Every sales manager can name their top rep but very few can explain what that rep does on a call that the others do not. That gap is where coaching stalls.
A manager who knows Sarah is the best closer but cannot describe the mechanics of her calls ends up coaching everyone toward “be more like Sarah,” which is a personality instruction rather than a behavior anyone can practice.
We built the Call Analytics Index to close that gap with numbers. This first study scored 6,209 real sales conversations across 12 dimensions on Insight7’s call evaluation platform. Of those, 427 rated Excellent. That is 6.9%.
What follows publishes every figure from the study, along with the sales rep performance metrics you can set as targets this week. The behaviors that separated the top tier are specific, measurable, and teachable.
How the conversations were scored
The method matters before the numbers mean anything.
Each of the 6,209 conversations was evaluated on Insight7 against 12 scoring dimensions, then assigned to a performance tier based on its composite score. Every score reflects a real recorded conversation rather than a survey response, a self-assessment, or a modeled estimate.
The 12 dimensions scored:
- Problem Resolution
- Goal Achievement
- Empathy & Rapport
- Turn Balance
- Actionable Outcomes
- Communication Clarity
- Information Density
- Complaints Resolution
- Negotiation
- Enthusiasm Score
- Overall Effectiveness
- Participant Satisfaction
Six of those need definitions, because the terms carry specific meanings in this dataset:
- Turn Balance Score: how evenly speakers take turns in a conversation.
- Actionable Outcomes: how often a conversation ends with clear next steps, decisions, or follow-ups.
- Goal Achievement: whether the core objective of the call was met, such as an issue resolved, a decision made, or feedback collected.
- Empathy & Rapport: how well the speaker connects emotionally and builds trust with the other party.
- Problem Resolution: how effectively issues were identified, addressed, and resolved during the conversation.
- Information Density: the quantity of meaningful information shared.
Naming these dimensions is what makes the findings usable. A manager who tells a rep to “build better rapport” has given feedback nobody can act on.
But saying the rep scored 2.8 on empathy and rapport, and pointing to the moment in the transcript where a customer concern went unacknowledged = giving that rep something to fix.
How 6,209 conversations sorted into tiers
The 6,209 conversations sorted into four tiers:
| Tier | Conversations | Share |
|---|---|---|
| Excellent | 427 | 6.9% |
| Good | 1,238 | 19.9% |
| Average | 3,979 | 64.1% |
| Poor | 565 | 9.1% |
Two-thirds of every conversation scored landed in the Average band. For a sales leader running 40 reps, that distribution means roughly 26 of them are producing conversations that are competent and unremarkable, while the coaching calendar treats them as one undifferentiated group.
The 565 Poor conversations are the ones a QA sample is most likely to miss.
At a 2% review rate, a team scoring 6,209 conversations a quarter would review around 124 of them, and fewer than 12 of those would statistically come from the Poor tier. The weakest performance in the operation is also the least likely to be seen.
What separated the top 6.9%
We compared the 427 Excellent conversations against every other tier. Four behavioral patterns showed up consistently, and each one can be measured on a call and coached the following week.
1. They asked more questions and asked them faster
| Metric | Excellent | Average | Delta |
|---|---|---|---|
| Questions per call | 14.3 | 10.4 | +37% |
| Question density (per minute) | 1.2 | 0.6 | 2x |
Top performers in the Insight7 dataset asked 14.3 questions per call against 10.4 for average performers, a 37% difference.
The density figure is the sharper one. At 1.2 questions per minute against 0.6, top reps asked twice as often within the same span of time, which means their questions were distributed through the call rather than clustered into an opening discovery block.
That distribution pattern is what makes the behavior coachable. A rep who front-loads ten questions and then presents for twenty minutes has run a qualification script. A rep who keeps asking at a steady rate is checking their understanding as the conversation develops, and that is where the timing advantage comes from.
The Excellent tier asked key qualifiers before offering solutions rather than after.
For an enablement lead at a 60-rep SaaS company, this is the easiest of the four to instrument. Question count per call is a countable event. Set a floor, report it weekly, and the number moves.
2. They shared the floor
| Metric | Excellent | Poor |
|---|---|---|
| Speaker dominance ratio | 1:1 (almost equal) | 1:8 (rep dominates) |
| Turn Balance Score | 0.4 | 1.2 |
The standard coaching instruction to take control of the conversation runs against what the data shows. Excellent conversations held an almost equal speaker dominance ratio. Poor conversations ran 1:8 in the rep’s favor.
A 1:8 ratio on a 30-minute call leaves the customer with under four minutes of speaking time. Objections, budget signals, and competing priorities all live in those four minutes, and a rep talking through them collects none of it.
Balanced talk time reduces defensiveness and gives the rep the material to personalize a recommendation, which is why the top tier listened early and led later.
Turn Balance Score is the cleaner metric to coach against, because it produces a single threshold. Below 1.2 is a conversation. Above it is a presentation the customer is waiting out.
3. They closed with a defined next step
| Metric | Excellent | Average | Delta |
|---|---|---|---|
| Actionable Outcomes | 4.8 | 3.3 | +45% |
| Information Density | +27% | baseline | 27% |
High-performing reps in the dataset scored 4.8 on Actionable Outcomes against 3.3 for average performers, a 45% gap. Excellent conversations ended with the next step named out loud rather than implied.
The Information Density figure sits alongside it for a reason. Top conversations carried 27% more meaningful information in the same span, so the clarity was not coming from talking longer.
A call that ends with “I’ll send the security documentation by Thursday and we’ll reconvene Monday with your CTO” has done something a call ending with “I’ll get back to you” missed.
For a sales manager reviewing pipeline, this behavior shows up as deal velocity. Every conversation without a defined next step becomes a follow-up sequence, and follow-up sequences are where deals go quiet.
4. They scored higher on empathy than on anything else
| Metric | Excellent | Poor | Delta |
|---|---|---|---|
| Empathy & Rapport | 4.7 | 2.8 | +68% |
| Enthusiasm Score | +35% | baseline | +35% |
| Exclamation count | 2.4x | baseline | 2.4x |
Empathy and rapport produced the widest single behavioral gap in the study at 68%, wider than question count, wider than talk balance, wider than outcome clarity.
That tends to surprise sales leaders who treat empathy as a personality trait rather than a scorable behavior. In this dataset it was scored on observable things: whether the rep acknowledged a concern before moving to a solution, or tone matched the customer’s, whether energy carried through the call.
Enthusiasm ran 35% higher in the Excellent tier and exclamation count 2.4x higher, both measurable properties of a transcript.
Emotion drives recall and response, which makes it the delivery mechanism for everything else the rep says. A technically correct recommendation delivered flatly gets forgotten by Thursday.
Three metrics predicted the tier better than the other nine
Ranking every dimension by how strongly it separated top reps from average reps produced a clear order:
| Dimension | Separation |
|---|---|
| Actionable Outcomes | 72.0% |
| Problem Resolution | 56.0% |
| Goal Achievement | 53.8% |
Actionable Outcomes at 72.0% is the strongest single differentiator across all 12 dimensions. For a sales leader deciding where to spend limited coaching hours, that ranking answers the question directly.
Coaching a rep to end every conversation with a named next step moves more than coaching tone, product knowledge, or objection scripts.
Why this stays invisible without full call coverage
None of these patterns appear in a QA sample.
McKinsey’s analysis of contact center quality assurance found that manual assessment is often limited to less than 5 percent of total conversations, with human bias affecting the accuracy of what does get reviewed. McKinsey puts manual scoring accuracy at 70 to 80 percent.
Applied to a 40-rep team, the coverage problem becomes concrete. At 5% review, a manager sees two or three conversations per rep per quarter. A 37% difference in question count and a 0.8-point difference in Turn Balance Score are not detectable in a sample that size.
They emerge across hundreds of conversations per rep, a volume no human reviewer reaches.
The consequence lands in coaching quality. A manager coaching from three sampled calls is coaching from the calls they happened to hear, and those calls may not represent how that rep sells.
Wider pressure makes this more expensive. The Bridge Group’s 2026 AE research, based on responses from 158 B2B companies, found that 48% of reps achieved annual quota in 2026, down from 51% in 2024, with more companies falling into the 0 to 30% attainment band.
When fewer than half of reps hit quota, the difference between a rep asking 14 questions and one asking 10 stops being a coaching nicety and becomes the variable separating a team that makes the number from one that does not.
How to make the top tier repeatable
The three steps below turn the findings above into a coaching program.
Step 1: Benchmark what great looks like in your own team’s voice
The figures in this study come from 6,209 conversations across multiple industries. Your team’s version of Excellent will differ, and the benchmark that matters is the one drawn from your own top performers.
Score your reps against four things first: speaker balance, question density, actionability, and empathy signals.
Then set individual micro goals from each rep’s own scorecard. “Next week, aim for 12 or more meaningful questions” is a target a rep can hit. “Improve your discovery” is not.
This is how average becomes excellent, through data rather than pressure.
Step 2: Build a clip library from real calls
Generic training content teaches reps what good sounds like somewhere else. A library of your own recordings teaches them what good sounds like on your product, against your competitors, with your buyers.
Pull calls that demonstrate three things:
- Strong question pacing, where the rep asks steadily rather than in a block
- Balanced speaker turns, with the customer talking close to half the time
- Next steps landed clearly at the close
Those clips become the training base. A new rep in week two can hear the behavior instead of reading a description of it.
Step 3: Make the metrics visible and coach to them
Four numbers, tracked weekly, tied to deal movement:
| Metric | Target | Why it matters |
|---|---|---|
| Average questions per call | Above 12 | Purposeful questions guide the call rather than filling it |
| Turn Balance Score | Below 1.2 | Above this threshold, one person is presenting rather than conversing |
| Actionability | Above 4.5 | Every call ends with a defined next step, no “I’ll get back to you” |
| Empathy | Above 4.5 | People respond to feeling heard, and it scored the widest gap in the study |
Tie each one to deal movement or resolution quality rather than tracking them as standalone activity metrics. A rep who games question count by asking twelve closed questions in a row will show up in the Turn Balance number.
What this changes about coaching
Sales teams coach from intuition because intuition is what has been available. Managers promote on feel, review the calls they happen to catch, and build training around the reps who seem to be doing well.
The 6,209 conversations in this study point somewhere more specific. The top 6.9% asked 37% more questions, held a 1:1 speaker ratio, drove 45% more actionable outcomes, and scored 68% higher on empathy. Every one of those is a behavior a rep can practice on Monday and a manager can verify on Friday.
Insight7’s AI call scoring evaluates every conversation against criteria your team defines, with each score linked to the moment in the transcript that produced it, so a rep can see what a 2.8 on empathy referred to. Where a score reveals a gap, AI coaching turns it into a practice scenario built on that specific behavior.
To see how your own numbers compare against the benchmarks above, the free call evaluation tools will score a handful of your recordings without a signup.
Cite this research: Insight7. The DNA of High-Performing Reps: Call Analytics Index. 2026. https://insight7.io/sales-rep-performance-metrics/
Media and analyst enquiries: hello@insight7.io
Methodology in brief: 6,209 real sales conversations scored across 12 dimensions on Insight7’s call evaluation platform, then grouped into four performance tiers by composite score. Every figure reflects scored behavioral data from recorded conversations.
Download the full report | More from the Call Analytics Index
Frequently asked questions
How many questions should a sales rep ask on a call?
In Insight7’s study of 6,209 conversations, top-performing reps averaged 14.3 questions per call against 10.4 for average performers. The report’s coaching benchmark is above 12 questions per call, with attention to density rather than total count, since top reps asked 1.2 questions per minute against 0.6.
What is a good talk-to-listen ratio on a sales call?
Excellent conversations in the study held an almost equal 1:1 speaker dominance ratio. Poor conversations ran 1:8 in the rep’s favor. Measured as Turn Balance Score, top performers scored 0.4 against 1.2 for poor performers, and the coaching threshold is to stay below 1.2.
Which sales rep performance metrics predict better call outcomes?
Ranked by how strongly they separated top from average performers: Actionable Outcomes at 72.0%, Problem Resolution at 56.0%, and Goal Achievement at 53.8%. Actionable Outcomes was the strongest single differentiator across all 12 dimensions scored.
Can empathy be measured on a sales call?
Yes. Empathy and rapport produced the widest behavioral gap in the study at 68%, scoring 4.7 for Excellent conversations against 2.8 for Poor. It was scored on observable behaviors including whether concerns were acknowledged before solutions were offered, tone matching, and sustained energy, alongside related measures like enthusiasm score and exclamation count.
