How much should a seller actually talk?

Ask any sales trainer how much a rep should talk on a call and you will get a number, delivered with total confidence. Forty-three percent. Or thirty. Or twenty, if they are quoting the "80-20 rule of selling," which has no source whatsoever and appears to have been invented by somebody's blog.

The honest answer is more interesting than the confident one.

What the call data says

The best available evidence comes from conversation-intelligence platforms that record and transcribe sales calls at scale. It is vendor research — those companies sell software to sales teams, and their findings conveniently support buying it — so treat it as directional, not as physics.

With that stated plainly: Gong's analysis of 326,000 sales calls found that in deals that closed, reps talked 57% of the time. In deals that were lost, they talked 62%.

Five percentage points. That is the whole famous finding.

You will see a more dramatic version everywhere — the "golden ratio" of 43:57, usually attached to a million calls. That was Gong's 2016 figure, and their own current page reports 57 and 62. The million-call sample belongs to a different study. If somebody quotes you 43:57 today, they have not read the source they are citing.

On questions, the same lab found that asking roughly 11 to 14 questions on a discovery call correlated with the highest success rate, from an analysis of over 519,000 calls. Past about fourteen, performance fell back toward average — a discovery call is not an interrogation. (The version of this stat that adds "…for a 70% success rate" is fabricated. The 70% appears nowhere in the source.)

There is also a finding we like, quoted precisely because it is easy to misquote: across 67,149 recorded demos, Gong reported that they did not observe a single closed deal that involved more than 76 seconds of uninterrupted pitching. That is a maximum observed, not an optimum. It does not mean "talk for 75 seconds." It means that somewhere past a minute of monologue, you have stopped having a conversation, and the data has never once seen that end well.

What the psychology says

Now leave the vendors and go to the peer-reviewed literature, where the claims are smaller and the evidence is much better.

Huang, Yeomans, Brooks, Minson and Gino (2017), in the Journal of Personality and Social Psychology, ran a set of experiments on question-asking in conversation. People instructed to ask at least nine questions were liked more than people who asked four or fewer. But the mechanism is the part worth reading twice: the effect was carried almost entirely by follow-up questions — the ones that prove you listened to the last answer rather than working through your own list.

In their field study, using speed-dating data, participants asked about ten questions per date, of which four and a half were follow-ups. Higher follow-up rates predicted second dates.

Be careful with what that does and does not show. It measures liking, not revenue. Nobody has run this study on sales calls. The sales numbers come from vendors, the rigorous work comes from psychology, and — we looked hard — nobody has joined the two. Anyone claiming a peer-reviewed link between active listening and win rates is claiming something that does not exist yet.

The finding that should actually change your Tuesday

Here is the one that reorganised how we teach.

The most methodologically defensible study in modern selling is not about talking at all. Matthew Dixon and Ted McKenna's JOLT Effect analysed 2.5 million recorded B2B sales conversations — observed behaviour, not a survey. They found that when a customer hesitates, 73% of reps respond by pushing harder on the cost of inaction — the classic "think about what this is costing you every month you wait."

In 84% of those interactions, doing that made the deal more likely to be lost.

Read that again, because it is the opposite of what almost every sales floor trains. The hesitating buyer is not under-motivated. They are afraid of being wrong. Turning up the fear turns up the paralysis.

And it is not a rare edge case. In their data, 40 to 60% of qualified pipeline is lost to no decision at all, and of those losses, 56% are lost to indecision rather than to a preference for the status quo. Win rates fell from around 30% at medium indecision to about 6% at high indecision.

The academic literature independently corroborates the shape of this. Verbeke, Dietz and Verwaal's meta-analysis (2011) found that the strongest predictors of individual sales performance are selling-related knowledge and adaptiveness — not personality, not likeability, not charm, which perform weakly. The seller who wins is the one who knows enough to make the buyer's decision safe, and who changes course when the room changes.

So: how much should a seller talk?

The percentage is a symptom, not a target. Chasing it produces reps who count their own sentences instead of listening to the answer.

What the evidence actually supports is narrower and harder:

• Ask, and then ask again about the answer. Follow-up questions are the mechanism, not question count.
• Never monologue past a minute. Not because 76 seconds is magic, but because past that point you are no longer in a conversation.
• When they hesitate, stop selling. The instinct to push cost-of-inaction is the single most expensive reflex in the trade, and it is trained into people.
• Adapt. The best-evidenced trait in the entire literature is the ability to change what you are doing mid-call.

None of that is learnable from a slide, and none of it survives being told to somebody once. It has to be practised, out loud, with someone stopping you at the moment you slip — which is precisely what we built Clario to do.