Quality Assurance Score
Definition
Quality Assurance Score
A quality assurance score is the rating a reviewer gives one piece of work when scoring it against a fixed checklist. It is quality judged against a written standard, rather than against whatever a reviewer happened to notice on any given day.
The score covers any reviewable work item. A call, a chat, a processed claim, and a coded medical chart can all carry one.
Two things make the score trustworthy. The scorecard has to be written down, and reviewers have to be calibrated against each other regularly.
Without calibration the number measures the reviewer — not the work. That is the failure mode nearly every quality programme hits within a year.
Key takeaways
- A quality assurance score rates one work item against a fixed, written scorecard.
- Calibration between reviewers is what makes scores comparable across a team.
- Auto-fail items should be separated from weighted points, not blended into them.
- Scores tied directly to pay reliably produce score inflation rather than better work.
How it works
Build a scorecard listing the behaviours or checks that matter, assign a weight to each, then have a reviewer mark one work item against every line. The weighted total becomes that item’s score, usually expressed as a percentage.
Most scorecards separate two categories. Weighted points cover craft, while auto-fail items cover things that make the work unacceptable regardless of everything else.
| Scorecard section | Typical weight | Examples |
|---|---|---|
| Accuracy | 30% to 40% | Correct data, correct outcome |
| Process compliance | 20% to 30% | Steps followed, notes recorded |
| Communication | 20% to 30% | Clarity, tone, confirmation |
| Auto-fail | Pass or fail | Data breach, regulatory omission |
Calibration keeps the number honest. Running the same work item past every reviewer each month exposes drift before it contaminates a full quarter of reporting.
The American Society for Quality distinguishes the two halves of the discipline, defining assurance as the planned activities that provide confidence requirements will be met, and control as the operational techniques used to fulfil them.
Structured measurement frameworks make the same point — the Baldrige Performance Excellence Program at the National Institute of Standards and Technology treats consistent measurement as inseparable from actual performance improvement.
Examples
Scorecards look different across support, finance, and clinical work, but the calibration problem is identical everywhere. Four cases show what changes and what stays the same.
A Manila support desk. Five contacts per agent are scored monthly against a 20-line scorecard. Team averages sat within four points of each other only after calibration sessions began.
A claims quality function. Auto-fail items cover regulatory disclosures, so a single omission scores zero regardless of everything else in the file. Blending it into the weighted total had previously hidden compliance risk.
A medical coding team. Accuracy carries 60% of the weight because a coding error has financial and clinical consequences. Communication lines were dropped entirely as irrelevant.
A financial services back office. Scores were tied to a monthly bonus and rose steadily — while customer complaints climbed. Decoupling the two restored the number’s usefulness within a quarter.
Related terms
A quality assurance score sits inside a wider quality function with its own roles, tools, and outcome measures. The terms below cover who scores, what they use, and what the score is checked against.
- Quality Assurance: the discipline the score belongs to.
- Quality Analyst: the role performing the reviews.
- QA Evaluation Sheets: the scorecards themselves.
- Call Quality Monitoring: the contact centre application of the score.
- Call Auditing: the sampling activity that feeds the score.
- Quality Assurance Manager: the owner of calibration and scorecard design.
- Customer Satisfaction Rating (CSAT): the outside view the score should track against.
FAQ
What is a good quality assurance score?
Most operations target 85% to 95%, with the exact bar set by how much a defect costs. A team scoring 99% is usually running a scorecard that is too easy.
How many items should be scored per person?
Four to eight per month is common, enough to spot patterns without consuming the reviewer’s week. Higher-risk work justifies larger samples.
What is calibration and why does it matter?
Calibration means reviewers score the same item and compare results. Without it, scores reflect reviewer preference rather than actual work quality.
Should scores be linked to pay?
Linking them directly tends to inflate scores and damage trust. Using them for coaching and trend analysis keeps the data honest.
How does it differ from a customer satisfaction score?
This score is an internal judgement against a standard, while satisfaction is the customer’s own rating. The two should be tracked side by side.
Who should design the scorecard?
Quality owns the design and operations reviews it. Involving frontline staff in drafting improves both accuracy and acceptance.
Compare quality programmes across delivery locations in the Outsource Accelerator BPO hubs.
Independent
