Home / Guides / Call QA scorecards
Method · 7 min read
What call QA scorecards measure, and what they structurally cannot
A QA scorecard is a good instrument pointed at the wrong target. It measures how well a person executed a call. It has no line for whether the customer ended up better off, because that usually is not the agent's doing.
What a typical scorecard measures
Vary the wording and almost every call QA form in the country covers the same ground:
| Criterion | What it is really testing |
|---|---|
| Greeting and identification | Compliance with a script |
| Verification | Compliance with a process |
| Active listening / empathy | Manner |
| Product knowledge | Competence |
| Ownership and next steps | Whether a promise was made |
| Call control and efficiency | Handling time |
| Closing | Compliance with a script |
Every one of these is about the agent. That is not a criticism - the form exists to support coaching, and coaching is about the agent. The problem is what happens when the same form becomes the organisation's only lens on call quality.
The line that is missing
Look down that list for the criterion that says "and the thing the customer needed actually happened". It is not there, and it cannot be, because on most calls the agent cannot control it.
The agent promised a callback. Whether the callback happened depends on a workflow. The agent submitted the change. Whether the change went through depends on a system. The agent escalated. Whether the escalation was picked up depends on someone else entirely.
A scorecard evaluates the part of the interaction that ends when the call ends. Most service failures happen after that point, or were already in motion before the call started.
What a perfect-score failure looks like
Consider a call that scores full marks on every criterion above. Correct greeting, clean verification, genuinely warm, knowledgeable, took clear ownership, set out next steps, closed properly, handled in under four minutes.
The customer is calling for the third time about the same unresolved matter.
Nothing on the form has a place to record that. The scorecard is evaluating this call in isolation, and in isolation it was handled well. The failure only becomes visible when you look across calls, over time, at the same customer and the same matter - which is exactly the view a per-call scorecard does not produce.
The most useful question about a call is rarely "how well was this handled". It is "is this the first time we have heard about this".
The cause mix nobody scores
When calls are reviewed for outcome rather than execution, the distribution of causes tends to surprise people. In a measured month at an Australian corporate travel agency, of the calls raised as genuine service failures, roughly two thirds were technology or system failures and around one in eight related to how the call itself was handled.
A programme that only runs scorecards would have found the one in eight and been structurally blind to the two thirds. Worse, it would have reported the one in eight as though it were the whole picture, which quietly turns an operational problem into a people problem.
Fixing the scorecard without breaking it
The answer is not a longer form. Adding outcome criteria to an agent scorecard makes it unfair, because you end up scoring people on things they did not control.
The answer is a second, separate view:
- Keep the scorecard for coaching, on a sample, scored by a human. It is good at that.
- Add an outcome review across all calls that asks whether the customer was left worse off, regardless of how well the call was executed.
- Route the two differently. Coaching findings go to team leaders. Process and system findings go to whoever owns the process or the system. If both go to the team leader, only one of them ever gets acted on.
- Never merge them into one score. A blended number hides exactly the distinction you built the second view to see.
A note on fairness
There is a real risk in adding any outcome-based review, and it is worth naming. If the alerts land on a team leader's desk and the only tool they have is a conversation with an agent, then in practice every alert becomes an agent problem - including the two thirds that were never about the agent at all.
Teams work this out quickly, and once they do, the value of the system drops sharply. The safeguard is structural rather than cultural: give the operational findings a route that is not the coaching route, and say out loud what the cause mix actually is. "Two thirds of what this finds is our systems, not our people" is both true and the most useful thing you can tell a team.
Common questions
What is a call QA scorecard?
A structured form used by a team leader to score a sampled call against criteria such as greeting, verification, empathy, product knowledge, ownership and closing. It is designed to support coaching of an individual agent.
Can a call score well on QA and still be a failure?
Yes, and it is common. The scorecard measures how the agent executed the call. Whether the customer's problem was actually resolved usually depends on a workflow, a system or another team after the call ends - none of which the form has a line for.
Should QA scorecards include outcome measures?
Adding outcome criteria to an agent scorecard makes it unfair, because agents get scored on things they did not control. A better structure is two separate views: a human-scored sample for coaching, and an outcome review across all calls routed to whoever owns the process.
See what a full review of your calls turns up
WiseSentry reads every recorded call and raises only the ones where your business let the customer down. Talk to us about a review of your own call history.