The Art of Call Center Management by Brad Cleveland, Incoming Calls Management Institute
Service Level Newsletter, 1996.

It was the turn of the century, and the dawn of a new age in communication. The telephone had been invented several decades earlier, in 1876, and telephone service was proliferating rapidly. The public was beginning to depend on and even expect reliable service.

As the subscriber base grew, telephone companies were contending with new resource planning problems. Automated central offices were yet to be invented, and human operators were required to establish connections for callers. One big question was, how can you know how many telephone operators are necessary? Too few, and service would be unacceptable to callers. But too many would be inefficient for telephone companies, and would drive costs upward for subscribers. Further complicating the issue: calls arrived randomly, driven by the myriad of motivations callers had for placing the calls.

In the years that followed, many bright people would grapple with these challenges. For example, A.K. Erlang, an engineer who worked for the Copenhagen Telephone Company, would develop the queuing formula Erlang C in 1917; Erlang C calculates on-phone staff required for specified service levels, assuming calls come in randomly and queue if all operators are occupied. Others would discover that calls arrive in relatively predictable patterns by season of year, day of week, and half hour of day, and would begin to develop disciplined forecasting methodologies. Others worked on scheduling approaches appropriate for these extended-hour environments.

Things have come a long way since the 1900's, and callers can direct dial around the world without the assistance of an operator. But if you manage a modern call center, there is probably a ring of familiarity to the challenges these pioneers faced. For example, forecasting calls accurately, staffing appropriately, and getting the right people and other resources in place at the right times are key objectives.

Incoming Calls Management Institute has a working definition of incoming call center management, that we've developed and used over the years. In the past five or so years, it has been published a handful of times in these pages: "Incoming call center management is the art of having the right number of skilled people (and supporting resources) in place at the right times to handle an accurately forecasted workload, at service level and with quality."

To accomplish that purpose, effective resource planning is necessary, and generally includes the following steps:

1) Choose a Service Level Objective. Service level, which takes the form of "X percent of calls answered in Y seconds," should be understood, taken seriously and adequately funded. And it should be appropriate for the services being provided and the expectations of the callers using those services. Service level is the critical link between resources and results.

2) Collect Data: The ACD and computer systems are primary sources of planning data, telling us how many calls we're getting, what the patterns are, how long they last, why callers are calling, and how the call mix is changing. But we also need information on things like what marketing and other departments are doing, legislative changes, competitor activities and changes in the customer base, to name a few.

3) Forecast Call load. Call load includes three components, average talk time, average after call work time (wrap up) and volume. A good forecast predicts all three components accurately for future time periods, usually down to a half hour.

4) Calculate Base Staff. This step determines the "on-phone" staff required to handle the incoming call load. Erlang C is built into just about all forecasting and staffing programs, but computer simulation is an alternative that is used by a small but growing number of call centers. Computer simulation has more flexibility to assess complex call handling routines.

5) Calculate Trunks (And Related System Resources). Trunking requirements are calculated after staff because staff impacts delay (the queue) and trunks must carry the delay.

6) Calculate Rostered Staff Factor (Shrinkage). This critical step adds realism into staffing, factoring in breaks, absenteeism, training, and non-phone work.

7) Organize Schedules. Schedules are essentially forecasts of who needs to be where, and when. They are detailed plans should lead to getting the right people in the right places at the right times.

The best managed call centers do a good job of resource planning. They have a process that is systematic, collaborative and accurate. But what about the future? What about the proliferation of new services, such as advanced computer telephony applications, integrated web services, and video? Have the rules change? Should the planning process change?

Not fundamentally. Consider integrated web services. Assume that customers and potential customers surfing the Web can click a button and be connected to the call center, to receive immediate live assistance. Now go to the top of the planning process. Start by establishing a service level that is appropriate. Next, collect the necessary planning data and forecast the load for those transactions. Then calculate on-phone staff, trunks and other required system resources, as well as rostered staff factor. Finally, organize schedules for the agents who will handle the Web-generated transactions.

How about video calls? Start by choosing a service level objective for video calls. Then collect data, and forecast the video call load. Next, calculate the base video-equipped agents required. Then, calculate the trunks and other system resources required to carry the video calls. Next, realizing that video equipped agents will take breaks, get sick, go on vacation, and handle other types of work, calculate rostered staff factor. Finally, organize schedules for video equipped agents.

The point is, the basics of resource planning have been around for quite awhile, and will continue to be important, as long as we need people in call centers and as long as some of our transactions queue for live answer. And it appears that will be the case for some time to come. Yes, technology is automating more and more of the calls agents used to handle. But, the absolute number of calls requiring live answer continues to grow in most call centers, as transactions of all types grow (see From The Observation Post, SLN, December 1995 ). And there are already indications that the Web may generate more calls than it displaces, as it opens up new possibilities and provides new services.

All told, an effective planning process will be as important as ever. We'll have more types of transactions and more complex transactions to handle. But the fundamentals of the planning steps will remain intact and forecasting, staffing and scheduling skills will be in high demand. And even though we will be operating in much more sophisticated environments, will be using better tools and will have a much broader base of knowledge about our craft, the fundamental resource planning challenge faced by young telephone companies almost a century ago will continue to be important challenge tomorrow: get the right people and supporting resources in the right place at the right times.