Score enough sales and service calls and a pattern turns up that you would probably not have picked in advance. Tone of voice is the highest-scoring skill on the phone. Genuinely working out what the customer needs, the discovery, sits at the bottom, roughly a third lower. We have seen this across more than a million scored behaviours in CallCoach, Icana's own call-analysis product.
Your reps sound warm. That part is handled. Where they thin out is the digging.
Warmth is the one skill every coaching loop you already run keeps reinforcing. You can hear it in any call you sample. A team leader picks it up in passing. Customers mention it back to you. So warmth gets rehearsed from a dozen directions every day, and it climbs to the top of the rubric.
Discovery gets none of that. A rep who asks 2 shallow questions and jumps to the pitch still sounds fine on the call. Nothing in the audio flags it. The cost turns up 2 weeks later as a stalled deal, and nobody traces it back to minute 3 of the first conversation.
What moves once you measure every call
Put a number on every call and the picture reorders itself.
| Skill | Agents with a meaningful lift within about 3 months |
|---|---|
| Needs analysis / discovery | roughly 1 in 3 |
| Closing | roughly 1 in 3 |
| Tone of voice | roughly 1 in 7 |
A meaningful lift means the agent's own 4-week average sits at least 5 points higher, on a 100-point scale, than their own first fortnight. Every agent is measured against their own starting point, never against a rubric or a colleague, since different organisations score to different rubrics. Behind the table sit hundreds of agents, hundreds of thousands of scored behaviours and multiple organisations. Wherever a lift happens, it runs on a similar clock: the median is about 8 weeks, whatever the skill.
Agents with too few scored calls are left out, to avoid survivorship bias, and the figures show correlation, not cause.
Part of the gap is arithmetic. Tone starts far higher, so it has less room to rise. That is the point. The weakest skills carry the headroom, and once agents see feedback on every call, that is where the improvement shows up.
Why coaching keeps missing it
Coaching follows what is audible. Without call-level measurement, your hours go to delivery, warmth and pace, the things you can catch while walking the floor. Discovery fails silently. A single call with thin questioning still sounds like a good call, so nobody flags it, and the habit sets.
So your coaching hours land on the skill already near its ceiling, where they buy the least movement. The quiet skills, discovery and closing, are the ones a third of agents lift within a quarter once someone measures them.
Two questions for whoever runs your quality program
Ask them which of the fundamentals each agent improved last quarter. Then ask how many coaching conversations were about tone. If the second number is large and the first draws a blank, your coaching budget is going where there is no headroom left to buy.
None of this works without the measurement first. You cannot coach a skill you only hear in a 2% sample. That is the same coverage problem we wrote about in the QA coverage gap: the calls you never score are the ones quietly setting your ceiling.