecast: 4 calls/minute
  • AHT: 3 minutes
  • AWT: 20 seconds
  • Average patience: 1 minute
  • Redials: 50%
  • ASA: 20 seconds
  • Number of agents: 12
  • SL: 65.07% on offered calls
  • Abandonments: 19.15%
  • Occupancy: 89.41%

3) Calculation of minimum number of agents required to satisfy given SL

See the example.
  • Forecast: 4 calls/minute
  • AHT: 3 minutes
  • AWT: 20 seconds
  • Average patience: 1 minute
  • Redials: 50%
  • SL: 80% on offered calls
  • Number of agents: 14
  • ASA: 7.34 seconds
  • Abandonments: 9%
  • Occupancy: 81.69%

4) Calculation of minimum number of agents required to satisfy given abandonment rate

See the example.
  • Forecast: 4 calls/minute
  • AHT: 3 minutes
  • AWT: 20 seconds
  • Average patience: 1 minute
  • Redials: 50%
  • Abandonments: 5%
  • Number of agents: 16
  • ASA: 2.84 seconds
  • SL: 93.24% on offered calls
  • Occupancy: 73.61%
Erlang Blending

The Erlang Blending model considers that agents can work on different types of calls: inbound and outbound.

  • It is assumed that there is an infinite amount of outbound calls to be done. When an agent finishes a call, they will always take the longest waiting inbound call from the queue, if there is any.
  • In case the queue with inbound calls is empty, a decision has to be made whether the agent takes an outbound call or remains idle so that they are immediately available for the next arriving inbound call.
  • This decision is modeled using a threshold. An available agent is allowed to take an outbound call only when the threshold number of agents is left idle.

This model uses the following input parameters from the historical data based on two different scenarios:

Scenario Example
Inputs Outputs

1) Calculation of SL based on a given threshold

See the example.
  • Forecast: 4 calls/minute
  • AHT for inbound calls: 3 minutes
  • AHT for outbound calls / emails: 6 minutes
  • AWT: 20 seconds
  • Number of agents: 16
  • Threshold: 4 agents
  • Occupancy: 89.30%
  • Outbound calls: 0.38/minute
  • ASA: 16.29 seconds
  • SL: 76.79% on offered calls

2) Calculation of the threshold value for agents based on SL

See the example.
  • Forecast: 4 calls/minute
  • AHT for inbound calls: 3 minutes
  • AHT for outbound calls / emails: 6 minutes
  • AWT: 20 seconds
  • Number of agents: 16
  • SL: 80% on offered calls
  • Threshold: 5 agents
  • Occupancy: 86.01%
  • Outbound calls: 0.29/minute
  • ASA: 13.68 seconds
Erlang Chat

The Erlang Chat model involves the following characteristics:

  • Agents can handle multiple chats simultaneously.
  • Parallel handling of chats can increase efficiency.
  • Average handling time might increase when working on multiple chats.

The Erlang Chat model is suitable for contact centers with agents handling chat interactions. It is essential to consider the impact of multiple parallel chats on average handling time and overall efficiency.

This model uses the following input parameters from the historical data based on three different scenarios:

Scenario Example
Inputs Outputs

1) Calculation of ASA, SL, customer abandonment, and agent occupancy based on given number of agents

See the example.
  • Forecast: 4 calls/minute
  • AHT for handling 1 chat: 3 minutes
  • AHT for handling 2 chats in parallel: 3.5 minutes
  • AHT for handling 3 chats in parallel: 4 minutes
  • AWT: 20 seconds
  • Average patience: 1 minute
  • Number of agents: 5
  • ASA: 13.45 seconds
  • SL: 74.46%
  • Abandonments: 15.44%
  • Occupancy: 86.78%

2) Calculation of the minimum number of agents required to handle chats within ASA

See the example.
  • Forecast: 4 calls/minute
  • AHT for handling 1 chat: 3 minutes
  • AHT for handling 2 chats in parallel: 3.5 minutes
  • AHT for handling 3 chats in parallel: 4 minutes
  • AWT: 20 seconds
  • Average patience: 1 minute
  • ASA: 15 seconds
  • Number of agents: 4.89
  • SL: 72.12%
  • Abandonments: 16.92%
  • Occupancy: 87.46%

3) Calculation of the minimum number of agents required to satisfy given SL

See the example.
  • Forecast: 4 calls/minute
  • AHT for handling 1 chat: 3 minutes
  • AHT for handling 2 chats in parallel: 3.5 minutes
  • AHT for handling 3 chats in parallel: 4 minutes
  • AWT: 20 seconds
  • Average patience: 1 minute
  • SL: 80%
  • Number of agents: 5.36
  • ASA: 10.47 seconds
  • Abandonments: 12.15%
  • Occupancy: 83.55%

4) Calculation of minimum number of agents required to satisfy given abandonment rate

See the example.
  • Forecast: 4 calls/minute
  • AHT for handling 1 chat: 3 minutes
  • AHT for handling 2 chats in parallel: 3.5 minutes
  • AHT for handling 3 chats in parallel: 4 minutes
  • AWT: 20 seconds
  • Average patience: 1 minute
  • Abandonments: 10%
  • Number of agents: 5.6
  • ASA: 8.48 seconds
  • SL: 83.62%
  • Occupancy: 81.40%
Multiple Skills

The Multiple Skills simulator can be considered as an extension of Erlang X with multiple skills-requiring calls but runs with a simulation.

  • The calls are under two classes based on their skill requirements.
  • Agents handle the demand from these classes based on their skill set.

Let us compare two scenarios to see that the same service level and occupancy can be achieved with fewer agents when they are multi-skilled.

Scenarios Parameters Result of calculation

1) Calculation of ASA, SL, customer abandonment, and agent occupancy, when there are two types of agents: all of them are single-skilled

See the example.
Type 1 Calls:
  • Forecast: 4 calls/minute
  • AHT: 5 minutes
  • AWT: 20 seconds (0.33 minutes)
  • Average patience: 10 minutes
  • Number of agents: 24
Type 1 Calls:
  • ASA: 0.01 minutes
  • SL: 94.75%
  • Abandonments: 5.25%
  • Occupancy: 78.67%
Type 2 Calls:
  • Forecast: 4 calls/minute
  • AHT: 5 minutes
  • AWT: 20 seconds (0.33 minutes)
  • Average patience: 10 minutes
  • Number of agents: 24
Type 2 Calls:
  • ASA: 0.01 minutes
  • SL: 94.67%
  • Abandonments: 5.33%
  • Occupancy: 78.89%

2) Calculation of ASA, SL, customer abandonment, and agent occupancy, when there are two types of agents: some of them are multi-skilled

See the example.
  • Forecast: 4 calls/minute
  • AHT: 5 minutes
  • AWT: 20 seconds (0.33 minutes)
  • Average patience: 10 minutes
  • Number of type 1 agents: 20
  • Number of type 2 agents: 20
  • Number of multi-skilled agents: 6
  • ASA: 0.17 minutes
  • SL: 82%
  • Abandonments: 1.73%
  • Occupancy: 84.7% for single-skilled agents
  • Occupancy: 90% for multi-skilled agents