What is the Problem?
My customer frequently asks if this model or that model can be used to direct call center traffic. Usually the models are projecting an acquisition or engagement window with too much of a gap to perform this function, and this is hardly part of their use case as acquisition models. However, I can perceive a model that does perform this function. Yet a model is not entirely what they need.
Models seem to be the magic “fix-all” in the Financial Services and Insurance Industry (FSI). Customers think that models will fix their product, marketing strategy and call center. The reality is: anything that pertains to analytics seems to be a model to them. That is an article for another day. For this one, I want to talk about the call center problem, namely, directing calls more efficiently.
A model might be the answer they are looking for. However, you cannot just build a model and leap into its use blindly, without testing. The best way to test in this situation is using discrete event simulation.
What is Discrete Event Simulation?
Discrete Event Simulation (DES) is the process of codifying the behavior of a complex system as an ordered sequence of well-defined events. In this context, an event comprises a specific change in the system’s state at a specific point in time (arrival at a bank, service by a teller, incoming call, etc.). Rather than stepping based on a time increment, like every second, DES advances based on events—events that may or may not be equally spaced in time.
Common applications of DES include:
- stress testing ( a form of deliberately intense or thorough testing used to determine the stability of a given system or entity)
- evaluating potential financial investments
- studying call center operations
- modeling procedures and processes in various industries, such as manufacturing and healthcare
- studies that support system design
- studying reliability, availability and maintainability (RAM)
- just about anything that involves queuing (arrivals, waiting time, service time)
Can We Simulate the Call Center?
One of the most applicable situations that lends itself to DES is the operations of a call center. When a customer calls, you have an arrival event. If they have to wait for a representative they enter a queue until such time as their call is answered, which begins a service event. If their call was rerouted upon entering the system–the initial arrival event–the rerouting creates a new arrival event for the call center to which it was transferred, and so on. One can study where bottle-necks occur, rerouting efficiency, waiting times, balking (customers deciding not to join the queue if it is too long), reneging (customers leave the queue if they have waited too long for service), etc.
An effective DES process must include, at a minimum, the following characteristics:
- Predetermined starting and ending points, which can be discrete events or instants in time (arrivals and departures, for instance).
- A method of keeping track of the time that has elapsed since the process began (waiting time, for instance).
- A list of discrete events that have occurred since the process began (begin service, for instance).
- A list of discrete events pending or expected (if such events are known) until the process is expected to end.
- A graphical, statistical, or tabular record of the function for which DES is currently engaged (plot of waiting times and service times, for instance).
Using DES, we can test the model with multiple parameter, or test multiple models, without disrupting the operations of a call center. Once we determine which model (or which parameters) work best, we can then use that particular model to direct calls, for instance. Continued monitoring and used of DES for simultaneous parallel testing can be used to maintain and improve efficiency.
DES can help administrators predict how a network will behave under extraordinary conditions, such as the Internet or claims call center during a major disaster, like a wildfire or earthquake. DES is also commonly used to monitor and predict the behavior of investments; the stock market is a classic example.
About the Author
Jeffrey Strickland, Ph.D. CMSP, is the Author of Discrete Event Simulation using ExtendSim and a Senior Analytics Scientist with Clarity Solution Group. He has performed predictive modeling, simulation and analysis for the Department of Defense, NASA, the Missile Defense Agency, and the Financial Services and Insurance Industry. Jeffrey has also taught numerous mathematics and statistic courses, as well as modeling and simulation, for several institutions of higher learning. He has presented award-winning tutorials in pre-conference workshops, and has often been a keynote speaker.
Jeffrey is a Certified Modeling and Simulation Professional (CMSP) and the author of over 200 articles and 20 books including:
- Predictive Analytics Using R
- Crime Analysis and Mapping
- Missile Flight Simulation
- Mathematical modeling of Warfare and Combat Phenomenon
- Predictive Modeling and Analytics
- Using Math to Defeat the Enemy
- Verification and Validation for Modeling and Simulation
- Simulation Conceptual Modeling
- System Engineering Process and Practices
- Weird Scientist: the Creators of Quantum Physics
- Albert Einstein: No one expected me to lay a golden eggs
- The Men of Manhattan: the Creators of the Nuclear Era
- Fundamentals of Combat Modeling