Service Level Notes: Skills-Based Routing: Top Problems and Solutions

Skills-based routing… Intelligent. Flexible. Real-time. The perfect answer to that proverbial call center challenge of getting the right call to the right place at the right time. At least that’s the way it’s supposed to work. Available in automatic call distributors (ACDs) for a decade now, this powerful routing capability is designed to match each caller with the agent who has the skill set best-suited to handle the call on a real-time basis. It’s been a boon to the efficiency and quality of services provided by call centers that, by nature, have overlapping groups or complex routing contingencies— and who have managed it well. Variations of skills-based routing are also working their way into multimedia queuing environments; e.g. groups that handle a variety of contact channels.

But in many cases, skills-based routing has also created difficult new problems that have tempered or obviated the potential benefits—difficult forecasting challenges, complex staffing puzzles and volatile service levels. As with most technologies, the benefits of skills-based routing are commensurate with the clarity of the purpose for which it is to serve, and the soundness of the underlying processes by which it is managed.

Enough call centers have used skills-based routing, and for long enough, that common problems and the lessons they teach have emerged. The top problems that hamper good results—and the corresponding antidotes—are summarized in this article. While this order doesn’t apply in all cases, it is very typical.

Top Problem #1: Rostered Staff Factor (Shrink Factor) Issues

The most problematic issue relates to that question so many call center managers have asked at one time or another as they’ve wandered the floors: "Where is everybody?" (Rostered staff factor, or shrink factor, numerically represents the percentage of agents not available to handle calls at given times.) Should this be any surprise? Breaks, lunch, meetings, projects, research, training, other transaction questions of mode usage… you know, the story. There are 101 things that can keep agents from the phones. Because call centers are so timing-sensitive, ensuring that people are in the right places working on the right priorities at the right times is a central call center management objective. If planning, communication or priority miscues create problems in normal agent groups, they tend to create havoc in skills-based routing environments. With necessary skills unavailable, calls end up with secondary and tertiary alternatives, sending a ripple effect through the process that can misappropriate available staff and send service level and quality plunging.

If there are times in the day that service level is volatile for some types of contacts, this issue is a likely culprit. The lesson? There’s no substitute for realistically planning and budgeting for the things that keep agents from the phones.

Solution #1:

Spend the time necessary to routinely and realistically anticipate and plan for the activities that keep agents from the phones, by time of day and by skill set. Also, educate agents on the importance and service level implications of timing and schedule adherence.

Top Problem #2: Not Forecasting Accurately at the Skill Level

The queuing formulas and simulation models available for calculating required staff are only as good as the accuracy of input they are analyzing. To anticipate staffing needs, you first need to know how many Spanish-speaking callers you’re going to get between 10:00 and 10:30, how the call mix will change throughout the day for the expert group handling call types A, B and C, and when your Mandarin Chinese-speaking agents will go on break. The inability to forecast accurately at the skill level is the Achilles heel of the powerful simulation programs available. Good forecasting is often the single most time-consuming aspect of managing skills-based routing well. But it is worth the investment. The forecast is the foundation on which all resource requirements and budgets are built.

Solution #2:

Invest the time necessary to forecast call load for each mix of calls requiring unique skill sets (e.g., Spanish-speaking calls for service A; Spanish-speaking calls for service B, etc.). Assess the accuracy of forecasted call load vs. actual; if it is routinely off by more than 5 percent to 10 percent by half-hour, consider combining skills to form more manageable groups.

Top Problem #3: Not Calculating Base-Staff Requirements Accurately

Purveyors of staffing and workforce management packages are continuously engaged in a titanic battle to capture your affections and budget. One of the defining issues is the methodology each uses to calculate staffing requirements in environments with complex routing routines. The classic queuing formula Erlang C and variations of it are still predominant in workforce management systems. However, since it is designed for straightforward environments where any agent in a group can handle any call, two other alternatives have emerged in recent years. One is call center computer simulation, which, like a flight simulator, allows you to test a wide range of variables and assumptions before actually implementing changes. The other alternative consists of various proprietary processes that utilize variations of existing queuing formulas and iterative routines to estimate requirements. With either alternative, a certain amount of trial-and-error and a healthy dose of intuition and experience are necessary to accurately model the environment. You will need to run through plenty (e.g., sometimes dozens) of "what-if" scenarios to get it right.

Solution #3:

Find out what capabilities you currently have for calculating staff. Invest the time necessary to run a wide range of "what-if" scenarios to assess your current capacity capabilities and requirements.

Top Problem #4: Poor Assumptions/Rationale

In general, skills-based routing works best in environments that require many skills and have many possible combinations of skill sets. Help desks handling a wide variety of complex issues and call centers handling many languages are common examples. Skills-based routing can also help to quickly integrate new agents by initially routing only simple calls or calls of a predefined nature to them. What skills-based routing can’t do is compensate for poor planning, inadequate training or poorly designed information systems. Remember, the core assumption of a call center is pooled groups, in which cross-trained agents are equipped to share the work load. All things equal, pooled environments are more efficient than those with specialized groups.

Solution #4:

Keep your eye on the prize, and create an environment that is as pooled as possible. This requires an incessant effort to hire the right people, improve training, improve information systems and reduce staff turnover In other words, go as far as you can toward obviating the need for skills-based routing.

Top Problem #5: No Skills-Based Routing Manager/Coordinator

If all of this sounds time consuming—it is. Even relatively small call centers have learned through tough, practical experience that it often takes the equivalent of a full-time person to keep skills-based routing running smoothly. Projecting requirements, assessing current capabilities, updating system programming and adjusting staffing plans and schedules to accommodate evolving circumstances are ongoing activities.

Solution #5:

Create a position for managing skills-based routing. Equip this person or team with the tools, information and authority necessary to predict requirements, make necessary changes to system programming and staffing plans and advise on future requirements.

There’s No Substitute for Good Planning

If there’s one lesson that towers above all others, it's this—skills-based routing is no substitute for good planning. To the contrary, good results are dependent on accurate forecasts, solid staffing calculations, realistic assumptions about staff availability and logical system programming. Skills-based routing is a powerful capability, but it must be managed appropriately to fulfill its promise.