Agent Coaching That Actually Moves the Numbers

Most call centers believe they coach. What most of them actually do is score calls, publish the scores, and hope. Scoring is measurement. Coaching is the thing that happens afterwards, and it is where almost all the value is.

Score against behaviour, not impressions

A scorecard should be answerable by two different people listening to the same call and arriving at the same result. If a criterion reads “built rapport”, it will not be. If it reads “acknowledged the caller’s stated concern before presenting the offer”, it will.

Write every criterion as an observable behaviour with a yes or no answer. Keep the card short — ten to fifteen items covering the behaviours that actually correlate with conversion in your vertical. A forty-item card produces compliant, unpersuasive agents and exhausted reviewers.

Calibrate the reviewers. Have three people score the same five calls each month and compare. Drift is constant and invisible until you measure it, and an uncalibrated scorecard quietly destroys agents’ trust in the whole process.

Separate systemic faults from individual ones

This is the single highest-leverage change most floors can make. Break scores down by failure category as well as by agent.

If eighteen of twenty agents miss the same criterion, that is not twenty coaching conversations. It is one scripting or training fix. Coaching it individually is slow, expensive, and reads to the floor as blame for something the business got wrong.

A pattern across the floor is a management problem. A pattern in one agent is a coaching problem. Treating the first as the second is the most common way QA programmes lose credibility.

One behaviour at a time

A coaching session that lists six things to improve changes nothing. People cannot hold six changes in their head while also doing their job.

Pick one behaviour. Play the recording where it went wrong and the recording where someone did it well. Role-play it until it sounds natural rather than recited. Agree how you will both know it has changed, and set a date to check. Then stop.

• Weekly, short and specific beats monthly and comprehensive.
• Always use real recordings — theirs and a good example from a peer. Abstract feedback is unactionable.
• Always set a follow-up. Feedback with no check-in is an opinion.

Coach the top of the floor too

Coaching effort almost always concentrates on the bottom quartile, because that is where the visible problem is. But a two-point improvement across your top performers usually produces more revenue than dragging the bottom two agents to average — and your best agents are the ones most likely to leave if they feel ignored.

There is a compounding benefit as well: recordings from your top performers are the raw material for training everyone else. A library of real calls that worked, indexed by situation, is worth more than any training deck.

Retention is a performance metric

Tenured agents convert better than new ones. That single fact means attrition is not an HR line item — it is a conversion problem with a delay on it. Any coaching programme that raises scores while burning people out is losing money it cannot see.

Track score movement and tenure together. If both are healthy, the programme is working. If scores rise while tenure falls, you are pressuring rather than developing.

What good looks like in a quarter

A short, calibrated scorecard. Weekly one-behaviour sessions with real recordings. Systemic faults routed to scripting and training rather than to individuals. A growing library of good calls. Conversion and tenure both moving in the right direction.

None of that requires new software. It requires someone to own it and the discipline to keep it short.

If you want help building it, our quality assurance and recruiting and training engagements do exactly this, and the wider picture is on our call center optimization page. Book a consultation to talk it through.