Teams that have never cold called usually fail in the same way. They buy a dialler, write a script, make thirty calls over two weeks, conclude that nobody answers the phone any more, and go back to email.
Thirty calls is not an experiment. It is roughly three connected conversations, in a channel where you are measurably bad for the first hour. The channel did not fail; the sample size did.
This is what the first two weeks actually look like.
Set the expectations first
Before anything is configured, agree these numbers, because they are what stops people quitting.
Most dials will not connect. Depending on market, seniority and data quality, somewhere between 10% and 30% of dials reach a human. That is normal. A dial that rings out is not a rejection, and treating it as one is what makes the activity unbearable.
Most connections will not go anywhere. Of the people who answer, most will say no quickly. That is also normal.
You will be bad for the first hundred dials. Everyone is. Tone, pace, handling an interruption, not sounding like you are reading, these are motor skills and they take repetition rather than preparation.
So judge nothing before 100 dials per person. Write that number down and commit to it. The single most common reason teams abandon calling is quitting inside a sample that could not have told them anything.
The setup
Four things, roughly an afternoon.
Numbers. One number per caller, in the region they are calling. A UK caller dialling UK prospects from a US number will connect at a materially lower rate because the number looks irrelevant and gets ignored. Local-presence numbers help, with the caveat that rotating aggressively through many numbers looks like exactly what spammers do. How local presence actually works covers the trade.
A dialler, and the right kind. For a first-time team, a power dialler, one call at a time, automatically dialling the next record. Not a parallel dialler. Parallel diallers call several numbers at once and produce the connection delay that makes a prospect say “hello? hello?” before you speak, which is the worst possible opening for someone still learning. Start with one at a time. Power versus parallel has the detail.
Disposition codes, kept short. Six to eight outcomes, no more: connected–interested, connected–not now, connected–no, gatekeeper, voicemail, no answer, wrong number, do not call. A long list produces inconsistent data and reps who stop logging. Designing dispositions covers why fewer is better.
Compliance, before the first dial. Do-not-call registration requirements, recording consent, and permitted calling hours all vary by jurisdiction and some carry real penalties. This genuinely needs checking for every country you will call, and it needs checking before rather than after. The cold calling compliance guide covers the landscape, and it is not a substitute for advice on your specific markets.
The first eight seconds
The only part of the call that needs to be scripted, because it is the only part that happens the same way every time.
Three things, in this order.
1. Say who you are, with your full name and company. Confidently and at normal speed. The instinct is to rush this; rushing it signals that you expect to be hung up on, which invites it.
2. Acknowledge the interruption without apologising for it. “I know I’m calling out of the blue” works. “Sorry to bother you” does not, it concedes that the call has no value, and the prospect agrees.
3. Give a one-sentence reason specific enough to be worth hearing. Not “I wanted to tell you about our platform.” Closer to: “I’m calling because you’re hiring three ops people and I work with teams at exactly that point.”
Then ask permission and stop talking. “Can I take thirty seconds to explain, or is now bad?”
That question is doing real work. It gives the prospect control, which converts better than any clever hook, and “now is bad” is useful information rather than a rejection, it is a scheduling problem, which is the easiest kind to solve.
Everything after those eight seconds is a conversation and should not be scripted. Scripts and openers in depth covers the structure.
The three things you will hear, and what they mean
“I’m not interested.” Delivered within four seconds, this is a reflex rather than an evaluation, they have not heard anything to be uninterested in. One attempt is reasonable: “That’s fair, you don’t know what this is yet. One sentence and then I’ll go.” Then honour it. The second attempt is where you become the reason people hate cold calls.
“Send me an email.” Usually a polite exit, occasionally sincere. The response that distinguishes them: “Happy to, what should I make sure it covers?” A sincere request answers the question. An exit does not, and you have your answer without an argument. Either way, send the email, because a called-then-emailed prospect responds better than one who only got email.
“How did you get my number?” Answer it plainly and accurately. “Your number is in a business contact database we license.” Evasion here turns a neutral call into a complaint, and in some markets a complaint has consequences. Honesty also ends the question immediately.
The first two weeks, structured
Week one: get the reps in
Two 90-minute blocks, not fifteen minutes a day. Calling in fragments does not work because the first twenty minutes of any block are warm-up. You get good somewhere around dial forty and then it stops being difficult. Fragmenting means never reaching that point.
Call the least valuable segment of your list. This is the advice people resist most and it is the most important thing here. Your first hundred dials will be your worst hundred dials. Spend them on accounts you can afford to handle badly, and keep the named target accounts for week three when you are competent.
Record calls where lawfully permitted, and listen back to three a day. Fifteen minutes of listening to yourself improves you more than an hour of coaching, mostly because of pace, nearly everyone talks too fast for the first week.
Target: 100 dials per person. Judge nothing.
Week two: start reading the data
Now the numbers mean something.
| Metric | Early-days range | What it tells you |
|---|---|---|
| Connect rate | 10–30% | Data quality and number setup, not skill |
| Conversation rate (of connections, got past the opener) | 25–50% | Your opener, this is the skill number |
| Meeting rate (of conversations) | 10–25% | Targeting and offer fit |
The useful diagnostic is which one is weak.
Low connect rate is a data or number problem. Bad mobile numbers, wrong region, calling the wrong hours. Not a skill problem and no amount of coaching fixes it.
Low conversation rate is the opener. This is the one you can fix fastest, usually by slowing down and cutting the first sentence in half.
Good conversation rate, low meeting rate means the calls are working and you are calling the wrong people. A targeting problem wearing a calling costume.
Fix one at a time, in that order.
What makes people quit, and the countermeasures
The silence. Twenty consecutive dials with no answer feels like failure and is statistically unremarkable. Countermeasure: track dials as the activity metric in week one, not connections. Dials are fully within the rep’s control; connections are not.
Doing it alone. Calling in a room with someone else calling is measurably easier than calling alone, which is one reason remote teams struggle to start. Countermeasure: everyone calls at the same time, on a shared call, muted between dials. It sounds artificial and it works.
The first properly rude person. It will happen in the first week. Countermeasure: say in advance that it will happen, and that it says nothing about the rep. A manager who makes twenty dials themselves in week one buys more credibility here than any amount of encouragement.
Where to point the calls
The highest-value decision, and it is not about calling technique at all.
Call people who have already engaged. Someone who opened two emails, clicked something or viewed your profile connects and converts at a much higher rate than a cold record. If you run email already, your call list should be drawn from engagement rather than built separately.
Call on fresh triggers. A company hiring for a role in your domain, a new leader in the relevant seat, a recent funding round. The call has a reason in its first sentence, which is what makes the opener easy to write.
Do not call your best accounts first. Already said, worth repeating, because it is the mistake with the longest tail.
The short version
Expect 10–30% of dials to connect and most connections to go nowhere, and expect to be bad for the first hundred dials. Judge nothing before then, quitting inside a 30-dial sample is the most common failure.
Set up one local number per caller, a power dialler rather than a parallel one, six to eight disposition codes, and compliance checks for every market before the first call.
Script only the first eight seconds: who you are, acknowledge the interruption without apologising, one specific reason, then ask permission and stop talking.
Week one is two 90-minute blocks against your least valuable segment, 100 dials, no judgement. Week two, read connect rate for data problems, conversation rate for your opener, meeting rate for targeting.
And point the calls at people who already engaged with an email. It is the easiest improvement available and it has nothing to do with calling.
Next: cold call scripts and openers, or power versus parallel diallers.
Frequently asked questions
How many cold calls before you can judge whether it works?
At least 100 dials per person. Thirty dials produces around three connected conversations in a channel where everyone is poor for the first hour, so it cannot tell you anything about the channel.
What should you say in the first seconds of a cold call?
Your full name and company at normal speed, an acknowledgement of the interruption without apologising for it, one specific reason for calling this person, then a permission question (“can I take thirty seconds, or is now bad?”) and then stop talking.
What is a normal cold call connect rate?
Typically 10–30% of dials reach a human, depending on market, seniority and data quality. A low connect rate points at data or number setup rather than skill, so coaching will not move it.
Should a new team use a power dialler or a parallel dialler?
A power dialler. Parallel diallers introduce the connection delay that makes prospects say “hello? hello?” before the rep speaks, which is the hardest possible start for someone still learning.