Free Training & Career Tips... Subscribe to Get Weekly Career Tips
By Subscribing You are Agreeing to Terms and Conditions
A QA scorecard is the form a reviewer uses to rate a customer interaction against agreed, weighted criteria, so that every call, email or chat gets a score you can compare fairly. In customer service QA, those scores show team leaders where to coach each agent and where the whole team needs training. If your team needs the skills the scorecard measures, see BOTI’s sales, marketing and customer care training courses.
If you manage a contact centre or service team, you probably already see the gap. Customers complain, satisfaction scores move up and down, and you have no consistent way to say why one agent’s handling works and another’s frustrates. A QA scorecard closes that gap. It gives you objective criteria, calibrated reviewers and a clear line from “what we measured” to “what we coach” to “what we train”. This guide covers what service QA is, how to build a scorecard, how partial scoring works, how to calibrate reviewers and how to coach from the results.
Need a starting point? Ask BOTI to help you build a customer service QA scorecard with weighted criteria, calibration notes and a coaching log that suit your team.
What is customer service quality assurance?
Customer service quality assurance (QA) is a repeatable system for sampling interactions, scoring them against agreed standards and feeding the results into coaching and training. It answers three manager questions: Are we delivering the service we promised? Where exactly are we falling short? Who needs help with what?
Call centre quality assurance covers voice and digital channels: greeting and verification, problem resolution, tone and empathy, compliance and call closure. The same principles apply to email, live chat and walk-in service teams.
A mature QA programme has four parts:
- A scorecard: the criteria and weights you score against.
- A sampling plan: how many interactions per agent, per period, and how they are chosen.
- Calibration: keeping reviewers consistent so that a “4” means the same thing to everyone.
- A feedback loop: coaching sessions and training that act on what QA shows.
Without the loop, QA becomes a scorekeeping exercise that agents resent. With it, QA drives steady service improvement.
How to build a QA scorecard
Your QA scorecard is the heart of the programme. It should reflect what your customers value and what your business needs, not a generic checklist. Build it in five steps.
1. Define your categories. Group criteria into a handful of themes. A typical service or call centre scorecard uses:
| Category | What it measures | Suggested weight |
|---|---|---|
| Opening & verification | Greeting, brand intro, identity/security checks | 10% |
| Communication & tone | Clarity, active listening, empathy, professionalism | 25% |
| Problem resolution | Accuracy, completeness, first-contact resolution | 30% |
| Process & compliance | Following procedure, POPIA/data handling, disclosures | 20% |
| Closing | Summarising, next steps, courteous close | 15% |
2. Write observable criteria. Each line item must be something a reviewer can see or hear, not a feeling. “Showed empathy” is vague. “Acknowledged the customer’s frustration and confirmed understanding before solving” can be scored.
3. Choose a scoring scale. Keep it simple: a 1-5 scale, or yes/no/partial per item rolled into a weighted percentage. Reserve “auto-fail” flags for non-negotiables (rudeness, a POPIA breach, incorrect financial information) that zero the whole interaction.
4. Weight what matters. Resolution and tone usually carry the most weight because they drive satisfaction and retention, with compliance weighted enough to be taken seriously. Avoid spreading weights evenly. That tells agents everything matters equally, so nothing stands out.
5. Pilot and refine. Score 15-20 interactions, review the results with two or three team leaders, and fix any criterion that is ambiguous or rarely used. A scorecard is a living document, so revisit it quarterly.
If you would rather not start from a blank page, ask BOTI to help you set up a QA scorecard with these five categories and weights you can adjust.
Partial scoring and calculating the QA score
Many teams use a yes/no/partial scale for each line item because it is quicker and more consistent than a 1-5 scale. “Partial” is the option reviewers argue about most, so define it in writing before anyone scores.
- Yes: the agent did everything the criterion describes. Full points.
- Partial: the agent did part of it, or did it late. For example, the agent verified the caller’s identity but only after discussing account details. A common rule is half the item’s points.
- No: the behaviour was missing. Zero points.
- Not applicable: the criterion did not arise on this interaction. Remove it from the total rather than scoring it zero, so agents are not penalised for situations that never happened.
Write one example of “partial” for each criterion into your scorecard’s guidance notes. That single step removes much of the disagreement you would otherwise see in calibration.
To get the overall score, work out each category as a percentage of its available points, multiply by the category weight, then add the results. Using the weights in the table above, an illustrative interaction might score:
| Category | Category result | Weight | Contribution |
|---|---|---|---|
| Opening & verification | 100% | 10% | 10 |
| Communication & tone | 80% | 25% | 20 |
| Problem resolution | 70% | 30% | 21 |
| Process & compliance | 100% | 20% | 20 |
| Closing | 60% | 15% | 9 |
The interaction scores 80%. If an auto-fail item had been triggered, the score would be zero, whatever the other categories showed. Keep the category results visible to agents as well as the total. “Resolution 70%” gives a coach something to work on, while “80% overall” does not.
QA scorecards and customer scorecards
A QA scorecard is your internal view: a trained reviewer judging an interaction against your standard. A customer scorecard is the customer’s view: satisfaction survey results, complaints and repeat contacts about the same issue. You need both.
- High QA scores and unhappy customers usually mean the scorecard measures the wrong things. Agents follow the script, but the customer’s problem is not solved. Shift weight towards resolution.
- Low QA scores and happy customers often mean the scorecard is too strict on process items customers do not notice. Review those criteria, but keep genuine compliance items.
- Both low points to a real skills gap that coaching or training needs to close.
Comparing the two every month keeps your QA scorecard honest and tied to what customers actually experience.
Calibrating quality: making scores mean the same thing
A scorecard is only as trustworthy as the people using it. If two team leaders score the same call 90% and 65%, agents lose faith in the whole programme. Calibration fixes this. It is a recurring session, fortnightly or monthly, where everyone who scores interactions reviews the same call independently, then compares results and talks through the differences until they share one interpretation of every criterion.
Run calibration like this:
- Pick one or two interactions in advance (mix an easy one and a contentious one).
- Each reviewer scores it alone, with no discussion.
- Compare scores line by line, flagging any criterion where reviewers differ by more than one point.
- Debate the gap until you reach a shared rule, then write that rule into your scorecard’s guidance notes.
- Track your calibration variance over time, aiming to keep reviewer agreement within an agreed tolerance.
Calibration is also where you catch a vague scorecard. If reviewers keep disagreeing on the same line, the criterion is the problem, not the reviewers.
Coaching from QA data
QA scores are not a stick. Their purpose is to drive better conversations between team leaders and agents: the score tells you where to look, and coaching is what you do about it.
A simple, effective coaching rhythm:
- Start with the agent’s view. Have them self-assess first. Gaps they find themselves stick far better than gaps you point out.
- Focus on one or two behaviours. Pick the highest-impact fix, usually in resolution or tone, and work on it until the next review.
- Use the recording, not just the number. “Here is where the customer asked twice. What could we have done?” works better than “Your resolution score was 3.”
- Set a clear, observable goal and date the next check-in.
- Log it. Keep a short coaching record per agent to show progress and to spot agents who need formal training, not just coaching.
Patterns across the team matter as much as individual scores. If most of your agents score low on the same criterion, that is not a coaching problem. It is a training gap.
Linking QA to training
Your scorecard data is a continuous training-needs analysis. Read it that way.
- Individual gaps that coaching cannot close: an agent who keeps mishandling difficult callers despite feedback needs structured skills training, not another pep talk. See Handling Difficult Customers: Training for Teams.
- Team-wide gaps: if resolution or empathy scores are low across the team, a structured programme usually works better than one-on-one coaching. Our guide to Customer Service Training That Lifts Retention covers the tone, empathy and resolution behaviours QA measures.
- New-hire ramp: feed your scorecard criteria straight into onboarding so agents know the standard from day one.
- Channel-specific gaps: voice teams often need different coaching from chat teams. Our Call Centre Training guide covers voice-specific behaviours.
The loop is simple: QA reveals the gap, coaching closes small gaps, training closes structural ones, and QA confirms it worked. For how to choose and run training around your QA results, see our guide to customer service QA training, and for the wider picture, Customer Service & Call Centre Training (South Africa).
Putting it together: your first 90 days
| Phase | Weeks | Focus |
|---|---|---|
| Set up | 1-3 | Build and pilot the scorecard; agree the sampling plan |
| Calibrate | 4-6 | Run the first two calibration sessions; lock reviewer consistency |
| Coach | 7-10 | Begin per-agent reviews and coaching logs |
| Train | 11-13 | Translate recurring gaps into a training plan |
By the end of one quarter you will have a calibrated scorecard, a coaching rhythm and an evidence-based training plan, all pointing at the metrics your business cares about.
Build your QA programme with BOTI
Setting up service QA is easier with help from people who have done it before. Two ways to start:
- Ask BOTI to help you design your QA scorecard, including weighted criteria, partial-scoring rules, calibration notes and a coaching log.
- Request a quote for QA, coaching and service skills training for your managers and agents. For the agents themselves, see the Call Centre and Customer Service Training course.
BOTI is a QCTO-accredited Skills Development Provider and a Level 1 B-BBEE contributor, and works with teams at organisations including Sasol, Glencore and the City of Johannesburg.
Frequently asked questions
What is a QA scorecard?
A QA scorecard is a form that reviewers use to rate customer interactions such as calls, emails and chats against agreed criteria, each with a weight. The weighted result gives every interaction a comparable score that shows team leaders where to coach and where training is needed.
What should a customer service QA scorecard include?
At minimum it should cover opening and verification, communication and tone, problem resolution, process and compliance, and closing. Each category needs observable criteria and a clear weighting, with auto-fail flags reserved for non-negotiables such as rudeness or a compliance breach.
How does partial scoring work on a QA scorecard?
On a yes/no/partial scale, partial means the agent did part of the behaviour or did it late, and it usually earns half the item’s points. Write one example of partial for each criterion into the guidance notes, and remove items that did not apply rather than scoring them zero.
How often should we calibrate QA scoring?
Most teams calibrate fortnightly or monthly. Reviewers independently score the same interaction, compare line by line, and agree a shared rule wherever they differ by more than a point. Regular calibration keeps scores trustworthy and the scorecard clear.
How do you coach agents using QA scorecard results?
Let the agent self-assess first, focus on one or two high-impact behaviours, use the actual recording rather than just the number, set an observable goal and log progress. When the same gap shows up across the team, treat it as a training need rather than a coaching one.
For managers and team leaders who run QA, the Call Centre Management Course helps you set service standards, measure performance and lead your agents from the results. Training runs in-house at your premises, at BOTI’s venue in Sunningdale, Johannesburg, or live online. This course is not accredited. It is delivered as a professional skills programme, and delegates receive a BOTI certificate of attendance. Call BOTI on 011 882 8853 or email [email protected], or request a quote.