Four annotated sales call scripts, with the load-bearing line marked

Gong’s data team ran the numbers on 90,380 cold calls. Opening with “How have you been?” produced a 10.01% success rate against a 1.5% baseline. One line, over six times the result.

Scripts almost never come marked that way. You get forty lines of dialogue laid out flat, as if the greeting and the close carry the same weight, when a call usually turns on two or three sentences and the rest is connective tissue. So the sales call script examples below are annotated. In each one the load-bearing line is marked, and where a published figure backs it up I’ve named whose figure it is and linked the page.

How to read these sales call script examples

Two conventions. Lines marked DOES THE WORK are the ones I’d rehearse until they’re boring. Lines tagged craft, no number are there because a call needs them, not because anyone has published a win-rate on them, and I’d rather say that than dress up a habit as evidence.

One caution before you copy anything. Neil Rackham, who built SPIN off Huthwaite’s research programme, was asked what actually produced the results. His answer, on Huthwaite’s own site: “It wasn’t the four little questions. It was the whole coaching process, which was also the research process.” He puts the measured lift at “17% higher sales volume” for the first thousand people trained against control groups. Read that as a warning label. A script you never review is a script that decays.

Cold call, first 45 seconds

• “Hi Dana, Sam here from Craqly. How have you been?” DOES THE WORK. 10.01% versus a 1.5% baseline in Gong’s 90,380-call set.
• Stop talking. Let it land. craft, no number
• “The reason for my call is that your team runs demos on Zoom and your reps are writing notes afterwards from memory.” Gong puts “The reason for my call is…” at a 2.1X lift.
• “Is that anywhere close, or have I got the wrong company?” craft, no number

Never open with “Did I catch you at a bad time?” Gong scored that one at 0.9% and reckons it makes you 40% less likely to book the meeting. It reads as polite. It’s actually an invitation to hang up.

Now the part that contradicts the advice you’ve been given. On a cold call you should talk more, not less. Gong’s cold-call research says “Reps take the burden of the conversation in successful cold calls, talking 55% of the time”, that successful cold calls run about 5 minutes 50 seconds, and that the average uninterrupted monologue in a good one is 53 seconds against 25 in a bad one. A stranger hasn’t agreed to be interviewed yet. Earn the questions first.

Discovery call, minutes 4 to 30

Target 11 to 14 questions across the call and roughly 46% talk time. Both figures come from Gong’s stats roundup, which puts average sellers “in the high 60s” on talk ratio. That gap between 46 and the high 60s is most of the difference between a discovery call and a pitch with pauses in it.

• “Before I ask anything, how long have you got? I want to leave you five minutes at the end.” craft, no number
• “Walk me through how a deal moves from first call to signature today.” Situation. Cheap to answer, and it buys you the vocabulary you’ll need later.
• “Where does that break?” Problem. Four words, and it does more than a paragraph would.
• “What does that cost you in a quarter, roughly? I’ll take a bad estimate over none.” DOES THE WORK. This is Huthwaite’s Implication question, which their methodology page describes as one that will “Explore the consequences and effects of problems to help customers understand the urgency and importance of finding solutions”.
• “If that number were half, what would change for you personally?” Need-payoff. Let them say the value out loud so you don’t have to.

The “roughly” and the “I’ll take a bad estimate” do real work in that fourth line. Buyers stall on impact questions because they don’t want to be quoted back a number they invented. Lower the bar and they’ll answer. Huthwaite is careful to describe the four question types as a framework “rather than a rigid sequence”, so treat the order above as a default you’re allowed to break, not a checklist. More on structure in our discovery call guide.

Demo and objections, whenever they arrive

The load-bearing line here isn’t a rebuttal. It’s a stall you’re entitled to.

• Prospect: “This looks like it’d be a nightmare to get past our security team.”
• “Before I answer that, can I check what’s behind the question? Is it a policy you’ve already been burned by, or a general worry?” DOES THE WORK. craft, no number. I couldn’t find a published win-rate for it and I’m not going to invent one.
• Then, and only then, the answer. One sentence, not four.
• “Does that clear it, or is there a version of the concern I’ve missed?” craft, no number

Half the objections you hear are two objections wearing one coat, and answering the outer one wastes the call. Our objection handling guide goes through the common pairs.

If you’d rather have the prompt during the call than the post-mortem afterwards, that’s what we built Craqly’s Sales Call mode for. It splits into Discovery, Demo, Negotiation and Closing sub-modes, and the help page states the talk gauge target plainly: “Target range is 35-55% talk time for the rep”, with framework prompts for BANT, MEDDIC and SPIN. Starter is $0 with 20 credits a month, where 1 credit is 1 minute of live session. Pro is $38 a month, or $19 a month billed yearly. Stealth is on every tier including the free one. Prices checked on craqly.com/pricing on 20 August 2026.

Closing, last eight minutes

Gong’s figure for the fastest deals: the seller “spent 53% more time discussing ‘next steps’ during the first meeting” than in average deals. Not the last meeting. The first.

• “Let me play back what I heard, and correct me where I’m wrong.” Then three sentences. Their words, not your feature names.
• “On price, you’re looking at roughly X for the shape you’ve described.” Same Gong roundup says win rates are 10% higher when sellers raise price on the first call, and that the strongest reps do it in the 38 to 46 minute window.
• “So the next step is 30 minutes with your CFO. Does Thursday afternoon work, or is Friday morning easier?” DOES THE WORK. Named person, named length, two dates.
• “I’ll send a recap tonight with the two open items in it.” craft, no number, though it’s the cheapest promise in sales. Templates for that recap are in our follow-up post.

What’s actually behind each marked line

Three of the five have a number. Two don’t, and I’ve said which. That split is the honest state of scripted selling in August 2026. Most of what gets sold as a proven script is two or three findings somebody genuinely measured, wrapped in a far larger volume of stage direction that has never been tested against a control group by anyone, and the untested wrapping is reliably the part that gets copied hardest. Say which is which.

What I’d argue with in my own post

An opinion that might be wrong: I think the cold-call opener finding is fragile in a way the others aren’t. Gong measured it on a corpus of calls where most reps weren’t using it. Once a line gets published in a research post and copied into a few hundred sequences, the pattern interrupt stops interrupting anything, and I’d expect the 10.01% to sag toward baseline without anyone publishing the correction. The Implication question doesn’t have that problem. Asking a buyer to price their own pain works whether or not they’ve seen it before.

Rehearse the marked lines and improvise the rest. Then do the thing Rackham says produced the 17%, which is get someone to sit with a recording of your call and tell you which of your questions were Situation questions in disguise. Rapport is the easy half, and we’ve covered it in a separate post. The awkward half is counting.