Scoring a call

Review queue→a call

Taking a call from the queue opens Evaluate call. The call is yours while you have it, so nobody else scores the same one.

Play it, pause it, go back. Playback is logged — who listened to whose call, and when — which is normal and worth knowing rather than a reason to hesitate.

Listen before you score

Most of it, not all of it. Scoring from the first thirty seconds produces a judgement about the greeting, and the thing you were asked to assess is usually in the middle.

One answer per criterion, against the scorecard the owner wrote. You cannot change the criteria or their weights, which is what keeps the standard the same across reviewers.

1. 1Answer each criterion on what you actually heard, not on your impression of the agent.
2. 2Mark a criterion not applicable where the call genuinely gave no opportunity for it — it is then left out rather than counted as failed.
3. 3Flag a critical failure where one happened. It fails the call whatever else went well, and that is deliberate.
4. 4The total is worked out from the weights. You do not set it directly.

The score tells an owner something. The feedback is the only part that changes what the agent does next week, and it is worth more time than the scoring.

General feedback

Visible to the agent. The overall read: what this call was like.

Strengths

Visible to the agent. What they did well, specifically. Not padding — an agent who only ever reads criticism stops reading.

Areas for improvement

Visible to the agent. What to do differently, as a behaviour rather than a trait.

Improvement recommendations

Visible to the agent. The concrete suggestion: try this next time.

Private notes for supervisors

Not visible to the agent. For things a manager needs and the agent should hear in person rather than read.

Write it so it can be acted on

• "At 2:14 you answered the price question before asking what they needed it for. Ask first — it changes which price is the right answer."
• "Good recovery at 4:30 when they said they had been promised something else. You acknowledged it before explaining."
• "You ended without agreeing a next step. Try: so shall I call you Thursday morning?"

• "Needs to improve discovery." — true, and gives the agent nothing to do.
• "Good call overall." — pleasant, and teaches nothing.
• "Be more confident." — a trait, not a behaviour. Nobody knows what to practise.
• "As discussed." — a review the agent reads weeks later with no idea what was discussed.

Private notes, used properly

For a pattern a manager should know about, or something that needs a conversation rather than a written note. Not for the real feedback — an agent who later learns the honest assessment was in a box they could not see has reason to distrust every review they did get.

Completing it sends the score and the visible feedback to the agent, who sees it on their own page. It also feeds the owner’s figures, so a review finished today is in this week’s numbers.

• Reread the feedback once before finishing. It is going to somebody.
• A review that has gone to an agent cannot be quietly edited afterwards.
• Set a coaching follow-up where the feedback deserves one — see Coaching and calibration.

Fifteen to twenty minutes for a call of a few minutes: listening to most of it, scoring it, and writing something useful. If you are consistently doing it in five, the feedback is almost certainly too thin to act on.

If the queue demands more than you have time to do properly, say so. The honest failure is telling the owner the target is too high; the dishonest one is producing twenty skimmed reviews that look like twenty real ones in the figures.