---
title: "Cold Call Opening Lines: Why \"Did I Catch You at a Bad Time?\" Backfires"
description: "Gong measured this opener twice, six years apart, and got 0.9% and 2.15%. It lost both times, but the size of the penalty is unknown, and the sentence is the smaller lever anyway."
author: "Sarah Rodriguez"
category: "SDR"
date: 2026-10-14T09:10:00.700Z
canonical: "https://salesbrew.co/blog/cold-call-opening-lines-why-bad-time-backfires"
---

# Cold Call Opening Lines: Why "Did I Catch You at a Bad Time?" Backfires

![call center, headset, woman](https://cdn.pixabay.com/photo/2017/05/01/14/59/call-center-2275745_1280.jpg?w=1200&q=75)

> Gong measured this opener twice, six years apart, and got 0.9% and 2.15%. It lost both times, but the size of the penalty is unknown, and the sentence is the smaller lever anyway.

Nobody teaches a rep to open with "Did I catch you at a bad time?" out of malice. It is taught as courtesy. The trouble is that it is a yes-or-no question, the honest answer is almost always yes, and you handed the buyer that answer before you said what you wanted.

Gong has measured this line twice, on two different datasets, six years apart. It came last both times. Where the two studies disagree is more instructive than where they agree, and almost nobody quoting the figures has noticed.

## Where One Dataset Undercuts the Other

Start with the newer one. Armand Farrokh's write-up, [built on 300M+ cold calls recorded on Gong](https://www.gong.io/blog/the-best-and-worst-cold-call-openers-backed-by-data-from-300m-calls), validated four openers. "Did I catch you at a bad time?" finished last at 2.15%, which the page describes as barely above the average baseline. "How's your day going?" reached 7.6%. A permission-based line, asking for thirty seconds to explain the call, came in at 11.18%. The winner, asking whether the buyer has heard the company's name mentioned, landed at 11.24%.

Now the older one, which is where most of the numbers circulating online actually originate. [Gong's 2018 analysis of 90,380 cold calls](https://www.gong.io/blog/cold-call-opening-lines) put the same bad-time question at 0.9% against a stated baseline of 1.5%, and described it as making a rep 40% less likely to book the meeting. On that dataset "How have you been?" reached 10.01%, a 6.6X multiple on the baseline, "How are you?" managed 5.2% or 3.4X, and leading with a proactive reason for the call lifted success by 2.1X.

Put the two side by side and the bad-time opener goes from 40% worse than baseline to slightly better than it. Neither Gong page mentions the other's figure. The datasets are six years apart, the call populations differ, and the only defensible reading is that this line is reliably the worst of the ones tested, while the size of the penalty is unknown. If you want the 2018 dataset in detail, we take it apart in [what 90,380 connected calls really show](/cold-calling-success-rate-what-90380-connected-calls-show).

## Where These Success Rates Actually Come From

Here is the thing that makes every one of those percentages less impressive than it sounds: they are conversation-level rates. Somebody already picked up. Getting to that point is the harder arithmetic, and it has its own numbers.

Belkins published benchmarks drawn from more than 175,000 dials logged across 2025. Per [that dataset](https://belkins.io/blog/cold-calling-benchmarks), the per-dial connect rate was 9.9%, 24.5% of unique prospects connected at least once across a sequence of attempts, and 4.6% of the conversations that did happen ended in a booked meeting.

Lay the two together honestly. Roughly one dial in ten turns into a conversation, and your opening sentence is only competing inside that tenth. Lifting a line from the bottom of [Gong's validated range](https://www.gong.io/blog/the-best-and-worst-cold-call-openers-backed-by-data-from-300m-calls) to the top of it changes the conversations you already won, not the nine dials in ten that never got answered. Reps who blur those two denominators end up disappointed by a change that genuinely worked.

## Reactance Is Real, but Not for the Reason Reps Are Told

The folk explanation for why the bad-time opener fails is that people enjoy saying no because it restores their sense of control. It is a pleasing story and it is not what the research measures.

Zixi Li and Jingyuan Shi published a set of meta-analyses in Human Communication Research covering 28 articles, 33 studies and 146 effect sizes from 2005 to 2024. Per [the published record](https://scholars.hkbu.edu.hk/en/publications/message-effects-on-psychological-reactance-meta-analyses/), language that strongly threatens someone's freedom raised anger at r = .21 across 28 samples and reactance at r = .20 across 53, while anger was negatively associated with being persuaded at r = minus .23 across 42 samples.

Read what that actually says. Reactance is not a pleasure in refusing. It is irritation produced by being told what to do, and irritation predicts not being persuaded. The mechanism runs through annoyance, not through satisfaction. Note also that [the same analyses](https://scholars.hkbu.edu.hk/en/publications/message-effects-on-psychological-reactance-meta-analyses/) found no meaningful difference between gain and loss framing on reactance, which should temper any team planning to rewrite its whole script around framing.

The effect sizes are small. Nobody should claim a cold call turns on an r of .21. What the evidence supports is modest and useful: a controlling opener costs you a little, consistently, and the cost arrives as annoyance rather than as a power game.

## A Weak Opener Is a Research Problem Wearing a Script

The standard response to a thin week of calls is a script rewrite. We think that is almost always the wrong order, and the winning opener in Gong's own newer data is the evidence.

Look at what the top line actually does. Asking whether the buyer has heard your name mentioned only works if there is some chance they have. Asked by a rep from a company nobody in that market has encountered, it is a bluff, and a buyer who answers no has now established that the caller was fishing. The permission-based line finished within a rounding error of it and makes no claim at all. So the measured gap between the best opener and a plain request for thirty seconds is almost nothing, while the gap between either of those and the bad-time question is large. The expensive mistake is not choosing the second-best words. It is calling somebody you have no reason to call.

That is why our order is: fix the list, then the reason, then the sentence. A rep with a real trigger can open badly and still get a conversation, because the second sentence rescues them. A rep with no trigger cannot be saved by any opener in the study, and will spend the quarter rotating scripts. For the research that produces a genuine trigger, see [our template for a brief a rep will actually read](/the-account-brief-your-rep-actually-needs-pzhr), and for what qualifies as a reason, [our piece on relevance that is not trivia](/personalization-is-not-knowing-their-alma-mater-uubf).

Show us a clean test, with the list and the trigger quality held constant, where opener choice still dominates the result, and we will concede the point. That test is worth running precisely because it would settle the argument. Until somebody runs it, the measured spread between the two best openers says the sentence is the smaller lever.

## Three Parts, No Fourth

A workable opener has three jobs and no fourth. Name yourself, give a reason, make a small ask.

- **The identifier.** Your name and your company, said slowly enough to be understood. Rushing this reads as evasion.

- **The reason.** One clause that ties the dial to a fact about this particular person: a role started last month, a team being built, a complaint their peers keep repeating. In [Gartner's June 2025 release](https://www.gartner.com/en/newsroom/press-releases/2025-06-25-gartner-sales-survey-finds-61-percent-of-b2b-buyers-prefer-a-rep-free-buying-experience), drawn from 632 buyers, 73% said they steer clear of a supplier whose outreach is not relevant to them. Dialling does not buy you an exemption from that.

- **The ask.** Thirty seconds, not a meeting. The small ask is what the permission-based opener is doing, and it tested within a whisker of the best line Gong found.

What to leave out: any question whose easiest answer ends the call, any apology for existing, and any implication that you have spoken before when you have not.

## The Line Where Craft Meets the Rulebook

That last item deserves its own paragraph, because the best-performing opener in the 2018 dataset, "How have you been?", works partly by sounding like a continuation of a relationship that never existed. Gong is candid that it functions as a pattern interrupt.

It is worth knowing the terrain here. The FTC's Telemarketing Sales Rule carries a business-to-business exemption at [16 CFR 310.6(b)(7)](https://www.law.cornell.edu/cfr/text/16/310.6) for calls made to induce a business to buy goods or services, which is why B2B callers sit outside the Do Not Call machinery. The same provision states that the exemption does not extend to the requirements of 310.3(a)(2) and (4), the parts that prohibit misrepresentation and false or misleading statements made to induce a purchase. We are a sales publication and not your counsel, so take the detail to someone who is.

Our point is about craft and priority rather than compliance. An opener whose performance depends on a buyer briefly misremembering you is a short-term trade: it buys a conversation and spends the goodwill you were calling to build. There are lines that test nearly as well and claim nothing, which is the argument we make at length in [authenticity in prospection](/authenticity-in-prospection).

## Testing It Without Fooling Yourself

- **Transcribe the opening line from thirty recordings.** Write each down word for word. Count how many end with a question the buyer can answer to make you go away.

- **Change one variable.** Same rep, same list, same hour of the day, different first sentence. Anything else moving makes the result unreadable.

- **Measure meetings per hundred conversations, not per hundred dials.** Opener quality cannot touch a connect rate, and mixing the denominators is how teams talk themselves into the wrong conclusion.

- **Wait for enough calls.** At the rates in these datasets, twenty calls tells you nothing. Give it two or three weeks before anybody declares a winner.

## What Not to Chase

One popular shortcut does not survive measurement. Orum analysed over 1.5 million calls made on its platform during autumn 2021 and tested whether dialling from a local area code helps. Per [that analysis](https://www.orum.com/blog/local-presence-dialing-impact-connect-rates), a chi-square test found no significant difference, with local calls connecting at 4.6% against 5.5% for non-local. The data is from 2021 and no newer version of it exists publicly, so treat it as dated rather than settled. Still, it is the only real evidence on the question, and it points away from the trick.

Expect some discomfort as a team drops the apologetic opener. Reps who have spent years softening the first line will feel abrupt. Tell them plainly that brevity reads as respect, and that a no on a cold call is weather rather than a judgement on them. Be kind about it, because the volume of no is the hardest part of this job and nobody needs a lecture on top of it.

## Where to Start Tomorrow

Pull your last thirty call recordings, read the opening sentences aloud, and delete every one that hands the buyer an exit. Replace it with a name, a reason that would survive the buyer asking "why me?", and a request for thirty seconds. Then fix the list, because that is the lever these studies keep pointing at. No script guarantees a meeting, and the published rates come from other people's markets, so test your own and judge it on conversations rather than dials.

## FAQ

### How bad is "Did I catch you at a bad time?" really?

It finished last in both Gong analyses, but the two disagree on the margin: 0.9% against a 1.5% baseline in the 2018 dataset of 90,380 calls, and 2.15% in the later 300M+ dataset where the page calls it barely above baseline. Treat it as reliably the weakest line tested and treat any specific penalty figure with suspicion.

### What should replace it?

A name, one clause explaining why this person and why now, and a request for thirty seconds rather than a meeting. A plain permission-based line tested within a rounding error of the best-performing opener Gong found, and it has the advantage of claiming nothing that is not true.

### Does calling from a local area code help?

The only real evidence says no. Orum tested over 1.5 million calls from autumn 2021 and a chi-square test found no significant difference between local and non-local connect rates. The data is dated and nothing newer has been published, so it is the best available answer rather than a final one.

### How long should an opener test run?

Long enough that the numbers mean something, which at these success rates is weeks rather than days. Hold the rep, the list and the time of day constant, change only the first sentence, and judge the result on meetings per hundred conversations so that connect-rate noise stays out of it.


---
Source: https://salesbrew.co/blog/cold-call-opening-lines-why-bad-time-backfires