A cold call script is a structure, not a page you read aloud. Five parts carry it: an opener, the reason for your call, one or two qualifying questions, rehearsed objection lines, and a specific ask for the meeting. The wording changes with each prospect; the sequence does not, and the opener is the part the call data says matters most.
What should a cold call script include?
Five parts, in order, and very little else.
- The opener. Your name, your company, and one line that earns the next twenty seconds.
- The reason for the call. One sentence that is true about this account and not about the other two hundred on your list.
- A qualifying question. One or two open questions about how they handle the thing today.
- Objection lines. Three rehearsed responses to the objections you actually get, written before you dial.
- The ask. A specific day and time, not "would you be open to a chat".
What the ranking pages give you instead is volume: Zendesk publishes 14 templates, lemlist 21. Templates are useful once you can already run the sequence, and close to useless before that, because the thing that fails on a real call is never the absence of a scenario you had not thought of. It is talking too long in part two.
That part has a number on it. Cognism, which sells sales intelligence and so has a stake in the answer, reports in its 2024 cold calling study that only 31% of calls get past the pitch. Two calls in three end where the rep starts explaining the product.
What should you say in the first ten seconds of a cold call?
The opener is the only part of the script with real outcome data behind it.
Gong analysed 90,380 cold calls and scored openers by whether the call ended in a booked meeting, against a 1.5% baseline.
| Opening line | Success rate in Gong's sample |
|---|---|
| "How have you been?" | 10.01% |
| "How are you?" | 5.2% |
| Baseline, all openers | 1.5% |
| "Did I catch you at a bad time?" | 0.9% |
Two caveats before you rewrite your opener. Gong sells conversation-analytics software, so this is vendor-published research. And it is observational: these are the rates for calls that happened to open that way, not a controlled test, so some of the gap is the kind of rep who opens warmly rather than the words themselves.
The direction still holds up against how the calls go. An opener that invites "yes, it is a bad time" gets that answer. A short human question buys you the twenty seconds you need for part two. Gong's same analysis found that stating the reason for your call raises the success rate as well, which is the cheapest change in this article: say why you called before you say anything about your product.
What do you say during a cold call after the opener?
Say why you called, in one sentence, then stop talking.
The reason has to be specific to them. "We work with operations teams on inventory accuracy" is a category. "You have three warehouses and you went live with a new WMS in the spring, and the counts usually go sideways in the first quarter after that" is a reason to stay on the phone. If you cannot write that sentence for an account before you dial, the account is not ready to call.
Then ask one question and let the silence sit. Not a list of discovery questions, one: how they handle it today, or who owns it. The answer tells you whether to keep going or thank them and hang up, and hanging up early is a win on a day with eighty dials in it.
The pitch comes after the question, if at all, and it is two sentences. Everything you want to say about the product belongs on the meeting you are calling to book.
How do you handle the first objection on a cold call?
Write three lines down before you dial, because the moment of most pressure is the worst moment to improvise.
Almost every cold call objection is one of three. The pattern that works on all of them is the same: acknowledge it, ask one question, make a smaller ask than the one you came with.
- "Send me an email." "I will. So it is not a brochure, can I ask one thing first: who picks up inventory counts at the moment, your team or the 3PL?" Then send the email, and say in it that you called.
- "I'm not interested." "That is fair, you have not heard anything yet. The reason I called is that teams running three sites usually eat a week of stock variance after a system change. If that is not happening to you, I will leave it there."
- "We already have someone for that." "Most people I call do. The thing I would want to know in your seat is whether the variance actually closed after they came in. What are you seeing on that?"
None of those are clever. They are short, they hand the conversation back, and they can be said out loud without sounding like a page being read.
Should you read a cold call script word for word?
No, and no one has published a number that settles it either way.
No study comparing verbatim delivery against adapted delivery turned up in the sources checked for this article. The ranking pages are consistent in their advice without data behind it: lemlist tells readers to treat it as a framework, not a transcript. That is practitioner judgement, and it matches what call recordings sound like, but treat it as judgement rather than a finding.
The version that works in practice: read it close to verbatim for your first dozen calls, while the sequence is still something you have to think about. Then drop the paper and keep the order. A rep who knows the five parts can be interrupted at any point and still get back on track, which is the actual job.
How is a cold call script different from an inbound call script?
An inbound call hands you the reason for calling, so the first two parts collapse into one.
Someone who filled in a form or asked for a callback already has a question in their head. The opener becomes a thank-you and a confirmation of what they asked about, the reason-for-the-call sentence is theirs rather than yours, and the qualifying question moves earlier because they are expecting to be asked something. Objection lines still matter, but they are different objections: timing, budget and who else needs to be involved, rather than "who is this". Running an outbound opener on an inbound lead throws away the one advantage you have, which is that they put their hand up first.
What can a cold call script not fix?
A script is the smallest part of the problem when the list is wrong.
The arithmetic in Cognism's own report makes that concrete. From 55,701 dials, 9,247 were connected, producing 5,265 conversations and 254 meetings. The 4.82% they publish as the average success rate is meetings measured against conversations: 254 out of 5,265 is where it comes from, which is our own arithmetic on their published counts. Against dials rather than conversations, on the same arithmetic, it is nearer half a percent. Most of the outcome is decided by who is on the list and how many numbers connect, long before anyone says a word.
So three cases where a better script changes nothing. You are calling accounts with no reason to care, and no opener invents one. Your contact data is poor, and the connect rate is the ceiling on everything downstream. Or the phone is simply not where these buyers are, in which case the honest move is to test another channel rather than rewrite the opener a fourth time.
If the conclusion is that the function should sit outside the building, what cold calling services cost and how to vet one covers that decision. If it is that software should place the calls, what AI cold calling can and cannot do is the more useful page.
Frequently asked questions
What is the 80/20 rule in cold calling?
It is the Pareto principle borrowed, not a cold calling finding. Vendors state it as a small share of effort producing most of the results, and name the economist Vilfredo Pareto, whose observation concerned the distribution of wealth rather than sales calls. No study specific to cold calling establishes that split, and the pages asserting it offer practitioner anecdote in place of data. Use it as a reminder to concentrate on your best accounts, not as a measured ratio.
What are the 5 stages of cold calling?
There is no agreed set. Sales training sites publish their own and the counts disagree: one names five steps, another six components, another four. None cites a common original study, because each author drew the boundaries differently. The sequence worth memorising is the one in this article: open, say why you called, ask one question, handle the objection, ask for a specific time.
How long should a cold call last?
Shorter than most reps expect. Cognism puts the average cold call length at 83 seconds across its sample. That average includes fast rejections, so it is not a target to pad towards. The useful read is that the decision happens early: if you have not earned interest inside the first minute, more talking rarely recovers it.
How many times should you call a prospect before giving up?
Cognism's report names 3 as the optimal number of cold call attempts for connecting with a lead. Spread them across different days and times rather than three times in one morning, and pair them with another channel so each attempt is not a cold start. If three well-spaced attempts produce no connection, the number itself is usually wrong.
Should you leave a voicemail when the prospect does not pick up?
Leave one, once, and keep it under twenty seconds: your name, your company, the one-sentence reason for the call, and that you will follow up by email. Do not pitch to an answering machine, and do not leave a message on every attempt. Its job is to make your name familiar on the next call, not to get a call back.