Best practices: writing LinkedIn messages that get replies

A LinkedIn message gets a reply when it's short, clearly written for the person reading it and easy to answer. This guide turns that into habits you can apply with Swimlane's templates, variables, AI suggestions and reminders.

1. Write for one person, even in a template

A template saves time, but your contact should never feel it. Use variables for the facts, then add one line that only makes sense for this person.

Swimlane recognises exactly seven variables: {{firstName}}, {{lastName}}, {{fullName}}, {{company}}, {{title}}, {{headline}} and {{location}}. Keep three rules in mind:

Anything else goes out as typed. {{city}} or {{industry}} reaches your contact with the braces. Use the variable buttons under the template field rather than typing from memory.
An empty value disappears. A contact with no company on record turns "How is the team at {{company}} handling this?" into "How is the team at handling this?". Prefer sentences that survive a blank, or use a static template only on contacts whose record you've checked.
Variables typed straight into the inbox message field aren't replaced. In the inbox, use them through a template.

Example of a static template that stays natural:

Hi {{firstName}}, thanks for connecting. I saw you're {{title}}. Are you the one handling outreach tooling, or is that someone else on your team?

Before sending, read the result in the message field: a job title in capitals or a very long headline is easy to fix by hand.

2. Keep it short, with one question

First message: two or three sentences. The AI icebreaker deliberately writes a single sentence. Use it as a base and add your reason for writing.
Replies: match the length of what you received. The AI draft sticks to one or two sentences for that reason.
One question per message. A precise question is easier to answer than "Let me know what you think".
Connection request notes are shorter still: LinkedIn allows 200 characters on a free account and 300 on Premium.

3. Give a reason, not a pitch

Your first message should say why you're writing to this person now: a post they published, a role they just started, a problem their headline mentions. The pitch can wait until they reply.

Use AI templates when the idea stays the same but the wording must fit each person. Keep static templates for messages where every word counts, such as an offer or a meeting link.

4. Share one link, and watch the click

Links starting with https:// are tracked automatically. One link per message tells you exactly what interested the contact. When they click, [First name] clicked your link appears in their Activity tab. That's a natural moment to follow up. Swimlane doesn't track whether a message was opened, only clicks.

5. Follow up on purpose, then stop

Silence is rarely a no, but three follow-ups in a row is too many. A simple rhythm:

First follow-up after three to five working days, adding something new: a resource, a short example, a simpler question.
Second follow-up about a week later, short and easy to answer.
Then stop. Set a reminder a few weeks out, and come back only with a real reason.

Set the reminder the moment you send the message: Info → Quick actions → Create a new reminder, or Get a reminder in Swimlane Analysis, which suggests a date. Reminders don't send notifications, so check the Reminders page every day (see Best practices: keeping your inbox under control every day).

A poke can replace a follow-up once, when the thread has gone quiet. Don't repeat it.

6. Let the AI prepare, but keep the last word

Feed it. The AI draft reads your notes, your pending reminders and Settings → Personalization. A note like "Wants to decide before the summer" makes the next draft far more relevant.
Edit before sending. Drafts you send unchanged don't teach Swimlane your style. Your edits do.
Check the language and register. The icebreaker is written in your Swimlane language, and in French it uses tu. Adjust when the relationship calls for something different.

7. When they reply, take over

Campaigns don't stop on their own when someone replies. If a contact is in a sequence, the next steps still run unless your sequence branches on the Replied to message condition. Build that branch into every campaign.
Reply by hand, quickly. Manual replies from the inbox go out immediately, even outside your virtual office hours.
If they say no, archive the contact. Their conversation leaves your inbox and every campaign stops reaching them.

Quick checklist before you click Send

Would this message still make sense if you swapped the name for someone else's? If yes, add a personal line.
Is there exactly one question?
No leftover braces, no empty gap where a variable was.
At most one link, starting with https://.
A reminder is set if you expect a reply.