# Updated LinkedIn Best Practices Guide

**Edition:** 2026 · version 1.0
**Last updated:** 29 September 2026
**Planned review:** annually, in September
**Scope:** organic content for personal profiles and company Pages
**Documentary basis:** Metricool newsletters received in Outlook, Metricool’s LinkedIn Study 2026 and official LinkedIn documentation.

## 1. The golden rule

LinkedIn should not be managed as a noticeboard. The strongest strategy combines:

- **People who start conversations:** specialists, spokespeople and employees bring experience, opinions and approachability.
- **A company Page that adds scale and credibility:** it publishes corporate assets, data, case studies and messages that are easy to share.
- **Useful content before promotional content:** every post should teach, clarify, compare, demonstrate or help someone decide.

Metricool’s study observed a 63% higher engagement rate on personal profiles than on company Pages once clicks were excluded to make the samples comparable. The difference also appears in the response type: people generate considerably more comments, while company Pages receive many more shares. This allows each voice to have a clear function: specialists and spokespeople interpret, offer opinions and start conversations; the company Page provides sources, data and institutional content others can redistribute. The practical conclusion is to coordinate the channels rather than choose one.

### Connecting employees’ voices with the brand

LinkedIn associates a person with a company when they add or edit their role in **Experience** and select the organisation’s official Page from the dropdown. Mentioning the company in every post is not necessary for that relationship to exist.

When the aim is to build authority through an employee’s knowledge, the recommended practice is:

1. The expert posts from their personal profile, retaining authorship, their own voice and the conversation with their network.
2. They may mention the Page with `@` when natural — for example, to identify the company, a project or a source — but a mention does not replace the employment association or guarantee an automatic transfer of authority.
3. The company Page amplifies the post through a repost with its own context, a comment or a complementary post. Reposting keeps the original authorship visible.
4. If the Page creates a post and mentions the employee, it remains a company post. That is appropriate for recognition or corporate communication, but is not equivalent to a piece published by the expert.

LinkedIn does not document a specific algorithmic signal that turns every employee post into brand authority. The relationship is built through the correct employment association, thematic consistency, interaction between both voices and transparent amplification. For longer articles, Pages with Career Pages features can also display employee-written articles in the team perspectives section.

## 2. What to publish

### High priority

1. **Original knowledge:** data, project lessons, frameworks, analysis and supported opinions.
2. **Carousels and multi-image posts:** in Metricool’s full report, carousels received **11 times more interactions than single images**. This is a finding from that sample, not a guarantee for every account. Use them to explain processes, compare options, summarise studies and turn complex topics into sequences.
3. **People-led case studies and experiences:** problem, decision, outcome and learning, without turning the story into an advertisement.
4. **In-depth articles:** for evergreen topics, research, methodologies or questions needing depth.
5. **Video when the message justifies it:** demonstrations, interviews, visual explanations or an expert voice. It should not be the default; it was widely used in Metricool’s sample but did not lead performance.

### Recommended editorial mix

- 40% practical and educational knowledge.
- 25% cases, results and lessons.
- 20% human perspective: teams, experts and culture demonstrated through facts.
- 15% product, offers or direct conversion.

This mix is a starting point, not an algorithmic rule. Adapt it to your objectives and data.

### Originality and authority

Sharing others’ content can help sustain a conversation but should not form the core of a strategy. Drawing on external and internal analyses of visibility in AI search, LinkedIn states that 95% of citations of LinkedIn content come from original posts rather than reposts. This figure must be interpreted within that specific scope — retrieval by AI systems — but it reinforces a broader recommendation: original analysis, experience and frameworks build more authority than redistribution alone.

When sharing an external source, add an interpretation: what it means, why it matters, how it applies or where you disagree. Demonstrated expertise matters more than follower count. People with first-hand knowledge should participate, even when they do not have the largest audiences.

## 3. Writing a post

### Basic structure

1. **A clear, specific first line:** introduce the topic or conclusion. LinkedIn notes that this line often influences the post URL, so include the main keywords and avoid starting with hashtags.
2. **One central idea:** a post should not try to solve five topics.
3. **Scannable development:** short paragraphs, lists where helpful and concrete examples.
4. **Demonstrable value:** a figure, argument, example, procedure or lesson the reader can use.
5. **A natural closing action:** a conclusion, next step or relevant question.

To support human reading and visibility in AI search, LinkedIn recommends posts of around **200–300 words**, with clear questions and answers, simple lists and educational value ahead of promotion.

### Questions and invitations to comment

Metricool observed 77% more comments on posts with questions and 80% more when there was an explicit invitation to comment. This does not validate every engagement formula: the question should develop the topic and invite considered answers. Avoid generic endings such as “What do you think?” when you can ask about a specific experience, decision or criterion.

### Hashtags

- Use **1–5**, always relevant.
- Prioritise topics the audience recognises over invented tags with no community.
- Do not open a post with them.
- Do not rely on hashtags as the main driver of reach.

In Metricool’s study, posts with at least one hashtag performed above average, but interactions declined as their number increased. Their value appears to lie in classifying content and helping interested audiences find it, rather than accumulating tags.

## 4. Formats: decisions, not formulas

### Carousel or document

Use it for a sequence: steps, comparisons, a checklist, a process, before/after or report findings. The format lets an idea develop gradually and asks the reader to take a small action to continue. However, it does not automatically turn weak content into good content: each page must provide a unit of information and a reason to continue, while the cover should promise a specific benefit.

### Multiple images

Suitable for events, visual cases, journeys, teams or stories in stages. Metricool observed the highest engagement rate for this format on personal profiles. It is particularly useful when there is a visual sequence but no need for a designed document or carousel.

### Text

Works when the idea, experience or argument is the main asset. Avoid dense blocks; write for quick reading without impoverishing the content.

### Video

Include subtitles, a quick visual or verbal opening and a duration proportionate to the value. Metricool suggests pieces under a minute as a reference, but retention and clarity should guide the decision rather than a fixed duration. Video was among the most used formats in the sample, but did not lead performance. The most produced or eye-catching format is not necessarily the most effective: choose it for its ability to explain the content.

### Article or newsletter

LinkedIn recommends **800–1,200-word** articles to develop topical authority. An effective structure includes a question-based title, opening summary, subheadings, examples, definitions and next steps. Articles and posts play complementary roles: an article is a deep, stable source; posts distribute its ideas, test angles and generate conversation. One research project can become a main article and three to five posts, each focused on one finding.

## 5. External links

There is insufficient evidence to claim that LinkedIn always penalises links.

In Metricool’s sample:

- Company Pages with links received more impressions and interactions than their average.
- On personal profiles, posts with links performed worse on impressions and interactions.

These are correlations, not causal guarantees. Recommended practice is to:

- Include a link when essential to complete the value: a report, registration, resource, case study or demo.
- Provide enough value within the post; the link should not be its only content.
- Measure company Pages and personal profiles separately.
- Compare equivalent posts over several weeks before changing strategy.

## 6. Frequency, timing and conversation

- Start with **2–3 quality posts per week**, increasing only if you can maintain usefulness and consistency.
- For an authority strategy, LinkedIn suggests combining an in-depth article with several derived posts.
- Metricool observed that around half of a post’s lifetime impressions arrive within its first 48 hours. Roughly 40% of interactions on personal profiles and 35% on company Pages happen on the first day. Publish when the team can respond, add information and moderate the conversation, rather than only when a tool indicates an “optimal time”.
- The 09:00–12:00 slot was the most used in Metricool’s sample, but this measures posting volume; it does not show that it is best for every audience.
- Choose timing using your own analytics and your real availability to participate after posting.

## 7. Recommended editorial system

### Monthly

1. Choose two or three topics on which the brand wants to build authority.
2. Publish an in-depth piece for each topic: an article, report, interview or guide.
3. Divide each piece into three to five posts, each with a distinct idea.
4. Assign voices: company Page, internal experts and relevant employees.

### Weekly

1. Review the three best and three worst posts.
2. Analyse topic, format, author, objective and presence of a link separately.
3. Identify real questions in comments to inform new content.
4. Schedule the following week while retaining room for current events and conversation.

AI can speed up analytics extraction and comparison, generate variants or help structure a first draft, but it should not decide independently what to publish. Conclusions need review, and final content needs judgement, experience, verifiable examples and a recognisable voice.

This care also responds to changing expectations. In Sprout Social’s global Q1 2026 survey, 66% of consumers said they were becoming more selective about the content they consume, and 83% said they encountered low-quality AI content at least occasionally. These are cross-platform findings, not exclusive to LinkedIn, but they help explain why production efficiency should not be confused with relevance or credibility. Hootsuite reaches a compatible conclusion: AI is already part of the workflow, while human authenticity acts as a differentiator.

### Build series, not just isolated posts

When a topic is strategic, turn it into a recognisable series: a monthly analysis, frequently asked question, expert conversation or regular project lesson. Continuity enables depth, builds anticipation and associates recognisable voices with an area of expertise. A series can retain a name, approach and frequency without always repeating the same format.

Sprout Social found audience interaction and original content series among the main expectations of brands. As a cross-platform reference, this should be used as an editorial principle and then validated against your own audience’s behaviour.

## 8. What to measure

### Main metric by objective

- **Authority:** saves, sends, qualified comments, profile views and citations or mentions.
- **Reach:** members reached, impressions outside the network and follower growth.
- **Traffic:** link visits and the quality of the subsequent session.
- **Conversion:** registrations, leads, meetings or attributable actions.
- **Community:** recurrence and quality of conversation, beyond reaction counts.

### Minimum record for each post

Record the metrics LinkedIn provides for each account and post type; some will not be available on all profiles, Pages or formats. Keep the same metric selection when comparing results over time.

- Date, time, author and account.
- Topic and objective.
- Format and presence of a link.
- Impressions, reach and percentage outside the network.
- Reactions, comments, shares, saves and sends.
- Followers and profile visits generated.
- Link visits and conversions where applicable.

### Engagement rate

For company Pages, LinkedIn natively defines engagement rate as:

> **Interactions ÷ impressions × 100**

Interactions include clicks, reactions, comments and shares. This is the definition used by LinkedIn’s own analytics, rather than simply a metric proposed by Sprout Social.

This rate is not necessarily lower than a follower-based rate. If impressions exceed follower count, the result will be lower; if only a fraction of followers saw the post, it may be higher. The formulas answer different questions:

- **Interactions over impressions:** measures content’s ability to prompt action among actual exposures. This should be the main rate. To compare posts or use external references, ensure both the included interactions and denominator are equivalent.
- **Interactions over followers:** can remain a complementary internal metric for how a post mobilises the total follower base, but must be clearly labelled and not compared with impression-based benchmarks.

Alongside the rate, distinguish visible interactions from clicks and relate results to the content objective. A post can have a high rate and generate no visits, leads or relevant commercial conversations.

LinkedIn cautions that several metrics are estimates. Analyse trends and consistent comparisons; avoid treating one isolated figure as an absolute truth.

Performance is not limited to the signals visible below a post. Metricool observed that while likes, comments and shares fell, clicks and total engagement rose. Opening a link, swiping through a carousel, saving or sending content can also express interest. A piece should therefore not be judged solely by its public reactions.

## 9. Practices to avoid

- Publishing corporate news without explaining why it matters to the reader.
- Copying the same message across the company Page and several personal profiles.
- Automating comments or forcing artificial interaction.
- Sacrificing clarity for an exaggerated hook.
- Starting with a line of hashtags.
- Using too many or irrelevant hashtags.
- Treating a winning format as a permanent guarantee.
- Measuring success only through likes or impressions.
- Publishing with AI without verification, first-hand experience or human editing.
- Turning study correlations into universal algorithmic rules.

## 10. Pre-publication checklist

- [ ] Is there one clear central idea?
- [ ] Does the first line explain the topic and use the right words?
- [ ] Does the reader gain value without clicking?
- [ ] Does the format help explain the content?
- [ ] Are figures and claims verified and dated?
- [ ] Does the voice fit the person or brand publishing?
- [ ] Is the closing question specific and relevant?
- [ ] Are hashtags few and relevant?
- [ ] Has the post’s main metric been defined?
- [ ] Can someone respond and moderate during the first few hours?

## Methodological note

Metricool’s study analysed 673,658 posts from 63,108 accounts connected to its platform, comparing January–February 2025 with January–February 2026. Its findings help formulate hypotheses, but come from Metricool-connected accounts, cover a specific period and cannot establish causation. Validate each recommendation against your audience, sector, account size and own data.

Different Metricool materials summarised carousel performance using different figures — 17 times and 11 times more interactions than images. This guide adopts **11×** because it is the figure in the full report. It is also the more conservative figure, but that alone does not make it more rigorous. The discrepancy is recorded, and the outcome is not presented as a universal rule.

## Sources consulted

### Metricool newsletters in Outlook

- **[STUDY] What’s working on LinkedIn in 2026**, 14 April 2026.
- **[PLAYBOOK] Why 93% of LinkedIn company pages didn't grow last year**, 8 June 2026.
- **Your new superpower: AI + your social media**, 17 September 2026.

Duplicate copies, automatic connection notices and emails mentioning LinkedIn solely to add a certificate to a profile were excluded as an editorial basis.

### Public sources

- [Metricool LinkedIn Study 2026](https://metricool.com/wp-content/uploads/Linkedin-Study-2026-EN.pdf)
- [Maximising post visibility in AI search](https://www.linkedin.com/business/marketing/blog/ai-search/how-to-maximize-ai-visibility-for-your-linkedin-posts)
- [LinkedIn’s guide to visibility in AI search](https://www.linkedin.com/business/marketing/blog/ai-search/how-b2b-marketers-can-dominate-ai-search-on-linkedin-2026)
- [How LinkedIn associates employees with a Page](https://www.linkedin.com/help/linkedin/answer/a546860)
- [Reposting a person’s content from a Page](https://www.linkedin.com/help/linkedin/answer/a551424)
- [LinkedIn Page content analytics](https://www.linkedin.com/help/lms/answer/a564051/content-analytics-for-your-linkedin)
- [Available post analytics](https://www.linkedin.com/help/linkedin/answer/a523040/)
- [LinkedIn content recommendations](https://www.linkedin.com/help/linkedin/answer/a7142408)
- [Hootsuite Social Media Trends 2026](https://www.hootsuite.com/research/social-trends)
- [Sprout Social 2026 social media trends](https://sproutsocial.com/insights/social-media-trends/)
- [Sprout Social statistics library](https://sproutsocial.com/insights/stats/)