Turn the script into observable criteria.
A script says what an agent should do, and a scorecard says how a reviewer can tell that it happened. This illustrative rubric scores a message taking call out of 100, with every weight open for editing below.
Change a weight, flip a result, or mark a criterion as critical, and the score on the right recalculates.
Points 100 / 100
Score for CALL 028
20 + 15 + 25 + 10 + 10 = 80
- Client business name20 / 20
- Offer to help15 / 15
- Information captured25 / 25
- Callback number confirmed0 / 20
- Closing10 / 10
- Courtesy10 / 10
Sample scorecard. Adjust the weights to your own script.
The rubric as a table.
Five criteria, one observable behavior each, and points that add up to 100. A criterion earns its full points when the behavior is present on the call and zero when it is missing.
| Criterion | Observable behavior | Points |
|---|---|---|
| Greeting | Names the business and offers help | 20 |
| Information | Captures required request details | 25 |
| Callback | Repeats the callback number | 20 |
| Closing | States the agreed next step | 15 |
| Courtesy | Uses clear and respectful language | 20 |
Document the exceptions too.
Every rubric meets a call it did not anticipate, and the review process decides in advance how the reviewer records each one. Those decisions belong next to the rubric, since this sample supplies a scoring method rather than a compliance standard.
- A requirement that does not apply to the call, such as a callback number on a wrong number call.
- Evidence the reviewer cannot confirm from the transcript or the audio.
- A critical failure that sets the score to zero whatever the other points add up to.
Bring your real scorecard.
Bring the rubric the team already scores against, and see how it reads once the transcript and the evidence sit beside each criterion.