LinkedIn cold message examples, and how an operator rewrites them

01The limits that shape every LinkedIn connection message

Before any rewriting, the numbers. Many template pages still say a note can run to 300 characters and an InMail to 8,000. LinkedIn’s own help pages say otherwise, and on a few points they don’t fully agree with each other either.

We read these pages logged out on 28 and 29 September 2026. The page that says three free notes a month was updated about five months before that; the one that says five, about a year before.

LinkedIn changes limits in the app before its help pages catch up, and signed-in members, Premium members especially, can see different numbers. If the counter in your own composer disagrees with this table, believe the composer.

02Ten messages founders delete, rewritten

Each example gives the context first, the way it would appear in a reason log: who sent it, who received it and the signal behind the rewrite. The signals are the part most templates skip, and they’re covered in more depth in the buying signals behind each first line.

1. The empty compliment

Sample Sender: founder of a 9-person company that sells customer onboarding software. Prospect: VP of Customer Success, 60-person HR software company, Denver. Signal: her post on Tuesday about losing two renewals to slow onboarding.

What’s wrong

• The praise fits every profile on LinkedIn, so it lands on none of them; Laura can’t tell whether you read a word she wrote.
• There’s no “why me” and no “why now”, which makes accepting a favor to you.
• There’s nothing to answer: “I’d love to connect” is a statement, and statements don’t get replies.

Why it works: the compliment survives, but now it’s earned, because “painfully honest” points at something she actually wrote. The question offers two causes she has probably argued about at work already, so answering takes her ten seconds and tells the sender which problem to talk about. The product doesn’t appear, and it doesn’t need to until she answers.

2. The pitch in line one

Sample Sender: founder of a 14-person company that sells security compliance software. Prospect: CTO, 35-person fintech API company, New York. Signal: a “Security and Compliance Lead” job on the company’s careers page, posted last week.

What’s wrong

• Line one is about the sender: in the preview, Mark sees a vendor category before he sees any reason to care.
• A demo is a 30-minute ask from a stranger, and nobody books one from an invitation note.

Why it works: the signal comes from outside LinkedIn, from a page anyone can check, which makes it harder to fake and easier to believe. Either answer tells the sender whether the product is relevant. If it is, the pitch goes in message two, built on Mark’s own words.

3. The broken merge field

This is the message most founders picture when they hear the words “LinkedIn automation”, and it’s why so many of them delete anything that smells of a template.

Sample A typical automated message; we wrote it to show the pattern.

What’s wrong, marked the way I’d mark it up for a client:

1. The merge field broke, so now they know it’s a mail merge.

2. An opener that says nothing.

3. Praise with no detail, which nobody believes.

4. A big claim with no proof.

5. It asks for a meeting before giving a reason to take one.

Example context for the rewrite: Priya, CTO at a 45-person SaaS company in Manchester, posted on Sunday about a 300-question security questionnaire.

Why it works:

1. It starts from her post, so she knows a person read it.

2. One real detail, which a template can’t fake.

3. A question she can answer in one line.

4. No pitch and no meeting ask, because those come after she replies.

4. The long self-introduction

Sample Sender: Chris, co-founder of a revenue forecasting startup. Prospect: Head of Revenue Operations, 50-person logistics software company, Chicago. Signal: his comment on another founder’s post, saying his team rebuilds the forecast by hand every Monday.

What’s wrong

• The only person described is the sender: Daniel learns about Chris’s career and nothing about why Chris picked him.
• It doesn’t fit, because the box stops at 200 characters and this note hits the wall mid-word, before it ever reaches a reason.

Why it works: Chris’s credentials are one click away on his profile, and Daniel will click if the question lands. A comment makes a strong signal because it’s a person naming the problem out loud without being asked. And “how many hours” invites a number, which is where a business case starts.

5. The meeting ask first

Sample Sender: founder of a 20-person company that sells finance close software. Prospect: CFO, 30-person marketplace startup, Austin. Signal: she was a guest on a finance podcast last month and said month-end close takes her team 12 days.

What’s wrong

• It asks for 15 minutes before giving one reason to spend them.
• “Tuesday or Wednesday” is an old sales move: two choices, neither of them no. Founders have seen it enough times to resent it.

Why it works: I think podcasts are one of the most underused signals there are. The prospect chose to talk about the problem in public and at length, and quoting one number from the episode proves you listened. The meeting ask moves to the second or third message, after she has confirmed the problem is real and current.

6. “Quick question”

Sample Sender: founder of a 10-person company that sells sales hiring services. Prospect: founder and CEO, 12-person developer tools company, Seattle. Signal: his post saying he’s still on every sales call and wants help by Q1.

What’s wrong

• “Quick question” is one of the most recognizable openers in cold outreach, and the question behind it is rarely quick.
• It’s a routing question. It asks Tom to do your research, and if he answers “me”, you’ve learned nothing you couldn’t have found in two minutes.

Why it works: it keeps the short, easy feel that “quick question” was reaching for, but the question is now about a decision Tom is actively making. Either answer tells you something real about his next quarter.

7. “We’re both in a group”

Sample Sender: founder of a 16-person pricing software company. Prospect: co-founder, 20-person procurement software company, Boston. Signal: her answer, in a founders group they both belong to, to a question about pricing a first enterprise deal.

What’s wrong

• Thousands of people share that group. It’s the weakest reason there is to contact someone, and she knows it.
• “Share ideas” gives her nothing to reply to.

Why it works: the group turns into the place where you saw her, not the reason you’re writing. The reason is what she said there, and the question picks up her own thread, which people are usually happy to finish.

8. Funding congratulations, then a pitch

Sample Sender: founder of an 8-person outbound consultancy. Prospect: CEO, 25-person data infrastructure company, Atlanta. Signal: a Series A three weeks ago; the announcement said the money goes to “doubling the sales team”, and the careers page lists three account executive roles.

What’s wrong

• Every vendor with a funding alert sends this in the same week; the congratulations is a wind-up, and James can see the pitch coming from the first line.
• “Build out your outbound engine” describes a need he never mentioned.
• Funding on its own tells you he has money, not what he plans to spend it on.

Why it works: it skips the congratulations he has already had plenty of and goes straight to what the round means, in his own announcement’s words. New reps with no pipeline is a problem a founder feels in month two, so the question arrives before he has solved it.

9. “Bumping this”

Sample Sender: founder of a 12-person compliance software company. Prospect: VP of Engineering, 70-person health-tech company, London.

She accepted the invitation a week ago and didn’t answer the first message. Since then, her company has published a post about its new SOC 2 report, mentioning a five-month audit.

What’s wrong

• It adds nothing: it asks her to reread a message she has already decided not to answer.
• “In case you missed it” is a polite fiction, and you both know it.

Why it works: a follow-up needs a second reason, not a reminder of the first. If nothing new has happened, don’t send one. On client accounts our rule is one follow-up per person, written by the operator, and then we stop.

10. The InMail wall of text

Sample Sender: founder of a 22-person company that sells employee onboarding software. Prospect: VP of People, 150-person software company, Chicago. Signal: her post two weeks ago saying new engineers take 90 days to ramp and the first three weeks go on waiting for system access.

[Four more paragraphs follow, with a case study, a bulleted feature list and a calendar link: about 1,700 characters in all.]

What’s wrong

• The subject line says nothing: “Partnership opportunity” reads as a pitch in any inbox, and the subject is the first thing she sees.
• The body uses most of the InMail limit (1,900 characters on one LinkedIn help page, 2,000 on a newer one; see the table above), and that’s a ceiling, not a target.
• Four paragraphs in, Olivia still doesn’t know why she was picked.

Even LinkedIn’s composer pushes back: “As your InMail gets longer, the character counter will change color and encourage you to keep your message concise.”

Why it works: the subject line is her own topic, so the message reads as a reply to her rather than a pitch at her. The body fits on one phone screen and asks one thing.

Related LinkedIn InMail vs connection request

03When they reply: three examples

A reply is where a first line turns into a meeting, and it’s the part most template pages never show. Answer within one business day, in the same voice as the note, and hold the pitch until they’ve told you the problem is real.

If you’re wondering who writes every first line on a GTME account: one operator writes those and every reply, so the voice never changes hands mid-conversation.

1. A positive reply. Priya, from example 3, answers:

Ha. Every month now, and sales keeps promising custom terms to close deals.

Answer (Example):

That’s the expensive part: the questionnaire shows up after the promise. Our product answers the standard questions from policies you already have, so your team only writes the custom ones. Worth 20 minutes to try it on one of your real questionnaires? Here’s my calendar, or send me two times that suit you.

Now there’s a reason to pitch, and the pitch uses her words, not yours. The call comes with an easy alternative to clicking a link.

2. “Not now.”

Not a priority this quarter.

Answer (Example):

Understood, and thanks for saying so. Mind if I check back in February? If it moves up before then, you know where I am.

Then stop, and put February in your calendar. Don’t answer “not now” with a case study.

3. “Is this automated?”

This one deserves a straight answer, because people ask it after they’ve been burned. The operator writes every note and every reply, and the client takes their own meetings, so the answer matches who actually wrote it:

Fair question. The invitation went out through a scheduling tool. The note and this reply were written by a person on my team, not a template, and I take every call myself.

A straight answer costs nothing. A dodge, once they spot it, costs the whole conversation.

04US and UK tone

GTME writes for founders who sell to US and UK buyers, and the same line can land differently in each. What follows is our house style, not research.

Invitation note (Example):

US: Hi Jordan, saw you’re hiring your first AE. Will they inherit your LinkedIn pipeline, or start from zero?

UK: Hi Oliver, noticed you’re hiring your first account executive. Will they be inheriting a pipeline, or starting from scratch?

Answer to “not now” (Example):

US: Makes sense. Want me to check back in February?

UK: Fair enough. Shall I check back in February, or would you rather I didn’t?

The UK versions leave more room: “noticed” instead of “saw”, the job title written out, and a plain way to say no. The US versions move a beat faster. Spelling follows the buyer’s market too, so a note to Chicago says “prioritize” and a note to Manchester says “prioritise”.

05Before you send: a checklist

Run every LinkedIn cold message through these six checks before it goes:

1. It’s under 200 characters, counted rather than guessed.

2. It names one reason they can check: a post, a comment, a job ad, a launch or a podcast.

3. It asks one question they can answer in a line.

4. There’s no pitch, no link and no meeting ask.

5. A person typed every name and detail, so there are no braces, brackets or blanks left.

6. If they reply “why me?”, the answer is already in the note.

Before you send one of your own, check your own note’s length and first line with our free checker.