How Is a Call Evaluated? The QA Scorecard, Item by Item
What does a quality analyst listen for when scoring your call? The usual scorecard sections, their weights, critical errors, and a template you can start from.
Almost every contact center has a quality scorecard: a list of items an interaction is scored against, adding up to 100. Scorecards vary, but most good ones are built on the same sections. This article explains them, whether you're an agent who wants to understand how you're scored, a team leader, or building your first scorecard.
The six sections of most scorecards
| Section | What is scored | Suggested weight |
|---|---|---|
| 1. Opening and verification | Greeting, introduction, verifying the customer's identity per procedure | 10 |
| 2. Understanding the request | Listening, asking the right questions, summarizing the issue before solving it | 15 |
| 3. Resolution and accuracy | Correct information, actually solving the problem, no need for the customer to come back | 35 |
| 4. Communication and empathy | Clarity, tone, empathy, avoiding internal jargon | 20 |
| 5. Closing | Confirming the fix, explaining next steps, asking "anything else?" | 10 |
| 6. Documentation | Logging the interaction, its reason and the action taken correctly | 10 |
The weights are suggestions; each center sets its own. But notice that resolution and accuracy carries the most weight. The customer called to fix a problem, not to hear a perfect greeting. The effect of first contact resolution on satisfaction is covered in 10 Daily KPIs.
Critical errors: when the rest of the score doesn't matter
Some errors are serious enough that the call fails no matter how good the rest was. These are critical errors, usually in three groups:
- Errors that harm the customer: wrong information that costs them, or a promise that can't be kept.
- Errors that harm the organization: a regulatory breach, or disclosing customer data without verification.
- Serious conduct errors: being rude to the customer, or hanging up on them.
Everything else is non-critical: it costs points but doesn't fail the whole evaluation. A clear line between the two keeps scoring fair: forgetting the customer's name isn't the same as reading out someone else's account number.
How to write the items
Items you can observe
"Agent showed professionalism" is vague, and every evaluator will read it differently. "Agent summarized the customer's request before starting to solve it" is something any evaluator can judge yes or no. The more observable the items, the less evaluators disagree, which is what calibration sessions address.
Yes / No / Not applicable
Most items work better with three answers: yes, no and not applicable. "Agent explained the service fee" doesn't apply to a call asking about an appointment time. Items that don't apply are left out of the calculation, so agents don't lose points for something that wasn't needed.
A comment with every deduction
A score alone teaches nothing. Every deduction needs a written comment: what happened, why it matters, and what to do instead.
What many scorecards miss: the customer's view
A high quality score doesn't always mean a satisfied customer, because the scorecard measures what the company wants from the inside, not what the customer feels. SQM's numbers on this are in What Does a Quality Analyst Do? So:
- Compare quality scores with customer satisfaction on the same calls whenever you can.
- If an item has nothing to do with customer satisfaction or compliance, ask why it's on the scorecard.
- Review the scorecard at least once a year, or when products and procedures change.
If you're an agent: how to use your evaluation
- Read the comments, not just the score.
- Ask to listen to the call with your team leader when you don't understand a deduction.
- Focus on resolution and accuracy first. It carries the most weight and has the biggest effect on customer satisfaction.
What this means for your team
Take your current scorecard, read each item aloud, and ask: could two different evaluators judge this the same way? Rewrite every item that fails the test.