Call centre quality assurance is the process of checking whether customer interactions meet agreed standards, then using the findings to improve the work. A useful scorecard tells a manager what happened, why it matters and which behaviour needs attention. It does more than turn a conversation into a percentage.
This guide gives you a practical quality assurance scorecard template, a calibration process and a coaching rhythm for an in-house, outsourced or blended team. If you need a review of the whole operation, start with our contact centre performance audit. Here, the focus is narrower: making interaction reviews consistent enough to act on.
What should call centre quality assurance measure?
Measure the actions that protect the customer and help resolve the reason for contact. A greeting can be correct while the answer is wrong. An adviser can sound confident while creating work for another team. A short call can become an expensive call when the customer has to start again tomorrow.
Build the form around five questions:
- Did the adviser understand the customer's actual need?
- Was the information and action accurate?
- Did the interaction move the issue towards a suitable resolution?
- Were the required controls followed?
- Does the customer know what happens next?
These questions apply across voice, email and live chat, but the evidence differs. Listening and turn-taking matter in a call; a written reply needs readable structure, complete instructions and the right links. Use channel-specific examples under a shared quality standard.
A quality assurance scorecard template
Download the editable quality-assurance scorecard (CSV). It opens in a spreadsheet and includes space for the rating, evidence, customer impact, action owner and review date. Add the interaction reference and form version using your approved internal system, without putting customer details in an uncontrolled copy.
Use the categories below as a starting form. For each criterion, record met, partly met, not met or not applicable, together with evidence and the next action. Define those ratings before the form is used. They are not industry benchmarks or a ready-made contractual standard.
Understanding the enquiry
Look for: a clear explanation of the issue, relevant questions and acknowledgement of information the customer has already supplied. The adviser should not make the customer repeat a history that is already available and safe to use.
Record: what the adviser understood and any missing fact that changed the response. “Needs better listening” is not specific enough. “Asked for the order number after the customer had already provided it, then answered for the wrong order” is something a coach can work with.
Accuracy and judgement
Look for: correct use of the knowledge base, appropriate policy interpretation and actions within the adviser's authority. Distinguish a knowledge gap from a policy that genuinely offers no workable answer.
Record: the incorrect advice or action, the approved alternative and whether the knowledge article itself needs changing. Do not mark down an adviser for following a published instruction that the business now dislikes.
Resolution and ownership
Look for: action that addresses the contact reason, a justified escalation where required, accurate case notes and a named owner for unfinished work. Escalation is not automatically a failure. An unnecessary transfer and a necessary specialist handoff are different events.
Record: the outcome, the remaining dependency and who owns it. Link repeat contacts back to the original issue where your systems and permissions allow it. Our contact centre metrics guide explains why quality scores need resolution measures beside them.
Communication and customer effort
Look for: clear language, an appropriate tone, useful explanations and a realistic expectation of what happens next. Reward clarity rather than a rigid script where the same meaning can be expressed well in different ways.
Record: the exact behaviour that helped or hindered the customer. Replace “show more empathy” with a concrete observation, such as acknowledging the disruption before explaining the options.
Control and documentation
Look for: required verification, authorised data access, appropriate disclosures and complete records for the work in scope. Security and compliance owners should define the controls that apply to the actual service.
Record: the control, the evidence and the escalation required if it failed. Do not let a critical failure disappear inside a good average for friendliness and presentation.
Separate critical errors from development feedback
A scorecard needs a clear critical-error rule. Agree which failures require immediate intervention, who confirms the finding, how the customer is protected and how the result is reported. A mandatory verification failure should not be treated as equivalent to an awkward phrase.
Keep two views: the interaction's developmental assessment and any confirmed critical failure. This avoids a single score concealing the reason an interaction was unsafe, inaccurate or incomplete. Give advisers a way to challenge an assessment with evidence.
If you use weighted percentages, document the weights, treatment of not-applicable criteria and pass rules. Changing a form changes the meaning of its score. Keep the version and effective date so a rise in scores is not mistaken for improvement when the scoring rules simply became easier.
Calibrate before comparing teams or suppliers
Calibration means different reviewers assess the same interactions independently, then agree how the standard applies. Start with a mix of ordinary, difficult and ambiguous contacts. Review both the rating and the evidence behind it.
Use this meeting sequence:
- Each reviewer scores the interaction before the meeting.
- Compare criterion-level differences, not just total scores.
- Decide whether the difference came from evidence, interpretation or unclear wording.
- Update the guidance and example library where needed.
- Recheck a comparable interaction to see whether the ambiguity has been resolved.
NiCE's quality-monitoring guidance discusses scorecards, calibration and keeping evaluation forms relevant. The practical point is simple: a form is not reliable merely because it has been approved in a meeting.
Choose a sample that answers the question
There is no universal number of calls that makes a review representative. A small coaching sample can help an individual practise a skill, but it cannot establish the quality of an entire operation. State what the sample can and cannot support.
Include a random baseline alongside targeted reviews of complaints, repeat contacts, escalations and new processes. Label targeted findings separately, because a sample deliberately chosen for problems should not be presented as the average customer experience.
When comparing an internal team with an outsourced team, align channel, contact reason, complexity, period and form version. A supplier handling difficult escalations should not be judged against a team answering straightforward order updates without accounting for the difference.
Automation can help locate interactions and surface patterns. Test its findings against human review, including missed issues and false alerts. Do not turn an unvalidated automated score into a high-impact employment decision.
Turn the finding into a coaching action
Choose one observable behaviour. Show the evidence, ask the adviser to explain their judgement, rehearse a better response and check it in later work. Keep the coaching action small enough to practise during a normal shift.
For example, if a customer leaves without a clear follow-up, the action might be to finish by confirming the owner, action and agreed update time. Review whether those elements appear in later eligible interactions. The action is not simply “improve the QA score”.
Supervisors need time and skill to do this. Frontline leadership training can strengthen coaching conversations, while customer service training can address specific interaction skills. If the barrier is a broken system or contradictory policy, fix that barrier as well.
Apply the same standard to an outsourced operation
Agree who owns the form, who reviews disputes and how quickly confirmed problems reach an operational owner. Both sides should be able to inspect the evidence they are authorised to see. A supplier's summary percentage is not enough to explain repeated customer complaints.
Connect the quality standard to your customer-service SLA and provider governance. Use separate measures for access, resolution and critical failures so speed cannot hide poor work. If late or weekend shifts perform differently, review the after-hours support model rather than assuming the same staffing and escalation arrangements work around the clock.
For independent help with calibration, supplier reviews and operational priorities, explore contact centre consulting.
Frequently asked questions
What is the difference between quality assurance and quality monitoring?
Monitoring is the observation and evaluation of interactions. Quality assurance is the wider system of standards, sampling, calibration, feedback and improvement that makes those evaluations useful. Organisations may use the terms differently, so agree the scope.
How often should a call centre scorecard be reviewed?
Review it when products, policies, channels or customer risks change, and when calibration exposes ambiguity. Keep a planned review date as well. Do not change scoring rules silently during a supplier comparison.
Should every adviser have the same scorecard?
Use shared principles with criteria matched to the work. Different channels or specialist roles may need different evidence. Keep comparable criteria consistent and explain which differences prevent a direct score comparison.
Can a good QA score prove customers are satisfied?
No. QA measures performance against the chosen form. Customer feedback, resolution, repeat contacts and complaints provide different evidence. Read them together before claiming that service has improved.
Can an outsourced provider run quality assurance for us?
Yes, if ownership, reviewer competence, calibration, access and independent checks are agreed. The client still needs enough visibility to challenge findings and act on problems that sit outside the provider's control.
About the author
Oliver Catt is the founder of The Cattalyst. He has spent 25 years inside contact centres, customer service, sales and training operations, and now helps organisations develop their leaders and build outsourced teams through a global partner network.
Start a conversation with Oliver
