We’re obsessed with metrics. We have a greater capability to measure business performance than ever before. We have baselines, benchmarks, and trends, arranged in reports and dashboards for easy consumption. AI analyzes every call to the contact center. Automation sends a survey after every support ticket is closed. We fixate on the results, creating scorecards and setting improvement targets.
But is the customer experience getting any better?
According to an IDC white paper published in January 2022, 87% of companies believe they provide an excellent customer experience—but only 11% of customers agree. Despite our obsession with metrics, we aren’t acting on them effectively.
Consider Chad, a newly promoted VP of Professional Services at a software company. He faced a difficult challenge: improving project turnaround times. Customers were complaining about the lengthy turnaround time to complete a custom programming project.
Chad began measuring the productivity of his programmers. On average, he found, a programmer could code 32 projects per month—but individual performance ranged from 4 to 48 per month. Clearly an opportunity for improvement.
Chad designed and implemented a performance management system. He gave each programmer a goal of completing 32 projects per month, which was reasonable considering average performance. A project was “complete” when the programmer sent the code to QA for review.
If programmers met or exceeded the target, Chad rewarded them with a cash bonus. If they missed the target for two consecutive months, he put them on a Performance Improvement Plan. If they missed the target a third time, he fired them.
At first, the system looked like a raging success. Programmer productivity skyrocketed, and so did morale—everyone loved the cash bonuses at the end of the month.
But the overall turnaround time of projects got worse. Programmers were sending more projects to QA than ever before, but customers were waiting even longer to get the results. What happened?
When we analyzed the root cause of the problem, we found a troubling statistic: QA’s rejection rate had increased—especially for projects submitted near the end of the month.
The crafty programmers had figured out a workaround to beat the system. At the end of the month, with the deadline looming, they submitted incomplete projects to QA. They achieved the 32-per-month target and earned the bonus instead of the punishment.
QA took several days to test and find the errors, and by the time they rejected the projects, it was a new month—and the programmers could finish the projects they’d started the month before and rack up some quick wins against the new month’s quota.
Can you really blame the programmers? If your job—your ability to feed your family—was on the line, what would you do?
By incentivizing programmers to achieve a metric-based target, Chad created unintended consequences. Specifically, Chad experienced a perverse result—unintended outcomes that were the opposite of what he intended. His goal was to improve the customer experience, but instead he made it worse.
Unintended Consequences of Metric-based Targets
Chad’s not the only well-intentioned leader to set metric-based targets and suffer from unintended consequences.
We know that customers calling a company for support want the most efficient call possible, without dead air and long wait times. Therefore, many contact centers measure Average Handle Time (AHT), include it on agents’ scorecards, and incentivize them to keep it low.
The rationale makes sense: shorter calls are good for customers who don’t want to wait, and good for the company to keep costs down. But when the metric becomes a target, it ceases to be a good metric. (That’s Goodhart’s Law.)
The intent of measuring agents on AHT is to reward efficiency and provide better customer service. But instead, agents hurry through calls and provide worse service—leading to callbacks and escalations.
In the Harvard Business Review article, Call Length is the Worst Way to Measure Customer Service, authors Pete Slease, Rick DeLisi, and Matthew Dixon show how incentivizing agents to keep AHT down leads to worse customer service.
When frontline agents are incentivized to lower AHT—and if they know that they’ll earn a reprimand from their supervisor for taking too much time with customers—they tend to rush their interactions, even if the customer’s issue demands more time.
When rewards or punishment is linked to a metric, employees will find a way to hit the target—even if it’s the opposite of what you hope they’ll do.
Solving Complex Operational Problems to Improve Customer Experiences
Whatever the business challenge—improving turnaround times, providing efficient service on the phone, or something else—it’s likely part of a complex system. The surface problem is rarely the real problem. We can’t simply identify a metric, create improvement targets, and expect that everyone will know and follow the best path to improve customer experiences.
Chad hoped that the programmers would write quality code, but he incentivized them to push mediocre or incomplete code to QA. And a deeper problem was lurking under the surface—some projects were never completed.
While Chad knew that every project required a different level of effort, he believed they’d average out in the end. The 32 project-per-month target allowed for a mix of easy and hard projects.
But the programmers—just like contact center agents—don’t think in averages. Faced with the penalty of missing the quota, they always worked on the easiest projects assigned to them. They would only work on a more difficult project when they’d exhausted their personal supply of easy projects. With a constant influx of new projects, that meant the most difficult projects languished until an irate customer escalated.
To solve Chad’s operational problems and improve project turnaround times, we first analyzed the root causes of the problem. We identified several factors contributing to lengthy turnaround times. Chad’s performance management system was a leading cause, but we also found process and cultural issues we could address through improvement initiatives.
We abolished the individual quotas and created team goals that aligned with customer needs. Instead of competing for bonus money, the programmers began helping each other. They worked together to accomplish their shared goal every day: completing the oldest projects in the queue. The new system, aligned with customer goals, improved turnaround times by over 40%.
Metrics show you how your business is performing—not how to improve that performance. Don’t set metric-based targets and hope everyone will do the right thing for customers. If you want to improve customer experiences, understand the root causes of operational problems and reward employee behaviors that align with customers’ needs and goals.