Call center agent scorecard template

What goes on a call center agent scorecard

The scorecard is the monthly picture of an agent's performance. It pulls together results from systems the agent touches every day: QA evaluations, post-contact surveys, the ACD or telephony platform, and workforce management. Most centers land on five to seven metrics. The usual set is QA score, customer satisfaction, first contact resolution, average handle time, schedule adherence and attendance, plus one program metric for sales or collections queues.

The QA number on this card comes from individual call reviews scored on a call center QA scorecard. Keep the two documents separate: one evaluates calls, the other evaluates a period of work. For the concept and how it differs from a dashboard, see our definition of an agent scorecard.

Setting weights that drive the right behavior

Weights tell agents what you actually care about, so they will optimize for them. If handle time carries 30% and quality 15%, expect agents to rush customers off the phone, transfer hard calls and create repeat contacts. We recommend at least half of the weight on quality and customer outcomes (QA, CSAT, FCR) and no more than 10 to 15% on handle time.

Treat AHT as a range rather than a race. Scoring "within 10% of target" as a 3 and penalizing both very long and very short calls stops agents from ending calls early to hit a number. Very short calls are often the problem: dropped customers, skipped verification, or calls released to get back into the queue.

Adherence and attendance deserve their own weight in high-volume centers because staffing plans depend on them. A 300-seat operation where adherence slips from 92% to 85% has effectively lost about 20 agents' worth of available time in every interval.

Worked example

Marcus works a billing queue. His period results and points:

• QA average 91% (target 85%): 4 points x 30% = 1.20
• CSAT 4.3 of 5 (target 4.2): 3 points x 20% = 0.60
• FCR 68% (target 72%): 2 points x 15% = 0.30
• AHT 412 seconds (target 420, within range): 3 points x 10% = 0.30
• Adherence 94% (target 90%): 4 points x 15% = 0.60
• Attendance, 1 occurrence (target 2 or fewer): 4 points x 10% = 0.40

Weighted score = 3.40, which rates as Meets. The review focus is FCR: his calls sound good, but customers call back. Pulling his repeat contacts usually shows a pattern, such as not confirming a refund timeline or leaving a ticket open, and that becomes the coaching plan.

Fairness, context and legal notes

Exclude time the agent could not control. If a CRM outage doubled everyone's handle time on a Tuesday, remove that day or adjust the target. Compare agents only within the same queue and tenure band; a nesting agent and a three-year veteran should not be on the same stack rank.

Attendance metrics need care. Absences protected under the FMLA, the ADA (as a reasonable accommodation), state paid sick leave laws, jury duty or military leave must not count against the agent. Have the scorecard pull from a time-and-attendance code set that separates protected leave from unplanned absences. Rules vary by state and city, so check local paid sick leave laws.

If you use scorecards in pay, bonus or termination decisions, apply the same metrics and bands to everyone in the role and keep the records. Scorecards that feed into a performance review become part of the personnel file.

Common mistakes

• Too many metrics. With ten items, each carries so little weight that nothing moves the total, and agents cannot tell what to work on.
• Changing targets mid-period. Set them before the period starts; change them only at the next one.
• Using team averages as targets. Half the team will always be below average by definition.
• Scoring QA from one or two evaluations. Require a minimum count, often four, before the QA metric counts; otherwise mark it as insufficient data.
• Treating the score as the review. The number opens the conversation. The value is in finding out why FCR is low and agreeing what changes.

Running scorecards across sites and large teams

At scale, nobody should be filling these in by hand. Build the scorecard in your reporting tool so each agent's card generates automatically at period end, and spend supervisor time on the review conversation instead. Reporting tools that pull from your telephony, QA and WFM data can produce a card per agent and roll them up by team.

For multi-site programs, compare team and site distributions, not just averages. If one site has 40% of agents rated Outstanding and another has 10%, check whether the QA evaluators are calibrated and whether the queues are the same before drawing conclusions. Site and zone performance views help here because they show the spread side by side.

Tenure matters too. Track scorecard ratings by hiring class and by recruiting source. If agents from one source consistently reach a Meets rating two months later than others, that is useful information for your recruiting and training plan.