Tackling Transferred Calls Through Your Agent Scorecard

A primary goal of contact center managers is to balance the cost of service delivery against service quality. For example, a classic decision is picking the contact center staffing level. The lower the staffing level, the lower the cost but the higher the waiting times, which leads to customer frustration and lower perceived quality.

While there are a myriad of practices and technologies to deal with the staffing decision, the decision of whether to transfer a call from one agent to another presents another cost/quality tradeoff that may not be receiving the attention it should. On the cost side, transferring from one agent to an expert agent may be expensive, given that experts often receive premium compensation. However, transferring could help lower cost if the expert can solve the customer problem fast enough to offset this premium. On the quality side, customers generally dislike transfers as they often have to repeat details of their problem to a new agent. Also, transferring can free up agents to serve other customers, which can bring down their waiting times and increase their service quality.

Agents often have discretion over whether and when to transfer, particularly when troubleshooting customer problems. After a series of failed resolution attempts, the agent may decide it’s time to move the customer on to the next level of support. This was the setting of a recent study that my colleagues and I at the Johns Hopkins University Carey Business School conducted with the contact center of an anonymous, mid-sized bank. The work was recently published in Operations Research and is titled “The Gatekeeper’s Dilemma: ‘When Should I Transfer This Customer?’”

In late 2017, the call center management team of this bank began examining agent-level transfer rates, and found what they felt was high variation in transfer rates, particularly for online banking access requests, with transfer rates ranging between 4% and 37%. Concerned about the cost and quality implications of excessive transferring, the team felt that they could drive this variation lower by incorporating agent transfer decisions into their monthly scorecard.

In Spring 2018, the management team implemented the new scorecard by changing the agents’ “productivity score” calculation. While previous versions of the productivity score included measures such as AHT and After Call Work, this new version was designed to encourage agents to quickly transfer issues requiring longer handling times, and discourage them from transferring issues requiring shorter handling times.

Specifically, their productivity score is a monthly average taken over all calls handled by an agent, with each call contributing between 0 and 100 points. The value of a call depends on whether the agent transfers the call and on the duration of the call prior to its resolution or transfer. All calls start with a base score of 80, but the agent may earn up to 20 bonus points if the call is resolved or transferred in less than 85% of the AHT for calls from that skillset. However, regardless of the handling time of the call, if the agent transfers, the score is reduced by 20.

We worked with this call center to capture how this scorecard change affected agent transferring, and found that it was pretty effective. For example, the median transfer rate in the online banking access queue dropped from 13.6% to 8.9% in the six months after the scorecard rollout, and we saw that the reduction maintained for well over a year past the study.

We hope that our findings will spur contact center managers to ask a variety of transfer-related questions, ranging from tactical to strategic. For example:

• How much does transferring cost us? What percent of our time is spent handling customers off between agents? How many agents does a customer speak to on average to finally get a resolution?
• How does transferring affect customer perceptions in our contact center? Can we see negative effects of transferring in our customer surveys? If so, how big of a deal is it?
• Is the design of our phone menu leading to unnecessary transfers? How do we balance the desire of customers to quickly navigate the phone tree against their desire to get to the correct agent to handle the call?
• Does our scorecard overemphasize AHT at the expense of transfers? Do agents have a loophole in that they can simply cut down their AHT by “passing the buck” to some other agent? How can we design the scorecard so that agents have at least some disincentive against excessive transferring?
• Finally, if an agent transfers far more than their peers, why is that? Is it lack of skill or lack of will? If skill, how do we train that agent up to transfer less? If will, how do we motivate the agent to take more ownership of the call?

In our data-rich world, finding answers to questions such as these can help contact managers design policies that minimize unnecessary transfers, helping strike the right balance between cost and quality.