How to build a QA scorecard
A QA scorecard is the list of criteria a contact center uses to judge a conversation, each with a weight and a scoring rule. A good one is short, observable and versioned. This guide covers choosing criteria, setting weights, writing definitions two reviewers will read the same way, and changing the scorecard without breaking your history.
Step by step
Six steps to a scorecard that holds up
Start from outcomes
List what a good conversation achieves for the person contacting you and for the business: the issue resolved, the policy followed, the next step clear. Criteria should serve those outcomes, not habit.
Pick observable criteria
Keep criteria a reviewer can point to in the conversation, such as "confirmed identity before discussing the account". Avoid ones that need a guess about intent, such as "was genuinely engaged". Six to twelve is usually enough.
Weight what matters
Give each criterion a weight so the total reflects priorities. Compliance and resolution usually carry more than greeting and closing. Weights that are all equal usually mean the priorities were never discussed.
Write a definition and an example
For each criterion, write one sentence on what earns full marks and one example of a miss. This is what makes two reviewers, or a reviewer and software, score the same way.
Test on real conversations
Have two or more reviewers score the same set independently. Criteria where they disagree need rewording before they are used at scale.
Version every change
When criteria or weights change, create a new version rather than editing in place. Scores made under the old version should keep it, or trend lines will compare different rules without saying so.
Wording
Criteria that score consistently
Hard to score
Easier to score
Hard to score
Showed empathy.
Easier to score
Acknowledged the problem in their own words before moving to a fix.
Hard to score
Good product knowledge.
Easier to score
Gave an answer consistent with the current policy, without escalating.
Hard to score
Professional closing.
Easier to score
Confirmed the next step and whether anything else was needed.
In Merivex
How a scorecard works in Merivex
Your criteria and weights
Merivex scores each conversation against the weighted rubric you define, with a rationale and a cited transcript turn per criterion.
A default to start from
A default "QA framework" template exists at the start, which you can adjust to your own criteria.
One active rubric per workspace
A workspace scores against one active rubric at a time. Plan around that if different programs use different scorecards.
Versioned
Changing the rubric creates a new version; earlier scores keep the version they were made under.
Questions
QA scorecards, answered
What is a QA scorecard?
A QA scorecard is the set of criteria a contact center uses to evaluate a conversation, each with a weight and a scoring rule, adding up to an overall quality score.
How many criteria should a QA scorecard have?
Enough to cover the outcomes you care about and few enough that each one is scored carefully. Six to twelve is a common range. More than that usually means overlapping criteria.
How should QA scorecard criteria be weighted?
By impact. Criteria tied to compliance and resolution usually carry more weight than courtesy items. Review the weights when priorities change, and version the scorecard when you do.
Can Merivex use a different scorecard for each client or program?
Not at the same time in one workspace: a workspace has one active, versioned rubric. If your programs use different scorecards, talk to us before you start so the setup matches how you report.
Keep reading
Bring your scorecard
Enter your criteria and weights, score a set of conversations, and see where the wording holds up.