Skip to content
Badgr Badgr Help

Connection messages that get accepted

The connection request note is the most-debated element of LinkedIn outreach. Some operators swear by sending notes; others insist no-note requests get higher acceptance. Both can be right, depending on the audience.

This article is the version of “what we’d tell a friend starting outreach today” — pragmatic, no formula worship.

The note vs no-note question

The trade-off:

  • No note → higher acceptance rate, lower reply rate from accepted connections. People accept out of curiosity. They then ghost.
  • Note → lower acceptance rate (usually 5–15% lower), higher reply rate from those who do accept. The note pre-qualifies; people who accept already know what you want.

For most B2B outreach to senior people, the note version wins on meeting-booked rate — the only metric that ultimately matters.

For volume-driven outreach to a broad audience (consumer-ish), no-note can be the right call.

What works in the note

Short. Specific. Doesn’t sound like a sales pitch.

The 200-character limit forces this anyway. Use it.

A formula that consistently works:

[Observation about them], [reason this matters to me], [low-friction ask].

Example:

Saw your post on agency operations — we’ve been writing about per-account safety for outbound and your line on “ops debt compounding” hit. Open to compare notes?

This works because:

  • The observation proves you actually saw something (not a copy-paste)
  • The “reason this matters to me” gives them a reason to engage
  • The ask is low-stakes (“compare notes” not “30-minute demo”)

What doesn’t work

The patterns we see fail consistently:

“Hey {firstName}, hope you’re well!” This signals “automation” even if the rest is hand-written. Skip the formulaic opener. Get to the substance.

“I help {industry} companies scale by…” This is a sales pitch in a connection note. It tells the prospect exactly what’s coming and they accept (if at all) braced for the inevitable pitch. Acceptance is fine; replies are awful.

“Hi, I love your work, let’s connect!” Generic praise reads as suspicious. If you actually loved their work, you’d reference something specific.

Asking for time in the connection note. “Free for a 15-minute call next week?” in the request itself is the fastest way to get ignored.

Auto-personalisation that’s slightly wrong. “Saw you’ve been at {company} for {years}+ years” when {years} resolves to “0.3” or “12.7” looks worse than no personalisation at all.

Templating without sounding templated

Most outreach is templated. That’s fine. The trick is making the template carry one piece of real specificity.

The pattern: write the connection note as a template with one mandatory variable that you fill in manually for each prospect.

Example template:

[SPECIFIC OBSERVATION], we’ve been thinking about [TOPIC] from a [YOUR_ANGLE] angle. Open to compare notes?

You write the [SPECIFIC OBSERVATION] yourself per prospect (15 seconds of looking at their profile or recent post). The rest of the template is shared. Acceptance rate jumps materially compared to pure variable substitution.

If 15 seconds per prospect is too much, you’re sending to too many people. Lower the volume and raise the acceptance rate.

Follow-up messages (after they accept)

Once accepted, the first real message comes 24–72 hours later. Keep it:

  • Conversational. Like a message you’d send a stranger at an event you both attended.
  • Specific. Reference something — the connection-request note context, something on their profile, a mutual connection.
  • Low-friction. Don’t ask for a meeting in message #1. Ask a question. Send a useful link. Open a conversation.

The meeting ask comes in message #2 or #3, after the conversation has earned the right to it.

What about message length?

Short messages perform better. Empirically:

  • Under 200 characters: high reply rate
  • 200–500 characters: solid
  • 500–1000: noticeable drop
  • 1000+: people scroll past

The longest messages I’ve seen perform well are ~600 characters and include a tactical question the prospect actually wants to answer. Otherwise, shorter is better.

A/B testing without overthinking it

Run two versions of a sequence in parallel — same audience, different opening note. After 50–100 invitations each, compare acceptance rate AND reply rate (not just one). Pick the winner. Don’t bother with 3-way tests at low volume; you’ll never get statistical signal.

After you’ve tested 4–5 variations across a couple of audiences, you’ll have a feel for what your audience responds to. The feel is more useful than any “proven template” you’ll find online.