IIDigital MarketingGuide
Writing LinkedIn Posts That Professionals Actually Read
Many businesses’ LinkedIn posts are announcements nobody reads or stories many like and nobody needs. This guide shows how to structure a post that professionals read to the end.
Key takeaways
- In the feed, readers decide in seconds whether to stop. Once they do, professionals read a post thoroughly. The elaboration likelihood model explains why a post needs both.
- An analysis of 800 LinkedIn posts found more reactions to posts with links, calls to action and emotion, not to factual, informative ones. But reactions are not the same as conversations.
- The four-paragraph post follows a fixed structure: observation, interpretation, evidence, question. It fits within the 3,000 characters LinkedIn allows for a post.
- Measure by comments from the right people and the conversations that result. Reach is only an intermediate value.
Why do professionals finish so few LinkedIn posts?
An owner decides to become more visible on LinkedIn. The first posts are announcements: new hall, new team member, trade fair participation. A few thumbs-up come from colleagues and acquaintances. Then he tries advice from the internet, a personal story with many paragraphs and a lesson at the end. Reactions increase. Only the purchasing managers and technical directors he wanted to reach react to neither.
The article LinkedIn social selling notes that on LinkedIn, selling happens in the conversation, not in the post. Posts still matter, because they show how someone thinks about a specialist topic before you talk to them. The question this article sets out to answer: How do you write LinkedIn posts that professionals in your target group actually read and that lead to conversations? To answer it, we first clarify a model from persuasion research, then what an analysis of LinkedIn posts shows, and derive structure, topics and measurement from it.
A post that the target group does not read to the end has no reach. It only has views.
The thinking model: two routes of persuasion in the feed
In the 1980s the psychologists Richard Petty and John Cacioppo developed the elaboration likelihood model. Elaboration here means how thoroughly a person thinks about a message. The model describes two routes by which people are persuaded. The article Persuading without tricks presents it in detail: on the central route, the recipient examines arguments thoroughly; on the peripheral route, they follow cues such as expert status or scarcity.
For LinkedIn, the model is particularly revealing because a post passes through both routes in sequence. In the feed, the reader scrolls quickly and thinks little. Here cues work: a known name, an image, the first words. If they stop and expand the text, they switch to the central route for a topic that affects them professionally. Now they check whether the arguments hold, whether the figures are plausible and whether the author knows what they are talking about.
In the feed: peripheral route
- Little attention, fast scrolling
- What works: name, image, the first lines
- Decision: stop or move on
- Risk: sensationalism that disappoints later
Expanded: central route
- High attention when professionally affected
- What works: arguments, evidence, experience
- Decision: is the author competent?
- Risk: empty claims that stand out
This leads to the basic rule of this guide: the first lines must make the reader stop, but with a promise that the rest of the post keeps on the central route. An opening that promises more than the text delivers wins the click and loses the reader.
What an analysis of LinkedIn posts shows
Yuen, Lee, Nguyen and colleagues (2024) analyzed 800 posts on the LinkedIn pages of the ten largest ship management companies according to an industry ranking. For each post they recorded features such as length, links, calls to action, images and videos, narrative perspective and the type of content, and related them to the reaction rate from likes and comments.
Nine features were associated with more reactions, in descending strength: links, company names in the text, calls to action, the length of the text, visible physical objects such as ships or buildings, content with a social dimension, emotional content such as congratulations, the first-person perspective and emojis. Hashtags, whether a post contained an image or video, and content that informed about services or factually about the company and industry showed no demonstrable influence. Just over half of the posts were informative.
Source: Yuen, Lee, Nguyen et al. (2024)
Two readings: reactions as the yardstick or as a distraction?
The first reading takes the result as instructions. Whoever wants more reactions writes longer, links, calls for action and shows emotion. The second reading asks what reactions stand for. Congratulations on a company anniversary get many likes from colleagues and partners but rarely a comment from a buyer with a problem. That factual information triggered hardly any reactions does not mean it was not read.
I consider the second reading more important, without dismissing the first. The study measures reactions on company pages in a single industry, not conversations or orders. Its pointers on form are useful: length does no harm, a link and a clear call to action do no harm, and the first-person perspective works. For content, the yardstick from social selling applies: measure by conversations held with the right contacts.
The four-paragraph post: observation, interpretation, evidence, question
From the model and the findings, I derive a structure, the four-paragraph post. It leads the reader from the peripheral route in the feed to the central route in the text and ends with an invitation to a conversation. LinkedIn allows up to 3,000 characters per post. A four-paragraph post usually needs 800 to 1,500.
1. Observation
The first one or two sentences describe something concrete from practice that the reader knows: a figure, a recurring customer question, a common mistake. No riddle, no exaggeration.
2. Interpretation
Two to four sentences on why this is so. Here the author shows how they think, and here the reader switches to the central route.
3. Evidence
An example from a project without confidential details, your own calculation or a named source. Evidence distinguishes an expert post from an opinion.
4. Question or conclusion
A genuine question to professionals that invites a comment, or a sober conclusion with a pointer to where it is covered in more detail.
| Part | Text |
|---|---|
| Observation | In almost every hall we inspected this year, the compressor was also running at the weekend. |
| Interpretation | The reason is rarely a defect but leaks. When production is idle, they do not hold the pressure, so the compressor starts. Because nobody is there, nobody notices. |
| Evidence | In a hall with 40 employees, this added up to around 9,000 kilowatt hours a year. We found it with a two-hour measurement on a Saturday. |
| Question | For those in maintenance: do you measure consumption at the weekend, or does it only show up on the electricity bill? |
Worked example
Topics that professionals read
The best topics for LinkedIn posts are the same as for articles on the website: questions customers ask repeatedly, mistakes that keep happening, figures from your own work, assessments of industry developments. The difference lies in scope. A LinkedIn post does not answer the whole question but one part of it, and refers to the detailed version if needed.
Worked example with freely chosen numbers. Counted are conversations with the right contacts that refer to a post. Company news often gets reactions but rarely conversations.
- Observations from projects, without naming customers who have not agreed.
- Your own figures and calculations showing how a problem translates into euros, hours or days of downtime.
- Mistakes and their causes, preferably your own. Whoever says openly what they have learned seems more credible than someone who only shows successes.
- Assessments of developments, such as a regulation or technology, with a clear position of your own and its limits.
- Answers to questions from comments, which become posts of their own.
An article on the website usually provides material for several LinkedIn posts. How one text becomes a series of short posts is described with the core-and-forms model in the article Repurposing content.
Form and style: what helps in the feed and what harms
The form of a LinkedIn post follows the same rules as any good text, with a few peculiarities of the feed. Only the first lines are visible without expanding. Many people read on a smartphone. And the reader often knows the author only from their profile.
- Short paragraphs. Two to three sentences per paragraph, with a blank line between them. One line per sentence, as many guides recommend, chops up the thought.
- First-person perspective. The author speaks as a person. The analysis by Yuen and colleagues found a positive association for this perspective.
- The readers’ technical terms, not in-house ones. What applies to website copy is described in the article B2B website copy.
- At most one link, at the end, with a sentence on what the reader will find there.
- Few or no hashtags. In the analysis they had no demonstrable influence.
- An image only if it shows something. A measurement curve, a photo from a job, a table. Symbolic images add nothing.
Texts from an AI tool can help with structuring. What they do not provide are the own observation and own evidence that make an expert post. Where AI helps in marketing and where it costs credibility is assessed in the article AI tools in marketing.
Rhythm and measurement: how to tell whether it works
The article The owner’s LinkedIn profile reckons with one short post of your own per week, and about half an hour for it, for a lively profile. More important than the number is regularity. One post a week for a year shows more expertise than ten posts in one month followed by silence.
- Impressions in the feed
- Expanded and read12% of previous stage
- Comments from the right contacts1% of previous stage
- Direct messages33% of previous stage
- Conversations40% of previous stage
Worked example with freely chosen numbers. The decisive stages are the last three. They can be counted without a tool if every message and every conversation referring to a post is noted.
Three figures per quarter are enough for the evaluation: comments from people in the target group, messages that refer to a post, and conversations that resulted. How posts, comments and outreach work together is described by the 5-10-5 routine in the article LinkedIn social selling. Whoever also maintains a company page for the business follows the rule from the article LinkedIn company page or personal profile: the person speaks, the page vouches.
What can be generalized
In the short term, LinkedIn rewards what brings many reactions. For a business that sells to professionals, this is a temptation. Whoever gives in gets attention from people who do not buy and loses the attention of those who examine. Whoever writes posts that hold up on the central route, by contrast, slowly builds something that cannot be bought: the reputation of having something to say on a topic. This is the same stance that applies throughout this website: evidence instead of assertion.
The limits are clear. The analysis by Yuen and colleagues concerns company pages in a single industry and measures reactions, not conversations. The elaboration likelihood model describes persuasion in general, not LinkedIn in particular. How the platform distributes posts changes constantly and is not public. What is robust is the conclusion from the model: a post needs a reason to stop and a reason to believe, and for professionals the second is the more important.
Conclusion and a first step
The opening question was LinkedIn posts must be written so that professionals read them and conversations result. The answer: with a four-paragraph post of observation, interpretation, evidence and question, with topics from your own practice, in the first person and without effects that promise more than the text delivers. Measure by comments from the right people and by conversations, not by reactions. The elaboration likelihood model explains why: in the feed, the first impression decides whether someone stops; in the text, arguments and evidence decide whether they trust.
- Note five customer questions or observations from recent weeks as topics.
- Write the first post following the four-paragraph structure, the first line last.
- Check that every post contains evidence: example, figure or source.
- Put a fixed weekday for the post in the calendar.
- Set up a simple list for comments, messages and conversations referring to posts.
- After one quarter, evaluate which type of post brought conversations.
A good LinkedIn post is not measured by how many people like it. It is measured by who calls afterwards.
Frequently asked questions
How long should a LinkedIn post be?
LinkedIn allows up to 3,000 characters. For an expert post with observation, interpretation, evidence and question, 800 to 1,500 characters are usually enough. An analysis of 800 posts found that longer texts received not fewer but rather more reactions.
How often should I post on LinkedIn?
For most owners, one post of their own per week is enough, supplemented by comments under posts from the target group. Regularity over months shows more expertise than many posts in a short time.
Which topics work on LinkedIn in B2B?
For conversations with professionals, observations from projects, your own figures and calculations, mistakes and their causes, and assessments of industry developments are suitable. Company news often gets reactions but rarely leads to conversations.
Do hashtags help on LinkedIn?
In an analysis of 800 LinkedIn posts, hashtags had no demonstrable influence on reactions. A few fitting hashtags do no harm; many tend to distract.
How can I tell whether my LinkedIn posts are working?
By comments from people in the target group, by messages referring to a post and by the conversations that result. Reach and likes are intermediate values that say little about orders.
Sources
- Petty, R. E.; Cacioppo, J. T. (1986): The Elaboration Likelihood Model of Persuasion. Advances in Experimental Social Psychology, 19, 123-205.
- Yuen, K. F.; Lee, J. D.; Nguyen, C. T. et al. (2024): Leveraging Social Media for Stakeholder Engagement: A Case on the Ship Management Industry. Information, 15(11), 693.
- LinkedIn Help: Share posts on LinkedIn (character limit for posts).
