Situation

Many back-office processing and sales support units in banking, health care, insurance, telecommunications, and other service environments struggle to maintain high operational efficiency in the face of extremely variable customer demand. For example, a global asset manager’s transfer agency (an internal group that processed 50 types of account-related transactions) had trained all its employees to handle every type of transaction, thinking that the resulting flexibility would more than offset higher training costs. As the company expanded its operations globally, however, executives were surprised to find the productivity—and service levels—of the unit’s teams eroding. At the busiest periods, up to 80 percent of them failed to meet their service-level agreements.

When the executives looked more closely, they found that the variety of tasks the employees undertook actually made it difficult for them to meet the promised service levels and for management to measure or manage their performance accurately. Frontline workers trained to do everything, for example, didn’t encounter some tasks often enough to do them really well. Certain workers therefore cherry-picked the easier assignments—a pattern that damaged morale and delayed the harder transactions. Worse, customer inquiries about service delays added to the volume of incoming calls, slowing turnaround times for all transactions and resulting in expensive overtime.

To determine the unit’s staffing requirements, the company measured the volume and frequency of the 50 transactions, ultimately grouping 30 of them into five sets by level of difficulty. The sets were distributed to “baseload” teams that handled the same assignments each day, thus simplifying staffing and performance management. The transactions in each set were similar enough for the company to train and manage these employees effectively, yet varied enough for them to avoid boredom. As their skills improved, they could graduate to baseload teams that handled increasingly difficult transactions. In addition, a small swing team was created to handle daily volumes exceeding the forecast baseload levels for the 30 transactions, as well as the remaining 20 infrequent (and more challenging) ones. This arrangement helped the unit meet its service-level deadlines in nearly all cases while reducing its frontline staff and management by 25 percent and overtime by 90 percent.

Implications

Power companies use “peaker” plants to manage spikes in electricity demand flexibly and cost-effectively. Likewise, managers in many back-office processing environments can make them more flexible and remove waste to boot by organizing transactions or activities according to their variability and then assigning different ones to baseload or swing teams (exhibit). The first step is to develop a detailed understanding of customer demand, since its patterns may help managers group operational responsibilities more efficiently. (Demand for tasks A, C, and D, taken together, say, may be less variable than demand solely for task B.) The correct principle for organizing tasks can vary by context; for example, it can make sense to group them by customer segment, value at stake, degree of difficulty, or regulatory requirements. Finally, the identification of baseload tasks and teams helps companies to create career paths for frontline workers, while the greater degree of specialization shortens learning curves for new employees—a benefit for companies that hope to consolidate sites or move activities offshore.