Are Your Agents Over-Skilled and Under-Utilized?

By Bill Durr, Contact Center Consulting

Skill routing is an important — often integral — practice in today’s contact centers. This ACD feature enables organizations to establish routing rules connecting customers having unusual or highly specific needs with agents who have the necessary skills and experience to serve them.

By matching caller needs with the appropriate agent skills, contact centers can improve first-call resolution, a primary driver of customer satisfaction. And because an ACD is typically the very first technology purchase by a contact center, the management team frequently attempts to leverage skill routing to the utmost.

In many centers, skill routing is well-defined and well-executed — meaning that a reasonable number of skills are assigned. But some contact centers take the concept to unworkable extremes, and the result can be far too many skill definitions, with far too many low-volume queues.

How many is too many? That’s not an easy question, because some centers must have an extensive number of skill definitions.

Technical support help desks are a good example, since any one person, no matter how gifted and well trained, is unlikely to have expertise that spans the complete range of the products being supported. Contact centers that support a significant number of languages and a broad range of products or services are another example. The combinations of language and product/service skills can produce scores of individual skill definitions.

There are, however, centers that have slowly created an environment with many skill definitions, often as a result of well-intentioned career pathing. Typically, in addition to defining language and product/service skills, they’ve established tiers within them to denote increasing proficiency levels with — presumably — related salary differentials.

Does this sound like your center? If so, do yourself a favor and think long and hard about the path you are on and the increasing problems you are likely to face.

Given scores — even hundreds — of skill definitions, it is likely that you have definitions describing only one, two, or three agents. The performance dynamics of small groups of agents typically are sub-optimal.

Why? Let’s use an example to illustrate the point. Imagine a North American agent who speaks French and Spanish and possesses unique product or service knowledge. Consequently, these call types will be handled only by this particular agent—which in turn means that we don’t want the agent to handle just any kind of call, for fear that he or she won’t be available when that highly specialized caller contacts us. As a result, this single, unusually skilled agent won’t have the same occupancy or productivity as less-specialized agents. This is most assuredly not a good practice to foster and propagate across multiple queues.

Service level goals can also yield a similar outcome. Forecasting accurately and staffing appropriately to meet strict, interval service level goals can be notoriously difficult with low-volume queues — and one or two bad intervals in low-volume queues can sink the day’s service level for the entire contact center.

Continuing with our example, it’s reasonable to assume that in North America, the number of Spanish-speaking callers will far exceed the number French-speaking callers. The French call queue is going to be a low-volume queue, and — because it’s so easy to mess up the interval service level in low-volume queues — the operations team will almost certainly exempt this uniquely skilled agent from handling ordinary calls, enabling him or her to be available when a French caller arrives. The need to meet service level goals drives their rationale.

Moreover, creating accurate forecasts in low-volume queues can be very difficult, and so can reconciling workforce management (WFM) service level projections against reality. The value proposition for WFM software is simple: Adhere to the schedules and the center will realize the service level projections provided by the software. But in low-volume queues, the forecast stands a greater chance of being off, and the schedules that are in place may have to contend with higher or lower demand than predicted in each interval level. The result can be a contact center with large swings in realized service levels from one interval to the next.

I am not advocating against skill routing — quite the contrary, since many centers would find it impossible to operate without it. But if the number of skill definitions creeps up too far, it can reach a tipping point which results in underutilized agents.

How can your center escape this “over-skilled and under-utilized” scenario? Start by cross-training your agents. Following this best practice can enable you to prune skill definitions, helping keep the number of low-volume queues in check.

Bill Durr is the principal consultant at Contact Center Consulting. He has been involved in the contact center industry in a variety of roles for over 30 years, leading seminars, educational sessions, providing consulting services and authoring four books and numerous articles on technology and management processes.He maintains a web site with free downloadable resources at