How to Write a LinkedIn Post: From a Note to a Draft

Most guides to writing a LinkedIn post start with the hook. By the time you're choosing an opening line, though, the post is already decided. Whether it will be worth reading depends on what you wrote down before that: the episode, the point you want to make, the details only you know.

This guide goes in the order the work actually happens. It starts with a note and ends with a draft you can publish, with an example all the way through.

Step 1. Start from a note, not a blank page

A blank page asks "what should I write about?", which is the hardest question to answer on the spot. A note answers it in advance. It's one line you wrote down when something happened at work, with one detail you'd otherwise forget.

Our example note:

Client asked again whether we do audits. Third time this month.

If you don't keep notes yet, start with the questions in 30 LinkedIn post ideas from your own work. Answer one, and you have a note.

Step 2. Pick one point

A note can lead to several posts. Pick one point and say it in a single sentence, the way you'd say it to a colleague. If you can't, the post isn't ready yet, and no amount of editing will fix it later.

For our note, a few possible points:

• We get the same question every week, so the website must be unclear.
• We started doing audits because clients kept asking.
• We don't do audits, and here's who we send people to.

Each is a different post. Choose the one you have the most material for. In the example, it's the second: three requests in a month made the team try a small paid audit.

Step 3. Answer the reader's questions before you write

This is the step most people skip, and it decides how specific the post will be. Before writing a single sentence of the draft, answer the questions a reader would ask. Write the answers as rough notes, in your own words:

• What exactly happened? Who asked, how, when.
• What did you do? The decision and the first action.
• What came of it? A number, a result, or "too early to tell", if that's the truth.
• What would you do differently? Optional, but often the most useful line.

For the example:

Three clients in October asked if we do audits. We said no to the first two. After the third, we offered a fixed-price two-day audit. Two of the three said yes. One of them turned into a three-month project. We should have priced it higher; we underestimated the reporting.

That's the post's material. Everything after this is arrangement. If the answers come out vague ("we improved our offering"), go back to the note. A draft written from vague answers will be vague too, whether you write it or an AI tool does.

Step 4. Write the first two lines

LinkedIn shows only the beginning of a post in the feed. The cutoff isn't documented and changes with the interface, but it's commonly measured at around 210 characters on desktop and closer to 140 on a phone (Eden). On a phone, that's one or two short sentences.

Those lines should tell the reader what the post is about and give them a reason to open it. The simplest way is to start with the episode or the result:

Three clients asked us for an audit in October. We had never offered one.

Compare that with typical openers:

Here's what nobody tells you about listening to clients.

The second line promises something and delivers nothing yet. Readers have seen thousands of openers like it, and many now read them as a sign of AI text (more on that in How to make AI writing sound human). The first one is plain, and it already contains the story.

Step 5. Write the body in the order things happened

The easiest structure for a post based on an episode is chronology: what happened, what you did, what came of it, what you'd change. You don't need a framework on top of it.

A few rules that keep the body readable:

• One thought per paragraph. Paragraphs of one to three sentences read well on a phone. A post where every line is its own paragraph reads like a poem, though, and it tires the eye.
• Keep the specifics from Step 3. "Two of the three said yes" carries more than "the response was great".
• Cut what doesn't serve the point. If a paragraph would fit a different post better, save it for that post.

LinkedIn allows up to 3,000 characters per post. That's a ceiling, not a goal. Length should follow the episode: a small story needs 100 words, a process with steps may need 300.

Step 6. End when the story ends

When the last point is made, stop. A summary that repeats the post ("So, listen to your clients!") adds nothing. The closing question ("What do you think?") works only if the post gives readers something specific to answer; otherwise it's a habit, and a widespread one.

A good last line often comes from the "what would you do differently" answer. It's concrete and honest, and readers remember it:

Next time we'll price the reporting separately. It took longer than the audit itself.

Step 7. Edit out loud

Read the draft out loud before you publish. You'll hear what you can't see on screen: a sentence you'd never say to a colleague, a phrase that sounds like a press release, a paragraph that repeats the previous one.

Then check it against the patterns that make a post read as generated: a "this isn't about X, it's about Y" pair, three short fragments in a row, a teaser line, a question the next line answers. The full list is in our guide to AI writing patterns.

The example, start to finish

From the note "Client asked again whether we do audits. Third time this month", through the steps above:

Three clients asked us for an audit in October. We had never offered one.

We said no to the first two. After the third request, we put together a fixed-price two-day audit and sent the offer to all three.

Two said yes. One of those audits turned into a three-month project, our biggest this quarter.

Looking back, those three questions were the clearest demand signal we got all year.

Next time we'll price the reporting separately. It took longer than the audit itself.

The post is under 90 words long and uses none of the usual formulas, and the reader still learns exactly what happened and what the team took from it.

Where AI helps, and where it doesn't

AI tools are good at arrangement: turning your answers from Step 3 into paragraphs, shortening a long draft, suggesting a plainer word. They can't do Steps 1 to 3 for you, because they don't know what happened in your work. Give a model only a topic, and it fills the gap with generic claims and familiar patterns.

So if you write with AI, the order matters. Write the note and the answers yourself, paste them in first, ask the model to keep your wording, and edit the result out loud.

Elynd helps you write posts like this from your own material - try it free

A short checklist

• The post has one point, and you can say it in one sentence.
• You answered what happened, what you did and what came of it before writing.
• The first two lines say what the post is about, without a teaser.
• The body keeps at least one number, name or quote from your answers.
• The ending is the last fact or the lesson, not a summary.
• You read it out loud and changed every line you wouldn't say.