This study used an applied ethnographic research method to investigate human-computer interaction (HCI) between call center agents and agent-facing software in the context of contact-center culture. Twenty semi-structured interviews were completed, along with non-participant observation at two contact centers, one that followed a user-centered design (UCD) process for software development and another that did not. Agent productivity and satisfaction at the non-UCD center were hampered by poor task-UI integration, ambiguous text labels, and inadequate UI standardization. Agents required multiple applications to complete a single task, leading to long task times and cognitive strain. In contrast, the UCD center used a unified UI that reduced task times and decreased cognitive strain. In both centers, the workflow was reported to be stressful at times; however, management at both companies employed high involvement work processes that mitigated this stress. Implications for possible high-involvement UI design are considered and a strategy for applied ethnographic research is discussed.
Practitioner’s Take Away
The study revealed several practical implications for practitioners to consider:
- Ethnographic research methods complement experimental methods, are effective for exploring new use environments, and may reveal unexpected data.
- The time required to apply traditional ethnographic research methods can be condensed by narrowing the focus using a four-phase approach that includes (1) reviewing existing ethnographies and related research, (2) phased ethnographic interviews, (3) on-site observation, and (4) documentation of artifacts and traces.
- Productivity of contact-center agents is extensively measured and very sensitive to the usability of agent-facing software (AFS). Shaving off a few seconds of task time per call can radically reduce contact-center operating costs, making user-centered design an attractive investment.
- Consider that the customer is an ever-present third party in the agent-AFS interaction. For example, the agent needs to sound competent to the customer while interacting with their system. Long pauses must be avoided and data exchange between agent and customer must respect social norms of conversation. To facilitate this, AFS must support rapid and easy data retrieval (to prevent long pauses) and flexible data entry (to allow the agent to record data while the conversation progresses naturally).
- To maximize agent efficiency consider organizing GUIs by agent task rather than by the underlying technology. Enhancing agent efficiency also enhances quality of service.
- Consider how software can be designed to support high involvement work processes. Any design decision that gives agents control over how their work is structured without compromising organizational effectiveness is likely to reduce agent stress and improve service quality.