Most discovery advice is useless because it’s directional. “Ask good questions.” “Listen more.” “Get to the pain.” True, and impossible to act on. What separates elite discovery from average discovery turns out to be surprisingly specific — specific enough to measure, because Gong did.

By analyzing hundreds of thousands of recorded sales calls against deal outcomes, they turned discovery from folklore into something closer to a spec sheet. Here’s what the spec says.

The number is 11 to 14

On successful discovery calls, top reps ask somewhere between 11 and 14 questions. Not five. Not twenty-five. There’s a real curve here, and both tails lose.

Fewer than about eleven and you’re staying on the surface — you’ve collected enough to write a generic proposal and not enough to understand the actual problem, the politics, or the stakes. More than about fourteen and the call stops feeling like a conversation and starts feeling like a deposition. The buyer gets defensive, the answers get shorter, and you extract less from more. The skill isn’t asking the most questions. It’s asking the right eleven to fourteen and letting each one breathe.

11–14 questions on a successful discovery call. Top reps also ask ~40% more questions than average reps overall — but concentrated on business impact, not features. Source: Gong, analysis of 519,000+ calls.

Where you ask them matters as much as how many

Average reps front-load. They open the call with a barrage — six questions in the first ten minutes — get their answers, and then spend the rest of the call presenting. It feels efficient. It’s backwards.

Elite reps spread their questions across the entire call. Discovery isn’t a phase that ends when the demo begins; it’s a posture that runs the whole conversation. They ask, the buyer answers, they follow the thread, they go deeper on what they just heard, and a new question emerges naturally three minutes later. The result is a call that feels like a dialogue and a buyer who keeps revealing more because they’re being listened to, not processed.

The talk-listen ratio is a law, not a guideline

On winning calls, top reps talk about 43% of the time and listen about 57%. Average and losing calls invert toward the rep talking more. This is one of the most replicated findings in the entire dataset, and it’s the single easiest thing to fix because it’s purely behavioral.

The mechanism behind it is the most important thing in this whole essay: the goal of a question isn’t the answer — it’s the next question, and eventually the long story. Gong found that getting the buyer to tell a longer story correlates strongly with closing. When a buyer talks for ninety uninterrupted seconds about how a problem actually plays out in their world, three things happen: you learn what no qualification framework would have surfaced, the buyer convinces themselves of the stakes by saying them out loud, and the trust deepens because they feel understood. Short answers are a symptom of bad questions and an interrupting rep. The long story is the prize.

The executive exception that proves the rule

Now the counterintuitive part. Everything above describes discovery with practitioners and managers. Take the same approach into a meeting with a C-level executive and you’ll lose.

Gong’s data shows that with senior executives, successful meetings had around four questions; unsuccessful ones had around eight. The relationship flips. Executives have discovery fatigue. They’ve answered “what keeps you up at night” a hundred times, they’re protective of their time, and a rep who shows up with a long question list reads as someone who didn’t do their homework and is now trying to do it on the executive’s clock.

With executives you earn the meeting by walking in with a point of view, not a questionnaire. You make a sharp, specific claim about their business, you’re visibly well-prepared, and you ask a small number of high-altitude questions that only matter because you already understand the basics. The four questions you do ask carry more weight precisely because you didn’t waste their time with the other ten. This is why executive selling is its own move — the discovery rules don’t transfer.

Where AI fits — and where it doesn’t

This is a near-perfect illustration of floor and ceiling. Before the call, AI now does the entire floor of discovery prep: it researches the account, drafts a tailored question bank, surfaces likely pain points from the industry and the persona, and during the call it transcribes every word so you never look down to take a note. There’s no excuse anymore for showing up unprepared. That part is solved.

But AI cannot hear the half-second of hesitation before a buyer answers “and how’s that working?” — the tell that you’ve found the real wound. It can’t decide, in the moment, to abandon question nine because the answer to question eight cracked something open. It can’t read the room well enough to know that with this executive, you should put the list away entirely. The questions are the floor. Reading what the answers actually mean is the ceiling.

So let the model build your eleven-to-fourteen. Then run the call like the number doesn’t exist — because by then you’re not counting questions, you’re following a human being toward the truth of their problem. That’s the part you get paid for.

The takeaway

Elite discovery has a measurable shape: 11–14 questions, spread across the whole call, a 43/57 talk-listen ratio, and a relentless push toward the long buyer story. With executives, invert it — four sharp questions and a point of view. Let AI prep the questions; your job is to hear what the answers really mean.