A connection request framework that survives the 300-character limit

The note has one job

The invite note is not selling your product. It is selling the act of clicking Accept, which is a much smaller thing, and the way to sell a small thing is to make it look cheap rather than valuable.

That is why most notes fail. They are written as compressed sales emails: a value proposition, a credential, a question about the reader's priorities this quarter. All of that signals that accepting has a price attached, so the reader declines a pitch they have not read yet rather than accepting and dealing with it later.

The reader is also making the decision in about two seconds while looking at your headline and photo alongside the note. Which means the note and the profile are one artefact, not two. A well-structured note on a profile that looks like a placeholder does worse than a mediocre note on a profile that reads like a real practitioner, which is the unglamorous reason profile work comes before copy work.

The four slots

Every note that works does four things, in roughly this order. Naming them as slots matters, because it turns editing from taste into a checklist: read the note, ask which slot each clause serves, delete the clauses that serve none.

The order is a default rather than a law. The relevance signal goes first because it is what stops the eye. The disclaimer goes last because it is the reassurance the reader needs after they have understood the ask.

Budgeting the characters

The invite note is short by design, and LinkedIn also limits how many notes some account types can send in a month, so confirm both against your own account before you build a campaign that depends on notes. Whatever the current limit is, the budget below holds in proportion.

• Greeting and first name: a small fixed cost, worth paying because omitting it reads as a broadcast
• Relevance signal: the largest share, because it is the slot doing the work
• Reason for reaching out: one clause, roughly the same size as the ask
• Low-commitment ask: one short sentence
• What you are not doing: the shortest of the four, a half-sentence
• Sign-off: your first name only, or nothing

What does not get a slot: your company's description, your job title, a link, a statistic, a compliment about their content that could apply to anyone, and the phrase "I'd love to connect". None of them advance the decision the reader is making, and every character they take comes out of the relevance signal.

Count characters on the final version rather than estimating. A note that is truncated mid-clause looks careless in exactly the place where you were trying to look careful, and it usually truncates the disclaimer, which is the slot you can least afford to lose.

Three rewrites

The fastest way to internalise the slots is to watch weak notes get repaired. Each of these is a paraphrase of a pattern you will recognise.

Before: "Hi Sarah, I'd love to connect and share how we help companies like yours scale their outbound pipeline with proven strategies. Open to a quick 15-minute chat this week?" — What is wrong: zero relevance signal, the reason is about the sender, the ask costs fifteen minutes, and there is no disclaimer, so the reader assumes a demo sequence starts on acceptance.

After: "Hi Sarah — saw you are hiring two SDRs in Manchester while running the team yourself. We work with sales leads in that exact gap, so I keep an eye on people doing it. Happy to just connect; I am not going to pitch you in the DMs." — the relevance signal is specific and checkable, the reason explains the sender's interest without a value proposition, the ask is the connection itself, and the disclaimer answers the fear directly.

Before: "Hi David, great post on AI in recruitment. Fully agree. Would love to be connected." — What is wrong: the relevance signal is real but weightless, because anyone can say "great post". There is no reason and no ask, so the note is polite noise.

After: "Hi David — your point about screening calls being the bottleneck rather than sourcing matched what we keep hearing from agency owners. That is the part of the process I work on, so I would rather follow you than not. No ask attached." — the relevance signal quotes the actual argument, which proves reading, and the reason places the sender in the same problem.

If your relevance signal would still make sense addressed to a hundred other people, it is not a relevance signal.

Before: "Hi Priya, I help B2B SaaS founders book 20+ qualified meetings a month. Let's connect." — What is wrong: it leads with a claim, which forces the reader to evaluate the claim rather than the person, and the sender has said nothing that shows they know who Priya is.

After: "Hi Priya — you have gone from founder-led sales to a first sales hire this year, which is the messiest handover in the business. That transition is what I spend my time on. Connecting is all I am after for now." — the relevance signal is a situation rather than a fact, which is harder to fake and more flattering because it implies understanding.

The slot everyone skips

The fourth slot — naming what you are not going to do — is the one most people cut for space, and it is the one doing the most work per character. The reader's hesitation is not "is this person interesting". It is "what happens to my inbox if I accept".

Answer that directly and the hesitation has nowhere to go. "Not going to pitch you" works. "No ask attached" works. "Happy to just be connected" works. What does not work is a softened version that still implies a sequence, such as "no pressure at all if the timing is not right", because that phrase is what pitches say immediately before pitching.

The obvious objection is that this is a promise you then break by sending a message later. It is not, if you keep it. The disclaimer covers the moment of acceptance, not the rest of your relationship, and the honest version of the promise is that you will not open with a pitch. Keeping that promise is what makes the first real message work at all, and it is why the first message after acceptance should set context rather than sell.

Adapting the slots per segment

The slots stay; what fills them changes. This is the practical reason to hold a framework rather than a template library — a new segment needs four decisions, not a new page of copy.

Notice that the ask barely moves. That is deliberate. Across every segment the ask is almost always just the connection, because the invite note is not where deals begin. What changes materially is the relevance signal, which is the slot that requires you to have built a list where each row carries a reason for being there.

Testing notes without fooling yourself

The measurement is simple: invites accepted divided by invites sent, calculated per segment and per sending profile. If 100 people receive a note and 34 accept, that is 34 per cent for that cohort, and the only comparison that means anything is against the same segment sent from the same profile.

1. Change one slot at a time. A rewritten note tells you nothing about which change mattered.
2. Keep the segment fixed while testing copy, and keep the copy fixed while testing segments.
3. Give each version enough recipients that a handful of decisions cannot flip the result. Small cohorts produce dramatic percentages that mean nothing.
4. Run versions concurrently rather than sequentially, so a quiet fortnight does not get attributed to the copy.
5. Record the winner alongside the segment it won for, because a note that wins with founders can lose badly with recruiters.

And track the second number: of those who accepted, how many replied to the first real message. A note that wins on acceptance while losing on replies has usually overpromised in the disclaimer or undersold the reason, which produces connections who do not remember why they accepted. If you want the whole loop run for you — note, sequence and reply handling — that is what our outreach service covers, and there are free message tools if you would rather draft your own.

Key takeaways

• Build every note from four slots: relevance signal, reason for reaching out, low-commitment ask, and what you are not doing.
• The ask should be satisfied by clicking Accept. Asking for a call in the invite note converts the decision into a negotiation.
• Name what you will not do next; it answers the reader's actual hesitation in the fewest characters.
• A relevance signal that would fit a hundred other people is not a relevance signal.
• Test one slot at a time per segment and per sending profile, and track replies as well as acceptances.

Frequently asked questions

Should I send a note at all, or invite without one?

Send one when you have a genuine relevance signal, and consider going without when you do not. A blank invite is neutral; a generic note is worse than neutral, because it proves the sender had something to say and it was nothing. Some account types also limit how many notes you can send per month, so check your own account before designing around them.

How long should the note be?

Shorter than the limit. Use the room the four slots need and stop, rather than filling the space because it is there. A note that ends before the character count is a note that respects the reader's two seconds, and it will never truncate mid-sentence.

Is mentioning a mutual connection a good relevance signal?

Only if the connection is meaningful to the reader and you can say why it matters. Naming a shared contact you have never spoken to is a weak signal that invites an easy check. A specific, verifiable detail about the reader's own situation outperforms a borrowed one.

Can the same note run across several sending profiles?

The structure can; the wording should not be identical everywhere. Each profile has a different headline and background, and the note has to sound like it came from that person. Rewrite the reason slot per profile at minimum, since that is the clause tying the sender's identity to the approach.