LinkedIn rewards one behaviour above all others: the save. People save posts they intend to use — instructions, checklists, decision rules. They like posts that made them feel something for two seconds and never think about them again.
That single fact explains most of the difference between a post that travels and a post that does not. Write something a reader can act on before their next meeting, and the mechanics take care of themselves.
The first two lines decide everything
LinkedIn truncates the post after roughly two lines. Everything after that is read only by people the first two lines convinced.
Three openings that fail:
- Context first. "Last week I was talking to a founder about pricing..." The reader was not there and does not care yet.
- Announcement framing. "Excited to share that..." This is the language of a press release, and it signals that nothing will be asked of the reader.
- Abstract noun. "Retention is one of the most important metrics in SaaS." True, unobjectionable, and impossible to disagree with — which is the problem.
One opening that works: state a claim the reader might resist.
"We killed the plan that made us the most money."
It creates a question in eight words. The reader stays to find out whether you are right, and by then you have their attention long enough to deliver the mechanism.
One idea per post
A post with three ideas is a post with none. The reader remembers the shape of three arguments and the content of none.
If you cannot summarise the post in one sentence, it is two posts. Splitting them usually doubles total reach, because each one can be picked up by a different part of the audience.
The test: read your post and ask what the reader should do differently. If the answer covers three areas of their work, you have written a newsletter, not a post.
The proof line
A claim without evidence is an opinion with better formatting. Every post needs exactly one piece of proof, and it should be specific enough to be checkable.
Three kinds work:
- A number that moved. Not "significant improvement" but the before and after.
- A named failure. "We shipped a config option for one customer and paid for it in documentation for three years."
- A mechanism. Not what happened, but why it happened. This is the most portable kind, because the reader can apply it to their own situation.
One piece of proof, not three. The second piece of proof is almost always the reason a post feels long.
The boundary line
State where your advice does not apply. One sentence.
"This holds for B2B with a sales team. Self-serve products have a different failure mode."
Two things happen when you add it. First, you pre-empt the objection that would otherwise be the top comment. Second, you signal that you have thought about the edges, which is the difference between advice and a take.
The ending
Do not summarise. A summary tells the reader the piece is over and they close the tab.
Instead, end with one of three things:
- An action with a verb. Two or three steps, each starting with a verb, each doable today without your help.
- A rule they can reuse. "If a plan's customers churn faster and file more tickets, its revenue is a loan." A rule travels; a conclusion stays in the feed.
- A specific question. Not "thoughts?" but a question with two plausible answers, so that replying is easier than scrolling.
Formatting that helps
- Paragraphs of one or two sentences. Whitespace is punctuation on this platform.
- No more than one link, at the end. Links in the body reduce reach and give the reader an exit.
- Lists only when the items are parallel. A three-item list where the items are sentences becomes a wall with bullets.
- No hashtag stacks. Three relevant ones at most, or none.
Three rules for LinkedIn specifically
- 150 to 250 words, one idea, one proof. Long enough to carry a lesson and its evidence, short enough to finish in a single scroll. A post that ends before the reader stops reading is doing its job; anything longer is a newsletter wearing a post's clothes.
- Lead with a statement someone could disagree with. A question invites the reader to answer silently and move on. A claim makes them stay long enough to check whether you are right — and that check is the read.
- One link, at the end, and never in the first line. Links reduce dwell time because some readers leave instead of reading. A link in the last line costs almost nothing; a link in the first line costs the post.
- Boring topics are usually just broad topics. "Hiring" is boring; "we withdrew an offer after the take-home and here is what it cost" is not. Narrow the instance, not the subject.