Contact Center: The Efficiency KPI That Hides Rework
A telecommunications provider runs a customer service center with 180 agents. It handles about 1,200 calls a day. Average handling time is 420 seconds, service level is defined as 80% of calls answered within 30 seconds, but actual service level is 68%. First-contact resolution is 74%, occupancy is 88%, schedule adherence is 81%, and quality audit scores average 82 out of 100. The operations director wants to cut average handling time by 10% and use the saving to reduce headcount. Team leaders warn that agents already skip after-call work to protect the service level, and repeat calls are rising. The finance department wants a single efficiency KPI for the monthly board pack. Customer complaints about repeat calls have increased 18% in two quarters. The center has no measure for backlog or for calls that return within seven days.
1. Critique the proposal to use average handling time as the single board-level efficiency KPI. What behaviour is it likely to drive?
Read the worked answer
Average handling time is a local input measure, not an outcome measure. It can be reduced by genuinely simplifying work, but also by transferring calls, ending calls before resolution, skipping after-call work, or discouraging agents from documenting cases. Those behaviours lower AHT while increasing repeat contacts, customer effort and total cost. The center already has evidence of this risk: repeat calls are rising and complaints about repeat calls are up 18%. Using AHT as the single efficiency KPI would make the board blind to the difference between a cheap call and a resolved call. Efficiency should relate outputs to inputs, for example cost per resolved contact, not seconds per call in isolation. A single KPI also creates a target for gaming. Agents and team leaders will optimize what is measured, especially if headcount depends on it. The finance department's desire for simplicity is understandable, but a single number should be a composite or a small scorecard with guardrails, not an unqualified AHT target. The board pack could lead with cost per resolved issue and show AHT as a diagnostic, alongside first-contact resolution, repeat contact rate and service level.
2. Design a balanced KPI set for the center. Include definitions and at least one guardrail metric. Explain how the metrics interact.
Read the worked answer
A balanced set should cover customer, operational and cost perspectives. Customer metrics: service level, defined as the percentage of calls answered within 30 seconds; first-contact resolution, defined as no repeat contact on the same issue within seven days; repeat contact rate; and customer satisfaction. Operational metrics: quality audit score, schedule adherence, occupancy with a target range rather than a maximum, and backlog in call-minutes. Cost metric: cost per resolved contact. Average handling time should remain a diagnostic, not a standalone target. Guardrails should include a minimum quality score, a maximum repeat contact rate, and a requirement that after-call work is completed within the shift. These guardrails prevent service level from being bought by skipping documentation or rushing customers. The metrics interact: cutting AHT may improve service level temporarily by freeing capacity, but if first-contact resolution falls, repeat contacts consume the saved capacity and cost per resolved contact rises. High occupancy looks efficient but leaves no slack for coaching, breaks or demand variation, and can increase errors. The scorecard should be reviewed together, with a rule that no headcount reduction is approved unless quality and repeat-contact guardrails remain within tolerance for at least two consecutive months.
3. The director wants to cut AHT by 10% and reduce headcount. What analysis would you do before agreeing, and what would you recommend if repeat calls are rising?
Read the worked answer
Before agreeing, I would analyse the call mix and the handle-time distribution by call type, customer segment and agent tenure. I would separate talk time, hold time, after-call work and transfer time. I would sample calls to identify avoidable contacts caused by poor bills, failed self-service, missing information or upstream process errors. I would compare first-contact resolution and repeat contacts for high-AHT and low-AHT agents to see whether longer calls are resolving issues. I would also review occupancy and shrinkage to check whether the center is already understaffed, since 88% occupancy with 68% service level suggests capacity strain. If repeat calls are rising, a 10% AHT cut is risky. The likely cause is that agents are already shortening calls to protect service level, and the work is returning as repeat contacts. My recommendation would be to pause headcount reduction, set a repeat-contact guardrail, invest in first-contact resolution, knowledge management and self-service for the avoidable contact types, and pilot any AHT initiative on a controlled team. Headcount should follow validated demand reduction, not an AHT target. The business case should show cost per resolved contact, not just seconds saved.