What LinkedIn outreach templates actually get replies?
The ones that name a specific signal and ask for almost nothing. Four lines maximum: the signal stated plainly, why it matters to them, who you are in six words or fewer, then no ask or a micro ask. Wait one to three days after acceptance. Three follow ups over ten to fourteen days, then stop.
The structure, in four lines
Line one is the signal, stated plainly: the observable thing that made you write to this person this week. Line two is why it matters to them, not to you. Line three is who you are, in six words or fewer. Line four is either no ask at all, or a micro ask that costs one sentence to answer.
Four lines is a ceiling rather than a target. The message is competing with a meeting and a phone notification, and the only thing that earns a second sentence is the first one being specific enough to be about them. Length is a signal in itself: a long first message tells the recipient that answering it will be work, and they are usually right.
The same logic covers the connection note and the first message after acceptance. The note is shorter because the character limit makes it shorter, not because the structure changes.
Three worked examples
Hiring signal: "Saw you are hiring two support leads. That usually means ticket volume grew faster than the team did. I work on deflection for B2B SaaS. No ask, happy to send the teardown we did for a similar queue."
Role change: "Congrats on the move to [company]. The first ninety days is usually when the reporting stack gets questioned. I build [one thing, plainly]. Not pitching, curious what you inherited."
Their own words: "You said on [person]'s post that [problem] is the part nobody solves. That is the part we work on. Small team, four years at it. Worth a reply or not, either is fine." Notice what is missing from all three: a calendar link, a paragraph about us, a compliment, and an emoji.
Openers we do not send
These are not banned for taste. Each one is a recognisable pattern that tells the recipient the message is one of many, and recognition is the thing that kills a reply rate before the content is even read.
"Pick your brain", "hope this finds you well", "quick question", "huge fan", "love what you're doing", and any emoji in a first touch. The emoji rule surprises people: it is not about formality, it is that emoji in a first message performs familiarity that has not been earned yet.
The rule underneath the list: if a sentence would survive being pasted into a message to anybody else, delete it. That single test removes most openers before they are written, including ones not named here, and it is the same test that makes line one of the structure work.
Timing, which matters as much as the words
Wait one to three days after acceptance. A message that lands the instant someone accepts is the clearest possible tell that a machine was watching, and it arrives before the person has any memory of having accepted.
Then at most three follow ups spread over ten to fourteen days, and stop. Not a fourth. A sequence that runs past that point is harvesting negative signal: unanswered messages, the occasional report, and a growing group of people who now associate your name with being chased. The stop is not a courtesy either. People who never replied are the same population whose ignored messages feed the pattern that gets invitations restricted.
A reply ends the sequence immediately. That sounds obvious and is worth verifying in whatever tool you use, because the alternative is a scheduled follow up landing in the middle of a real conversation, which is the worst message in the whole sequence.
Test the blank invitation before you test the copy
Notes are not automatically better. Blank invitations sometimes match or beat noted ones, because a note gives the recipient something to disagree with while a blank invitation from a plausible profile in an adjacent field is a cheap yes.
So the first test on a new audience is not note A against note B. It is note against no note, on comparable halves of the same list, read at the same elapsed time. Only once that is settled is it worth spending a week on wording.
Keep the result attached to the audience that produced it. A blank invitation winning with founders in one market says nothing about senior operators in another, and the teams that get this wrong usually adopted a blanket no note policy off a single test.
What the tooling does with this
Drafts, not sends. The agent reads the profile, writes against this structure, and queues the message behind a human approval gate. Creating a sequence sends nothing at all; enrolling people queues drafts for review. A reply stops the remaining follow ups, and caps are enforced server side, so a long list does not become a heavy day.
Suppression memory prevents a second first touch to anyone already contacted, including people imported from your own history on day one. That is usually the first thing a new workspace catches, because most teams have contacted more of their list than they remember. It is also why the first week in a new workspace is quieter than expected, which is the correct outcome rather than a setup problem.
What is not here: any CRM, Zapier, HubSpot, Salesforce, webhook or sync tool. If your templates or your pipeline live in another system, they get pasted across, not synced. The same applies to approvals: the gate is a person in the conversation, not a review queue in some other product.
Banned openers and what to write instead
| Phrase | What it signals | Write instead |
|---|---|---|
| "Pick your brain" | An unpaid meeting with no agenda | Ask the actual question in the message |
| "Hope this finds you well" | A mail merge behind the message | Open with the signal |
| "Quick question" | It is never a quick question | Ask it, in one sentence |
| "Huge fan" | Flattery with nothing behind it | Quote the specific thing you read |
| "Love what you're doing" | Nothing. It fits any recipient | Name the thing, or say nothing |
| Any emoji in a first touch | Familiarity that is not earned yet | Plain sentences |
Questions people ask next
What should the first LinkedIn message after connecting say?
Name the signal that made you write, say why it matters to them, identify yourself in six words or fewer, then either ask nothing or ask something answerable in one sentence. Four lines at most.
How long should I wait after someone accepts?
One to three days. An instant message on acceptance reads as automation and arrives before the person remembers accepting.
How many follow ups is too many?
More than three. Cap it at three follow ups over ten to fourteen days, then stop, and make sure any reply ends the sequence immediately.
Should I put a calendar link in the first message?
No. A link asks for a commitment before the person has decided you are worth a reply. Earn the reply, then offer the time.
Why this page exists: LinkedBoost message structure and cadence policy, against 861 conversations of our own sending
LinkedBoost is the LinkedIn MCP server: your agent sources, drafts, sends and works the inbox, inside caps the server enforces rather than suggests.
Related answers
30 to 50% is healthy for cold outbound. We average 41% across 861 conversations over three years and five seats. Under 20% means targeting is broken, 20 to 30% means relevance is thin, 50 to 60% is a strong signal match, and over 60% usually means a warm audience.
Partnership outreach asks for shared upside rather than a purchase, so the qualifying question is overlap, not budget. The list is tiny, the research is heavier per name, the first message names what you can send them, and none of it forecasts like a sales pipeline. Volume tactics actively hurt it.