The scorecard, with its weightings.
This is the instrument every contact is graded against. It is published in full because a quality process you cannot inspect is a quality claim, not a quality process.
Six things, weighted.
Weightings add to 100. Two criteria are marked critical: failing either one zeroes the contact regardless of the rest, because some mistakes are not survivable by a good score elsewhere.
| Criterion | Weight | What earns full marks |
|---|---|---|
| Accuracy Critical | 25 | The information given is correct and current. A wrong answer zeroes the contact. |
| Completeness | 20 | Everything the customer asked was answered, plus the obvious next question they did not ask. |
| Process adherence Critical | 20 | Verification, authorisation and escalation steps were followed. Skipping a verification step zeroes the contact. |
| Tone & clarity | 15 | Reads as the client's brand. Plain language, no jargon, no defensiveness under pressure. |
| Efficiency | 10 | No unnecessary transfers, holds or repeated questions. Handle time in a sensible band for the contact type. |
| System hygiene | 10 | Correctly tagged, notes usable by the next person, records updated. |
What each score actually triggers.
A band with no consequence attached is decoration. Each of these sets off a specific, dated action.
| Score | Band | What it triggers |
|---|---|---|
| 95–100 | Exemplary | Added to the calibration set as a reference example. Considered for tier-2 or mentoring. |
| 85–94 | Meets standard | Normal sampling continues. Weakest criterion noted in the weekly one-to-one. |
| 70–84 | Below standard | Named coaching action within 48 hours, re-review inside 7 days, sampling doubled. |
| Under 70 | Escalation | Team lead reviews within 24 hours. Root cause classified: training, documentation, system or judgement. |
The difference a scorecard makes.
Not to the number — to what happens after the number.
- A monthly average appears in a deck
- Nobody can see which contacts were sampled
- Scores drift because nothing is calibrated
- Coaching is generic: "improve your tone"
- The same errors recur for months
- Disagreements become opinion against opinion
- Weekly scores per agent, per criterion
- You can open any sampled contact yourself
- Both sides score the same contacts to calibrate
- Coaching names the criterion and the fix
- Repeat causes change the SOP, not just the agent
- The scorecard settles it — you approved it
Stopping the standard from drifting.
Two teams scoring separately will diverge within a couple of months. The provider's quality score stays healthy while the client's perception of quality falls, and nobody can explain the gap.
- Fortnightly at first. Your reviewer and ours score the same five contacts blind.
- Differences are reconciled criterion by criterion, not argued in aggregate.
- The scorecard is amended when the disagreement turns out to be a definition problem rather than a scoring one.
- Quarterly once stable, and immediately after any change to policy, product or team size.
Sampling volume
Established agents: a defined number of contacts per agent per week, spread across channels and across the week — not only the contacts that went wrong.
New starters: sampled at several times that rate for their first four weeks, with the threshold rising as they ramp.
Exact sample sizes are set per engagement against volume and risk.
Bring us your definition of good.
If you already have a quality standard, we will work to it. If you do not, this scorecard is a starting point we will adapt with you during Design.