Systems, Boundaries, and Starting Points
in the Call Center
Part Six in the Continuing Series, Getting Quality Right

By Cliff Hurst

November 2008

Let's review: In the first installment of this series we asked, "Why monitor calls?"  This initial question prompted four additional questions to be addressed in any effective quality monitoring program.

1)   How are we, as an organization, doing at representing our call center?

2)   What can we, as an organization, do to get better at representing our call center?

3)   How is this particular agent doing at representing our call center?

4)   What can we, as managers, do to help this agent to get better at representing our call center?

To address the first question, we uncovered four elements:

• Select a random sample of calls for monitoring 

• Select a sufficient sample size to achieve the intended level of accuracy and precision 

• Monitor calls using processes and a QM form that are reliable and valid 

• Achieve results that are approximately normal in their distribution 

Now we can move on to the second question, "What can we, as an organization, do to get better at representing our call center?"  The answer to that question starts with systems theory.  Some systems are closed, and others are open. 

Closed Systems: The most commonly recognized example of a closed system is a thermostat.  The thermostat takes feedback in the form of temperature.  It acts on that feedback to turn on the heat or air conditioning when needed until the temperature reaches a predetermined level, and then the thermostat turns off the heat or air conditioning.  The thermostat is a control mechanism, using a single-loop feedback to keep the variation in temperature within a narrow range.  A closed system operates all by itself and responds to only one variable.  Within call centers, a simple ACD that distributes calls to the next available agent is an example of a closed system. 

Open Systems: Most processes in a call center, however, are not so simple and are better described as open systems.  Open systems operate under the influence of feedback from multiple variables.  They also interact with, change, and are changed by those same variables; they interact with their environment, which is a larger system of which they are a part. 

Take a workforce management system, for instance.  You begin with calculations based on historical call volumes and call lengths.  You then adjust those calculations by forecasts that take into account nonhistorical variables.  Unfortunately, even the best plans can be thrown off by many things.  When this occurs, call volumes can unexpectedly spike, and service levels plummet.  In an open system, the notion of control goes out the window.  Think of all the mutually influencing variables involved, and you will see that a quality monitoring process is best described as an open system. 

With the theory of open systems in mind, we recognize that the best we can accomplish through quality monitoring is the ability to influence the outcome of the system.  To think that we can control the outcome is foolish.  Unfortunately, because the notion of control is ingrained as a management principle, understanding the influential nature of an open system is more difficult that it first appears

Boundaries: Systems have both boundaries and starting conditions.  As typically practiced, the boundaries of quality call monitoring extend from "hello" to "good-bye."  If your sole purpose in monitoring calls is to evaluate agent performance, then this is an acceptable boundary.  However, to answer question number two, "What can we, as an organization, do to get better at representing our call center?" broader boundaries are called for. 

To extend the starting boundary, consider including the speed of the answer and the caller's experience with your IVR.  Likewise, after "good-bye," the caller's experience continues.  How promptly was the order or request processed?  How accurate was it?  To extend the boundaries even farther, consider the reason that prompted the call in the first place.

In reflecting on what we might learn from the self-destruction of the outbound B-to-C industry in the United States, I realized that a broad conception of the boundaries of the outbound telemarketing "system" in this country would encompass the entire population of U.S. households.  This population can be thought of as a "common pool" resource that needed to be marketed with sustainability in mind; it wasn't.

Starting Conditions: An open system is extremely sensitive to initial conditions.  Change the initial conditions of a system even slightly, and the outcomes may vary wildly and unpredictably.  For call centers, one starting condition that may influence the quality of a caller's experience is the length of time that the caller has had to wait in queue before reaching a live agent.  Managing the queue is a day-to-day challenge for many call center managers.  Instead of directly lobbying for an increased head count, perhaps a more effective approach would be to build a rational argument based on a statistically sound analysis of the consequences of keeping callers in queue.

Queue time is just one example of the importance of initial starting conditions on quality scores.  Here are a few others:

IVR: What if you have just implemented a new IVR menu - or installed an IVR?  That's a change in starting conditions.  What impact does that change have on quality scores?

 Shift bidding: Suppose you practice shift bidding and your experienced reps prefer to work days.  This leaves your newest agents working at night.  You can determine if there is a correlation between longevity and quality scores, using those results to determine whether it might be worthwhile to provide additional supervision, coaching, or training to newer reps working at night. 

Cyclical events: Suppose monthly statements or mailings are sent out.  What happens the week after a redesigned form is mailed?

In your continuing quest for getting quality right, begin to view your call center as an open system, expand your understanding its boundaries, and factor in the starting conditions.  The result will be the ability to better represent your call center.

Read part 5 and part 7 in this series.

 Cliff Hurst is president of Career Impact, Inc, which he started in 1988.  Contact Cliff at 207-499-0141, 800-813-8105, or cliff@careerimpact.net.