How do you coach a rep who talks too much on discovery calls?

What over-talking actually signals, and why it costs you pipeline

A rep who dominates a discovery call is almost never doing it out of arrogance. Over-talking is a downstream symptom, and if you coach the symptom without finding the upstream cause, the behavior snaps back inside two weeks — usually right after the next slow pipeline day. Before you say a word to the rep, sort the cause into one of four buckets, because each bucket needs a different intervention and three of the four are not "coaching" at all.

Skill gap. The rep genuinely does not know how to ask layered, open-ended questions or how to hold a silent pause. They fill airtime because they have no next question loaded. Watch the transcript: skill-gap reps ask a question, get a two-sentence answer, and then immediately pivot to a new unrelated question or a feature. There is no second layer. This is the most common and the most fixable — it responds to drills, not conversations.

Will or mindset gap. The rep believes their job is to *convince*, so they pitch the moment a buyer breathes. Often this is a former top closer from a transactional or SMB motion where the "always be selling" reflex actually worked. Moving that rep upmarket into a multi-stakeholder deal breaks the reflex, but they don't know it yet. This one needs a reframe of what discovery is *for*, not more question training — they can already ask questions; they don't believe they should.

Knowledge gap. The rep over-explains the product because they are insecure about their own product knowledge, and they compensate by reciting everything they *do* know before the buyer can ask something they can't answer. You'll hear this as long, defensive product monologues triggered by any technical question. The fix here is product enablement and a certification, not a coaching cadence. Coaching a knowledge gap is one of the most common wasted managerial hours in sales.

System or nerves. Anxiety, a thin pipeline that makes every call feel life-or-death, or a call flow that front-loads a demo. If the rep has three deals in the funnel and needs all three, they will not risk a silence. That's not a rep problem — that's a top-of-funnel problem wearing a coaching costume. Fix the pipeline pressure or the script, and the ratio often corrects on its own.

The commercial cost is real and it compounds downstream. Discovery that runs seller-heavy produces thin, seller-authored qualification: the rep leaves with their own hypothesis about the pain rather than the buyer's words. That thin discovery flows straight into a generic demo, a proposal aimed at the wrong metric, a champion who can't repeat your value internally, and eventually a forecast full of deals nobody can articulate a reason to buy. Gong Labs' analysis of large volumes of recorded sales calls has consistently found that winning discovery calls skew toward the seller listening more than talking, while losing calls have the seller consuming roughly two-thirds or more of the airtime. Whatever exact number your own data produces, the direction is the same, and you can validate it inside your own CRM in an afternoon.

There is a RevOps angle worth naming here too, because it changes who owns the fix. Talk-to-listen ratio is a leading indicator that lives in the conversation-intelligence platform, while conversion is a lagging indicator that lives in the CRM. If nobody has joined those two systems, the manager coaching the rep is flying on anecdote and the VP reviewing the manager is flying on quota. RevOps' job is to make the join exist: call-level behavior data stitched to opportunity outcomes, so you can show the rep not just "you talked 68%" but "the discovery calls where you were under 50% advanced to next step at a materially higher rate than the ones where you weren't." That's the argument that changes behavior in a skeptical veteran rep, and it's an argument only ops can build.

One more diagnostic nuance most managers miss: over-talking on discovery is frequently *stage-specific*. The same rep who bulldozes a first call can be perfectly disciplined in a technical validation call or a negotiation. That's a tell. It means the reflex fires under uncertainty — when the rep doesn't yet know what the deal is about and fills the void with what they do know. If you see the ratio problem confined to first calls and disappearing later in the cycle, you are almost certainly in the nerves or will bucket, not skill.

The coaching sequence, start to finish

Run this in a scheduled 1:1 with a recorded call already queued. Never deliver it as a hallway drive-by, and never in a team meeting — public correction of a talk habit produces a rep who talks less because they're humiliated, which lasts about a week. Use the GROW model (Goal, Reality, Options, Will) so the rep does the thinking and owns the fix.

Open with safety, not a verdict. "I pulled up your discovery call with the Acme account from Tuesday. I want to watch two minutes of it together, and I want *your* read first. Sound good?" You are establishing that this is a review, not a sentencing.

Establish the Goal — let them name it. Ask: "What's the job of a discovery call, in one sentence?" Most reps say "qualify" or "understand the buyer." Good. Then: "So if the buyer should be doing most of the talking, what percentage of the airtime should be theirs?" The rep almost always names a number better than their actual performance. You've now got their own standard on the table, which is far more durable than yours.

Surface Reality with the recording, not your opinion. Play 90 seconds from a stretch where they talk over the buyer. Ask: "What did you notice about who was talking?" Let them answer before you show any data. Then show the number: "The platform clocked you at 68% talk time on this call. The buyer got 32%. What's getting in the way of letting them talk?" That last question is where the diagnosis pays off — their answer tells you which of the four buckets you're in, in their own words.

Explore Options — one, not five. "What's one thing you could do differently in the first ten minutes to flip that ratio?" If they're stuck, offer exactly one technique. Dumping five options is how managers feel helpful and reps feel overwhelmed. A good single option: after asking a question, count to three silently before speaking again. Silence pulls the real answer out, and the three-count is small enough to actually execute under pressure.

Lock Will — a specific, small, measurable commitment. "On your next three discovery calls, what number are you aiming for, and how will we both see it?" Land on something like: under 50% talk time on the next three, reviewed Friday. Notice what you never said in this entire conversation: "stop talking so much." You made the rep diagnose, name the number, and commit to the review. That is coaching. The other thing is telling.

Then run the loop. One conversation does not rewire a habit that took years to build.

Days 1–30, awareness. Weekly call review together. Before you show the platform number, the rep self-scores their own talk ratio out loud. The gap between their guess and reality is the single most powerful teaching moment in this whole program — most over-talkers guess 50% when they're at 70%. Goal for the month: consistently under 55%.

Days 31–60, skill build. Shift the focus from awareness to question quality. Awareness alone plateaus around 55% because the rep is white-knuckling silence without knowing what to ask next. Run the ask-then-shut-up drill in team practice. Target under 48% talk time and at least three open-ended questions before any product mention.

Days 61–90, independence. The rep reviews their own calls and brings *you* the clip they want to discuss. You spot-check rather than lead. Target the 43–47% zone and look for movement in discovery-to-next-step conversion. Fade the scaffolding deliberately — if you're still running the same weekly review at day 120, you've built dependence, not skill.

Drills that actually change the behavior

Skill changes through repetition, not advice. Advice is what managers give when they don't want to run practice. Here are the drills worth the calendar time.

The ask-then-shut-up drill. In a 1:1 or team practice, the rep asks one discovery question and must stay silent for a full five seconds. You play the buyer and give a short, slightly incomplete answer. The rep's only permitted next move is a follow-up question to that answer — never a pitch, never a "that's interesting," never a bridge to a feature. Five rounds. Then switch seats so the rep plays the buyer and feels how a pause actually lands from the other side. Almost every rep reports the same thing: silence that felt eternal as the seller felt like nothing as the buyer. That single reframe does more than three coaching sessions.

Assign it as five minutes of daily solo practice too. The rep records themselves asking a question and holding silence for ten seconds after an imaginary answer. It sounds absurd. It works, because the discomfort is the thing being trained away, and discomfort only fades through exposure.

Recorded-call autopsy with a highlighter. Take one of the rep's real recordings, pull a three-minute transcript, and have the rep highlight every sentence they spoke in one color and every buyer sentence in another. Do not summarize it for them — make them do the marking. The visual is brutal in a way that a percentage isn't. A page that's 80% one color ends an argument no dashboard can.

The three-to-one question ladder. Role-play where the rep must ask three layered questions — situation, then problem, then impact, borrowing the SPIN structure — before they're allowed to describe a single feature. Score them on whether they *earned* the pitch. This drill specifically attacks the will-gap rep, because it makes the pitch a reward for discovery rather than the point of the call.

Mirror-and-pause. The rep repeats the buyer's last two or three words back as a question — "…losing two deals a quarter?" — and then goes silent. It transfers airtime back instantly and costs the rep no cognitive load, which is why it's the best technique to give a nervous rep. It requires no preparation and no next question loaded.

The blind-note drill for the adjacent problem. Have the rep take notes on a call and then, afterward, write down the buyer's stated priority in the buyer's own words. If they can't do it — if the note says "needs better reporting" when the buyer said "my board asks me for churn by segment every month and it takes my analyst three days" — that's proof the over-talking cost them the actual discovery, not just the ratio. This drill connects the behavior to the deal, which is the connection reps care about.

Two adjacent applications worth stealing: the same drill set works nearly unchanged for customer success managers running renewal risk conversations, and for solutions engineers who over-demo. SEs are a particularly under-coached population here — an SE who takes over a discovery call and starts feature-touring undoes the AE's restraint in ninety seconds. If your ratio problem persists on calls with a second internal person present, coach the pair, not the rep.

What to measure, and what it costs to run

Coaching only counts if a leading indicator moves. Track these weekly, in priority order.

Talk-to-listen ratio on discovery calls. The headline metric. Look for a trend from wherever the rep starts down toward the mid-40s. Track the trend line, not any single call — one technical deep-dive with a skeptical engineer will legitimately blow up the ratio, and punishing that teaches the rep to game the number.

Count of open-ended questions per call. Conversation-intelligence platforms tag these automatically. A rising count usually precedes a falling ratio by a week or two, which makes it a useful early signal that the coaching is landing before the headline metric moves.

Longest customer monologue. Strong discovery calls contain at least one uninterrupted buyer stretch in the 30-to-45-second range or longer. If the buyer never talks for a sustained block, the rep is interrupting even when the aggregate ratio looks acceptable. This metric catches the rep who has learned to talk in short bursts to game the percentage.

Patience, or silence after questions. Most platforms report this directly. Rising patience nearly always precedes a falling talk ratio, because it's the mechanical cause.

Discovery-to-next-step conversion. The lagging proof, and the number that justifies the program to leadership. Better listening should lift the percentage of discovery calls that advance to a scheduled next step with a named stakeholder.

Do not coach to quota alone. Quota is a lagging number that moves months later and gets contaminated by territory, seasonality, and luck. The ratio and question count are the behaviors you can change this week.

Timelines. Expect measurable ratio movement in three to four weeks with a weekly review cadence, and habit durability — meaning the ratio holds without you watching — somewhere around the 90-day mark. If you see zero movement after four consecutive weekly reviews with a real recording each time, stop coaching and re-diagnose; you almost certainly picked the wrong bucket. The most common misdiagnosis is treating a knowledge gap as a skill gap, which produces a rep who gets more anxious every week because you keep drilling something that isn't the problem.

Time cost. Budget roughly 45 minutes of manager time per rep per week during the first 30 days: 20 minutes reviewing the recording before the 1:1, 20 minutes in the 1:1, 5 minutes logging the commitment. That drops to about 20 minutes weekly during days 31–60 and roughly 15 minutes biweekly in the independence phase. Across a team of eight reps with two in active coaching, that's a manageable load. Across a team of eight with six in active coaching, you have a hiring or enablement problem, not a coaching problem — no manager sustains six concurrent 90-day loops.

Doing it without a conversation-intelligence platform. Plenty of teams don't have one, and the program still runs. Have the rep self-score every discovery call for one week on a tally sheet: one mark every time they speak for more than 30 seconds straight, one mark for every open-ended question asked. Targets: no more than three long-speech marks per call, at least five open-ended questions. The act of tracking alone cuts over-talking noticeably, because the rep becomes conscious of the behavior in real time rather than in review. If the rep resists self-scoring, spot-check one call a week yourself — listen live or to a plain recording, time it roughly with a stopwatch, and give a single number: "You spoke about 70% of the time. Next call, aim for 55%." The number strips the emotion out of the feedback in a way adjectives never do.

The build-versus-buy question for RevOps. If you're deciding whether to fund conversation intelligence specifically to solve coaching, the honest framing is that the tool doesn't coach — it makes coaching cheap. Its value is manager time saved: without it, a manager spends 20 minutes finding the right 90 seconds of a call; with it, 2 minutes. Multiply that saving by reps times weeks and you have your business case. If your managers aren't already doing weekly call reviews with a stopwatch, buying the platform will not create the habit. Tools amplify a cadence; they don't manufacture one.

Where managers get this wrong

Coaching the symptom instead of the cause. Telling a nervous rep with four deals in the pipeline to "ask more questions" is asking them to take a risk they cannot afford. Fix the pipeline, or accept that the ratio won't move.

Rescuing the rep. Watching the call review and jumping in with "here's what you should have said" instead of letting them find it. It feels like generosity. It builds dependence, and it's the single most common failure mode among managers who were themselves great reps.

One-and-done. A single dramatic 1:1, no follow-through, habit returns in ten days. This is the most common outcome of coaching conversations in general, and the reason the weekly loop exists.

Coaching everyone identically. A skill-gap rep and a will-gap rep need opposite interventions. The skill-gap rep needs drills and would be insulted by a purpose reframe. The will-gap rep can already run the drills and will nod through them while changing nothing. Same speech to both wastes everyone's time and teaches the team that coaching is theater.

No number, no accountability. "Talk less" is unmeasurable and therefore uncoachable. "Under 47% on the next three, reviewed Friday" is a commitment with a review date.

Making the ratio a compliance metric. The fastest way to destroy this program is to put talk-to-listen ratio on a leaderboard or into a comp mechanism. Reps optimize what's measured — they'll ask a question, go silent, let the buyer ramble unproductively, and hit a beautiful 42% on a call that qualified nothing. Ratio is a coaching diagnostic, not a performance metric. Keep it in the 1:1 and out of the dashboard the VP reviews.

Missing the wrong-fit hire. If after a full 90-day loop with real recordings and real drills the rep cannot change, and particularly if they avoid or resist the recordings themselves, that is a performance conversation and possibly a formal plan — not a fifth month of coaching. Managers routinely spend two quarters coaching a rep who told them in week three that they weren't going to change.

The manager who also talks too much. Uncomfortable, but common. If the rep is mirroring the manager's style on joint calls, coaching the rep is futile. Have a peer or a second-line leader model the behavior on a joint call, or coach the manager first. Behavior on a sales floor propagates downhill; nothing you tell the rep survives contact with what they watch you do.

Deciding what to run, and when to stop

Pick the intervention by bucket, not by habit. Managers tend to reach for whichever tool they're most comfortable with — the ex-trainer runs drills for everything, the ex-rep tells war stories for everything — and the mismatch is why so much coaching produces nothing.

If the diagnosis is skill, run drills and the weekly review loop; expect four to six weeks to visible change. If it's will, start with a reframe conversation about what discovery is for and what it costs the deal downstream, and don't run a single drill until the rep agrees on the purpose — drilling an unconvinced rep is compliance training. If it's knowledge, hand it to enablement, set a product certification, and check back in three weeks; the ratio usually corrects itself once the anxiety source is gone. If it's system, the intervention isn't aimed at the rep at all — it's aimed at pipeline coverage or the call script, and it's a RevOps and management fix.

When two buckets are live simultaneously — which happens often, especially skill plus nerves — sequence them. Fix the system cause first, because a rep under pipeline pressure cannot absorb skill coaching. Then run the skill work. Trying both at once produces a rep who is being asked to change three things while worrying about their number, and nothing lands.

Know your stopping conditions in advance and write them down before you start, because in the moment every manager rationalizes one more month. Stop and escalate when: four weekly reviews produce no ratio movement and no change in question count; the rep declines to have calls recorded or reviewed; or the rep improves the ratio while conversion stays flat and the discovery notes stay just as thin — that last one means they learned to be quiet without learning to be curious, which is a different and harder problem.

Related questions

How do you coach a rep who talks too much only on executive calls?

That's a nerves pattern, not a skill gap — the reflex fires with seniority in the room. Run a pre-call plan with three executive-level questions written down, and have a second person on the call to hold the rep accountable to their own plan.

Should the talk-to-listen ratio target differ by deal size or segment?

Yes, modestly. Complex enterprise discovery with multiple stakeholders tolerates a slightly higher seller share because you're managing a room. Transactional SMB calls should skew even further toward listening because the qualification window is shorter.

Can you coach this in a team meeting instead of a 1:1?

Drills, yes — the ask-then-shut-up drill runs well in a group. Diagnosis and the ratio conversation, no. Public correction of a talk habit produces short-lived compliance and lasting resentment.

How does this apply to SDRs on cold calls?

Differently. Cold calls are short and the SDR legitimately carries more airtime early. The useful metric there is whether the prospect ever speaks in a sustained block, not the aggregate ratio.

What if the rep's talk ratio improves but deals still stall?

You fixed the airtime and not the curiosity. Shift coaching from quantity of silence to quality of questions — specifically whether they reach impact-level questions rather than stopping at surface problems.

FAQ

How long does it typically take to see improvement in a rep's talk-to-listen ratio?

Most reps show measurable change within three to four weeks of consistent coaching, but only if they review at least one recorded call weekly and practice the drill regularly. Without that cadence, the old habit generally returns inside two weeks. Durability — the ratio holding without supervision — takes closer to 90 days.

What if the rep gets anxious during silence on calls?

Very common, especially with newer reps. Start with three-to-five-second pauses in a safe role-play environment, then extend toward ten seconds on live calls. The mirror-and-pause technique helps because it gives the rep something to say that costs no preparation, which lowers the anxiety of not having a next question ready.

Can talking a lot ever be right on a discovery call?

Occasionally. A highly technical validation call, or a call where the buyer explicitly asks for an overview, will legitimately run seller-heavy. Judge the trend across calls rather than any single one, and check whether the rep can afterward state the buyer's priority in the buyer's own words — that's the real test, not the percentage.

How do you track this without buying a conversation-intelligence platform?

Self-scoring tally sheets plus one manager-listened call per week works. Mark every stretch where the rep speaks longer than 30 seconds and every open-ended question asked. Rough stopwatch timing on a plain recording is enough to produce a number, and the number is what makes the feedback actionable.

How do you handle a rep who believes they're asking great questions but isn't?

Count them together on a real recording, separating open-ended from closed and leading questions. Reps often mistake repeated "what else?" prompts for discovery. The fix is pre-scripting three to five layered questions before each call and reviewing afterward whether they got asked and what they surfaced.

What if the manager doing the coaching also talks too much?

Then the rep is likely mirroring, and coaching the rep won't hold. Have a peer manager or a second-line leader model the behavior on a joint call, or address the manager's own ratio first. Sales behavior propagates from whoever the rep watches most.

Sources

• Gong Labs: Talk-to-Listen Ratio in Sales Calls
• Gong: Discovery Call Tips Backed by Data
• Harvard Business Review: What Great Listeners Actually Do
• RAIN Group: Sales Discovery Questions
• Winning by Design: The SPICED Sales Methodology
• MindTools: The GROW Model of Coaching and Mentoring
• Chorus by ZoomInfo: Conversation Intelligence
• Harvard Business Review: The Surprising Power of Questions

Related on PULSE

• [How do you phrase a question that helps a rep recognize they are using too much jargon instead of customer language?](/knowledge/cg0896)
• [How much time should a sales manager spend coaching each week?](/knowledge/cg0002)
• [How do you coach a rep who freezes on executive calls?](/knowledge/cg0794)
• [How do you coach a rep who gets ghosted after strong first calls?](/knowledge/cg0790)
• [How do you coach a rep to talk less and listen more on calls?](/knowledge/cg0122)
• [How do you coach a rep to handle the 'it's too expensive' objection?](/knowledge/cg0071)