Gong analyzed over a million recorded B2B sales calls and found one of the clearest behavioral separators between top performers and everyone else: how much time the rep spends talking. Top-performing reps talk roughly 43% of the time. Average reps talk closer to 65%. The prospect fills the remaining time in both cases. That 22-point gap in listening is not a personality trait. It is a coachable skill, and the reps who close it close more deals.
The ratio matters not because listening is virtuous but because it is diagnostic. A rep who talks 65% of the time is almost always pitching before they fully understand the problem. They are broadcasting features to a prospect who has not yet confirmed those features connect to anything painful. The prospect can feel the mismatch. It triggers the same low-grade defensiveness that every one of us feels when we realize we are being sold to rather than helped.
The 43/57 benchmark and where it comes from
Gong's conversation intelligence platform records and transcribes sales calls at scale. Their data science team has analyzed talk-time ratios across industries, deal sizes, and rep tenure levels. The 43/57 split is their benchmark for closed-won deals: on calls that end in a signed contract or a clear next step toward one, the rep talks 43% of the time and the prospect talks 57%.
The research also found a hard ceiling: calls where the rep talks more than 65% of the time close at significantly lower rates across all deal types. The relationship is not linear. There is a range between roughly 35% and 50% rep talk time where close rates are roughly flat. Below 35%, the rep is probably not adding enough structure to the conversation. Above 50%, close rates begin to decline, and the decline accelerates past 65%.
Gong's finding is not isolated. Jim Keenan and Mike Weinberg both report similar patterns from sales call audits. Chris Voss, who built his negotiation framework on FBI hostage negotiation principles, makes the same point from a different angle: the person asking questions controls the conversation, not the person filling it with words. The data from multiple directions points the same way.
What an inverted ratio actually signals
When a rep's talk time crosses 60%, it almost always means one of three things is happening in the call. The rep is pitching before they have confirmed what the prospect actually cares about. The rep is uncomfortable with silence and filling it with more product information. Or the rep ran out of prepared questions and defaulted to explaining features to keep the conversation moving.
All three produce the same outcome from the prospect's perspective: a call that felt like a presentation, not a conversation. A presentation can be interesting. It rarely produces urgency or commitment.
If your average talk time across discovery calls is above 55%, you are likely pitching before you have gathered enough information to pitch relevantly. The problem is not the pitch. It is the sequencing. Features land when they connect to a problem the prospect just named in their own words. They bounce when the rep names them first.
There is also a second-order effect on objections. Reps who overtalk in discovery tend to generate more objections in the close, not fewer. The reason is that they have not surfaced the prospect's real concerns during the discovery phase, so those concerns arrive later in the cycle when they are harder to address. Listening more in discovery is a form of objection prevention.
Why reps overtalk
Understanding the cause is necessary before working on the fix. Overtalking on sales calls almost always traces to one of four root causes.
How to self-diagnose using call recordings
You cannot fix a ratio you have not measured. The self-diagnosis process takes about 30 minutes per call and produces a clear picture of where the overtalk is happening and why.
Run this audit across five consecutive calls and you will see a pattern. The overtalk almost always concentrates in the same moments: the call opening, the demo segment, and the first 60 seconds after an objection. Those are the three places to target.
Practical drills to improve your listen ratio
Talk-time ratios improve fastest when you target specific behaviors rather than trying to "listen more" in general. Each drill below addresses one of the root causes above.
Before every discovery call, write down eight to twelve open questions you want answered by the end of the call. Prioritize them: the first three should cover the problem, the next three the business impact, and the last few the decision process and stakeholders. Enter the call with that list visible.
The goal is not to run through all twelve sequentially. It is to have a fallback when the conversation stalls. Reps with a question list never run out of things to ask. Reps without one run out within the first five minutes and start pitching.
Example opening question: "Before I share anything about what we do, I want to make sure I understand your situation. What made you agree to this call today, specifically?"
After asking any substantive question, count to five internally before saying anything. If the prospect starts talking before you reach five, stay quiet until they finish. Then wait another two to three seconds before responding.
This feels unnatural at first because silence on a call is genuinely uncomfortable for the rep. But that discomfort is almost entirely the rep's problem. Prospects do not experience a three-second pause after a hard question as awkward. They experience it as being given space to think. That space is frequently where the most useful information comes out, because the prospect was about to self-edit something and the silence gave them permission to say it.
After asking: "What would have to be true for this to be a top-three priority for your team this quarter?" wait for the full answer. Do not rescue the silence. Do not rephrase the question. Just wait.
Set a personal rule: you will not speak for more than 90 seconds without asking a question. In practice this means interrupting yourself during demos and walkthroughs. After every feature you explain, ask a calibration question before moving to the next one.
The most natural version of this is a relevance check: after showing something, ask whether it connects to the specific problem the prospect named earlier. This does two things. It confirms that what you just showed was actually relevant. And it keeps the prospect engaged and talking rather than processing passively while you walk through a checklist.
After showing a feature: "You mentioned earlier that your team loses about two hours a week on this manually. Does this address that specifically, or does it solve a different part of the workflow?"
Before transitioning into any product explanation, state back the problem in the prospect's words. This serves as a confirmation check and a bridge. It signals to the prospect that you were listening, which increases receptiveness to what comes next. It also forces the rep to confirm whether they actually understood the problem before they start explaining a solution.
If you cannot reflect the problem back in the prospect's own language, you do not yet have enough information to pitch. Ask one more question instead of bridging to the product.
"Based on what you described, the core issue is that your reps spend the first 10 minutes of every call re-explaining context the prospect has already shared elsewhere, which kills the opening momentum. Is that a fair summary?" Then, only after confirmation: "Let me show you how we handle exactly that."
What call recordings show you that you cannot see in real time
One of the most consistent findings in call recording data is that reps dramatically underestimate how much they talk. In self-assessments, most reps estimate their talk time at 50 to 55%. When they listen to recordings, the actual number is often 65 to 75%. The gap between perceived and actual behavior is where coaching starts.
Call recordings surface four things that are invisible in the moment: the exact length of your monologues, how long you wait after asking a question, whether your pitch language mirrors the prospect's language or ignores it, and how often you ask questions in the demo versus during "discovery." Many reps who believe they run strong discovery sessions have essentially moved the pitch to the discovery slot. The questions they ask are not exploratory; they are setup questions for features they were planning to show anyway.
If you want to understand your baseline before working on it, pull five calls and run the audit above. The pattern across five calls is more reliable than any single data point, and it will show you exactly which of the four overtalk drivers is most active in your specific call style.
For sales teams that want to model how coaching on talk-time ratios affects pipeline conversion and close rates across a rep cohort, Numi's simulation layer lets you run those scenarios against a synthetic ICP before you invest in a training program or process change.