Cold calling scripts that survive contact with a real prospect

Most cold call scripts are written to be read and fail because they are heard. A prospect decides within a few seconds whether this is a call worth having, and they decide on tone and specificity, not on your value proposition. Which means a script's job is much smaller than most scripts assume: control the opening, prepare for four objections, and get out of the way.

The first fifteen seconds

Four beats, in order, each one short.

What none of the four beats includes: your founding story, your funding round, a feature list, or the phrase “I wanted to reach out”. Every one of those spends the attention you just borrowed.

The four objections that account for most calls

“I'm not interested”

Almost always a reflex fired before they processed anything. Do not argue with it. Acknowledge and give one concrete hook.

“Completely fair, you don't know what it is yet. The one thing I'd say is that teams your size usually find half their outbound numbers get labelled as spam within a quarter. If that's not happening to you, I'll leave you alone.”

That works because it is falsifiable. They can check. A vague benefit claim gives them nothing to test and nothing to say except no again.

“Send me an email”

Usually a polite exit, occasionally genuine. Accept it, and take something for it.

“Sure. So I send something useful rather than generic: are you dealing with answer rates, compliance, or cost right now?”

You get a qualifying answer either way, and you have earned a reason to follow up. Fighting to stay on the call after this converts almost nothing and costs you the email too.

“We already use [competitor]”

This is the best objection you can get. It confirms budget, need, and category awareness. Do not attack the incumbent.

“Good, they're solid. Most people who move do it for one specific reason: they're paying per seat for people who barely dial. Is that a live issue for you, or is it working fine?”

Naming the specific reason people switch is far stronger than asking whether they are happy. It gives them permission to admit a problem without conceding they chose badly.

“Now's not a good time”

Take it at face value and get a specific alternative. “When would be better?” invites a vague brush-off; two named options do not.

“No problem. Thursday morning or Friday afternoon: which is less bad?”

Lines to delete today

Structure it so it survives being read aloud

• Bullets, not paragraphs. A rep reading a paragraph sounds like a rep reading a paragraph.
• Script the opener and the objections. Nothing else. The middle of a good call is a conversation, and scripting it is what produces the robotic delivery everyone complains about.
• One page maximum. Two pages means nobody reads it after week one.
• Write in spoken English. Read it aloud while writing it. Anything you would not say to a colleague, cut.

The parts of a cold call the script cannot fix

It is worth being blunt about this, because teams often rewrite scripts to solve problems that live elsewhere.

1. Whether the phone is answered at all. If your numbers are labelled “Spam Likely”, the script is irrelevant; nobody hears it. That is a number reputation problem.
2. Whether the list is any good. No opener rescues a list of wrong people. Specificity in the second beat is only possible if the list supports it.
3. How many calls get made. A brilliant script used forty times a day loses to a decent one used two hundred times, which is what dialing modes are for.
4. Whether you are allowed to call. Consent, DNC, and calling hours are not script concerns; see the TCPA guide. But they end campaigns far more abruptly than a weak opener does.

How to actually improve a script

Not by workshopping it. By reading transcripts. Recording and transcription turn the question “does this opener work?” from an opinion into a measurement: pull fifty calls, mark where the prospect first interrupted, and look at what preceded it. The pattern is usually obvious within an hour, and it is usually not what the team assumed.

Change one line at a time. Scripts that change wholesale cannot be attributed, and every rep quietly reverts to their own version anyway.

Voicemail: leave one, but not the one you are leaving

Most cold calls reach voicemail, which makes the voicemail message a substantial share of your total output, and it is almost always an afterthought delivered by a rep who has stopped concentrating.

• Under twenty seconds. Long voicemails are deleted before the point arrives.
• Say your number twice, slowly, at the end. The person is reaching for a pen on the second listen, not the first.
• Give one specific reason, not a pitch. The message is buying a callback, not making a sale.
• Do not promise to call again “to follow up”. It tells them they can ignore this one safely.
• Say your name and company clearly at the start. Whispered attribution is what makes a voicemail feel like a scam.

One practical consequence for tooling: if reps are recording the same message forty times a day, voicemail drop both improves consistency and returns roughly half an hour per rep per day. The message gets better because it was recorded once, deliberately, rather than improvised at four in the afternoon.

Gatekeepers

The advice to “get past the gatekeeper” is both dated and counterproductive. Receptionists and executive assistants are paid to route calls correctly, and treating them as an obstacle is how you get routed to nowhere permanently.

“Hi, I'm probably calling the wrong person. I'm trying to find whoever owns outbound calling for the sales team. Would that be Sarah, or someone else?”

That works because it is true, it asks for help rather than access, and it gives them something they can actually answer. Naming a plausible person invites a correction, which is more useful than a yes or a no. Tricks (pretending to be returning a call, claiming a prior relationship) get remembered, and they get shared internally.