|
|
 |
 |
 |
 |
 |
|
REMS projects
Simulating Interruption Modeling and Analysis of Email Management for Improved Customer Relationship Management Stochastic Programming for Optimal Email Processing
Simulating Interruptions and Information Overload
due to emails Our current research is looking into email’s effect as an interruption to other work. We propose several strategies for minimizing email’s negative impact as an interruption. Strategies include continuously processing email messages, holding specific “email hours” while ignoring email outside of this time frame, and processing email only in batches. We simulate a knowledge worker’s environment to look at the effects of these strategies on the knowledge worker’s performance. Performance measures include the email response time and the realized utilization of the knowledge worker. A tradeoff exists between the prompt email response times that can result from continuously monitoring email and the improved efficiency of holding “email hours” or processing email in batches. Our aim is to analyze these tradeoffs and prescribe appropriate email processing strategies for various knowledge work environments. Queuing Models for Analyzing Customer Contact Center Operations
Considerable
research has been done in modeling call centers, where customer service agents
answer telephone calls from customers. Customer
call centers, which represent a multi-billion dollar industry, are evolving into
customer contact centers in which customer contact happens through other media
– e-mail, fax, and the Web. Contact
centers differ from the traditional call centers in one important way – they
allow the possibility of postponing some work (e.g ,responding to e-mail).
This has necessitated the development of new stochastic models to analyze
the performance of contact centers. We
model the operations of contact centers, where the customer contact is mainly
via e-mail. We present ongoing
research on the development of queueing network models to model the processing
of e-mails by a network of customer service agents.
Problem resolution time becomes an important performance measure in
analyzing such systems. Resolution
time is not normally addressed in call center models. Also, unlike call centers,
the agent can be interrupted by other knowledge work while processing e-mails.
The flexibility in postponing work also increases the likelihood of an
existing case being processed by an agent who previously handled it.
We show how some of these features can be handled within a network model
based on two-moment queueing approximations. Modeling and Analysis of Email Management for
Improved Customer Relationship Management We
have recently submitted “Modeling
and Analysis of Email Management for Improved Customer Relationship Management”
to the International Journal of Simulation and Process Modeling (IJSPM),
Special Issue on Modeling and Simulating Business Processes for E-Business .
This paper describes the simulation and analysis of a customer contact
center that employs email as the means of communication between customers and
customer service agents. Customer
call centers, a multi-billion dollar industry, are evolving into customer contact
centers, which now include
channels of communication other than the telephone. Email, a popular alternative
to the telephone for all types of communication, is different from telephone
communication in several ways. Of primary interest is the customer’s
expectation regarding response times. The added flexibility this provides can
represent potential for optimization of customer care. Applications currently
exist that allow management to implement and monitor email management
strategies. For example, incoming email can be automatically routed to specific
agents according to predefined routing rules. The objective of this paper is to
illustrate the use of discrete-event simulation in modeling and analysis of
email response centers. An email response center is modeled to study the
flexibility of the routing and prioritization of incoming email messages. The
results of the simulations indicate that the routing of incoming email messages
has little impact on performance, as defined by average response time and
average resolution time of different categories of email messages. The use of a
priority scheme by the email agent, however, is shown to result in a significant
improvement in the response center’s performance. A
Stochastic Programming Approach to Managing Email Overload This past January, we presented “A Stochastic Programming Approach to Managing Email Overload” at the INFORMS Computing Society meeting. Although email users often have little control over the arrival rate and content of email messages, they do have control over the order in which they process incoming messages. These decisions are often based on heuristics. For example, some workers process email as it arrives, while others process according to the importance of the email. Often times, the value of responding to an email message decreases with time. This rate of decrease may vary from one type of email to another. Also, the future arrivals and processing needs of email messages are stochastic events. The email user must take all of these things into consideration in deciding the order in which to process messages. Most efforts aimed at reducing email overload involve filtering incoming messages and/or automating responses to these incoming messages. Some efforts involve categorizing and prioritizing the email in need of processing. None of the efforts explicitly take into consideration the impact of possible future email processing demands. Stochastic Programming (SP) is an optimization tool that allows for the consideration of stochastic parameters. SP allows the user to find an optimal solution while taking into consideration possible future scenarios. The purpose of this paper is to explore the potential use of Stochastic Programming as a tool for finding the optimal processing sequence of email. We presented a simple example that considers several potential scenarios of arriving email. The results of this example demonstrate the potential usefulness of SP in planning the processing of email. |
| © Copyright 2003 Oklahoma State University. |