My time as an outsourced call center agent was when I most acutely felt the impact of being managed by metrics. It was one of the most dehumanizing experiences I$$ve had, feeling like a human being enslaved by the machine. This was only surpassed by my previous job as a child, where I harvested cucumbers on a mechanized farm.

Call center agents were measured on metrics like first call resolution $($FCR$),$ average handle time $($AHT$),$ and customer satisfaction $($CSAT$),$ based on post$-$call surveys $($Dialpad$,2023).$

Although these metrics are not inherently bad, their implementation was sorely lacking, leading many agents to game the system. Both agents and management misunderstood the value of AHT. Some agents would rush customers off the phone as the call length approached the target, resulting in more irate customers calling back. When a complex case required longer resolution, managers would pressure agents to shorten their calls. Some agents even lied to customers to boost their CSAT scores or ensured customers would only call back after the prescribed delay to avoid affecting their FCR scores.

The results in this environment were suboptimal, not because these metrics lacked value, but due to their poor implementation. This was made evident later on when I worked for Apple, where AHT was a reflection of knowledge and skill rather than an objective, and the most important metric was Quality, based on spot checks from a quality control team. The feedback loop between training and key performance indicators ensured a human$-$centric approach where call center agents had an opportunity to grow, rather than the fear of losing their job due to not reaching arbitrary targets meant to ensure the profitability of the third$-$party call center.

Perhaps scientific measurements of work processes are only as good as the person using them?