linkedin-article

name	linkedin-article
description	Writes LinkedIn-formatted thought leadership articles with professional credibility markers, structured arguments, and platform-specific formatting. Use when the user wants to publish a long-form article on LinkedIn, write thought leadership content for a professional audience, or create LinkedIn Pulse content. Do NOT use for short LinkedIn posts (use `linkedin-post`), general blog posts (use `blog-post-writing`), or Twitter threads (use `twitter-thread`).
license	Apache-2.0
metadata	{"author":"foundry-skills","version":"1.0.0","tags":"social-media writing content-marketing","category":"writing","subcategory":"content-marketing","depends":"","disclaimer":"none","difficulty":"intermediate"}

LinkedIn Article

When to Use

Use this skill when:

• The user wants to publish a long-form article through LinkedIn's native article editor (800--2,000 words), which displays on their profile, is indexed by search engines, and distributes to followers as a notification
• The user wants to establish or reinforce thought leadership on a specific professional topic -- building a durable, searchable body of work rather than feed-based visibility
• The user wants to reach a specific professional audience segment: industry peers, hiring decision-makers, potential clients, speaking conference organizers, or media contacts who research executives before outreach
• The user needs to process a professional experience -- a failed project, a career transition, a contrarian industry observation -- into structured argument with evidence
• The user is a subject matter expert who wants to translate technical or operational depth into digestible professional insight for a broader audience
• The user has a research finding, survey result, or internal data point they want to contextualize and publish without going through a trade publication pitch cycle
• The user wants to drive inbound: LinkedIn articles rank in Google search and persist on the profile indefinitely, unlike feed posts which decay to near-zero reach in 48 hours

Do NOT use when:

• The user wants a short LinkedIn feed post (under 300 words or a text-first scrollable update) -- use linkedin-post instead
• The user wants a general blog post intended for their personal or company website -- use blog-post-writing instead; LinkedIn articles have platform-specific structural requirements that differ from SEO-optimized blog content
• The user wants a Twitter/X thread optimized for rapid-fire sequential reading -- use twitter-thread instead
• The user wants a formal opinion piece or byline for a trade publication, business journal, or magazine -- use opinion-piece instead; those require different citation standards, editorial framing, and length conventions
• The user wants a press release or company announcement -- LinkedIn articles are first-person professional insight, not corporate communications
• The user wants to republish an existing article without adaptation -- flag that direct republication without LinkedIn-specific restructuring will underperform; offer to adapt instead
• The user wants a carousel post, video script, or document post (PDF attachment) -- those are separate LinkedIn content formats with different production requirements

---

Process

Step 1: Establish Professional Context and Credibility Assets

Before writing a single word, extract the raw material that makes a LinkedIn article credible. Generic professional observations fail on LinkedIn because readers can detect immediately whether the author has lived the experience.

• Ask for the user's current role, company stage (startup, scale-up, enterprise), industry vertical, and years in this specific function -- these details calibrate voice and determine what evidence is available
• Ask for the specific professional experience, decision, failure, or observation that motivated this article -- the more specific, the better: "I managed three platform migrations" beats "I have experience in cloud infrastructure"
• Ask what outcome or change they observed: numbers, timeframes, before/after states, team sizes, budget ranges, or any quantifiable result -- even approximate numbers ("roughly 30% faster," "from 14 weeks to 9") are more credible than no numbers
• Ask who the target reader is and what their role, seniority, and professional pain points are -- a Director of Engineering reads differently than a first-year analyst or a CFO
• Ask what the reader should believe, do, or reconsider after reading -- this becomes the article's thesis, which must be stated within the first 100 words
• Ask what the author would be willing to admit went wrong, what assumption they held that turned out to be incorrect, or where their advice has limits -- these moments of intellectual honesty are the single most trust-building element in a LinkedIn article
• If the user says "I am not sure what angle to take," use this diagnostic question: "What is one thing you believe about [topic] that a majority of your professional peers would push back on?" The answer becomes the thesis

Step 2: Identify the Article's Argumentation Type

LinkedIn articles fall into distinct argumentative structures, and the chosen structure should drive the entire outline. Match the user's content to one of these five proven patterns:

• The Reversal: The author held a widely shared professional belief, encountered evidence that contradicted it, and changed position. Structure: old belief → disconfirming experience → new position → practical implication. Example: "Why I Stopped Hiring for 'Culture Fit' After It Cost Me Three Strong Engineers."
• The Mechanism Expose: The author observed a professional phenomenon and traced it to a non-obvious root cause. Structure: observable symptom → common misdiagnosis → actual mechanism → corrective action. Most effective for operational leaders with process depth.
• The Pattern Report: The author has observed the same dynamic across multiple companies, clients, or years and can name a pattern others have not articulated. Structure: pattern identification → three instances → why the pattern exists → how to use this knowledge.
• The Decision Audit: The author walks through a high-stakes professional decision in real time -- what information they had, what they weighed, what they chose, what happened. Structure: decision context → options considered → choice made → result → what the decision revealed.
• The Useful Contrarianism: The author argues against a mainstream professional belief using evidence. Structure: name the consensus position charitably → present the counterevidence → argue the alternative position → acknowledge where the mainstream view is right.

Choose the structure before drafting. Mixing structures produces articles that feel unfocused and do not hold a reader's attention through the middle sections.

Step 3: Construct the Opening (First 150 Words)

The LinkedIn article opening is the highest-stakes writing in the document. LinkedIn's platform shows a preview before the reader clicks "Read more," and roughly 60--70% of people who open the article will decide whether to continue within the first three sentences.

• Open with the most specific, concrete sentence available -- a number, a date, a job title, a room, a decision -- not with context or background
• State the article's central claim or reversal within the first three sentences -- do not make the reader work to find the thesis
• Avoid every variant of these opening patterns: "In today's rapidly changing landscape," "As a [role] with X years of experience," "I am excited to share," "We all know that," "Have you ever wondered," "The question I get asked most is" -- these are invisible to readers because they signal nothing specific
• Establish credibility through specificity, not assertion -- "I have managed 47 product launches across three enterprise companies" is credible; "as a seasoned product leader" is not
• The ideal opening contains: one concrete professional situation, one surprising or counterintuitive element, and a clear signal of what the reader will learn -- all within 80--150 words
• If the user's experience does not immediately suggest a strong opening, lead with the disconfirming data point or the most counterintuitive result, then build backward to explain how the author got there

Step 4: Draft the Body Sections (3--5 Sections, 150--250 Words Each)

Each body section makes one claim and supports it with one type of evidence. Never stack two major claims in one section -- readers lose the thread and the article feels rushed.

• Use H2 headings that function as standalone argument summaries -- a reader scanning only the headings should be able to reconstruct the article's logic. "Background" and "Key Insights" are useless headings; "The Metric Most Teams Measure Is a Lagging Indicator" is a heading that carries information
• Evidence types in order of LinkedIn reader persuasiveness: (1) first-person professional experience with specific details, (2) peer-observed pattern across multiple organizations, (3) industry data with source context, (4) expert reference or cited research, (5) logical inference from first principles -- use type 1 or 2 as the primary evidence for at least two sections
• Keep paragraphs to 3--4 sentences maximum -- LinkedIn's mobile reading experience punishes dense paragraphs; white space is a feature, not wasted space
• One section in every article should include a moment of intellectual honesty: a caveat, a failure, an admission that the author's advice has limits, or a place where the author got it wrong before getting it right -- this is the section most likely to generate comments and shares because it signals trustworthiness
• Use "I" freely -- LinkedIn rewards first-person narration; passive voice and institutional "we" create distance that reduces engagement
• Concrete details that belong in body sections: headcount, budget ranges, duration in weeks or months, percentage changes, geographic scope, decision-maker titles, industry names -- any specific detail that grounds the observation in reality
• Transition between sections using the "so what" connector: end each section with one sentence that sets up why the next section matters, rather than leaving readers to figure out the logical progression themselves

Step 5: Write the Nuance or Caveat Section (Optional but Recommended)

This section is optional only for articles making low-controversy claims. For any contrarian, prescriptive, or experience-based argument, include it.

• State explicitly when the article's advice does not apply -- if the author recommends eliminating a practice, name the conditions under which that practice remains valuable
• Use "This does not apply when..." or "The exception is teams that..." framing rather than defensive hedging -- this signals confident expertise rather than uncertainty
• Keep this section to 100--180 words -- it is supporting argument, not the main event
• Use this section to address the obvious objection the target reader is most likely thinking -- readers who feel their context is unacknowledged will discount the article's entire argument
• This section improves algorithm performance because articles that acknowledge nuance tend to generate higher-quality comments, which the LinkedIn algorithm weighs more heavily than generic reactions

Step 6: Write the Closing and Discussion Invitation

The closing does two jobs: crystallizes the article's argument into its most memorable form, and initiates the comment conversation that determines algorithm amplification.

• Crystallize the key insight into one or two sentences that the reader could quote in a meeting -- if the closing cannot be reduced to a memorable statement, the article's argument is not sharp enough
• Invite discussion with a specific question tied directly to the article's content -- not "What do you think?" but "Which of these three patterns matches your organization's current situation?" or "Has anyone run this experiment with a team over 50 engineers?"
• The discussion question should be answerable by the target reader in two or three sentences -- questions that require the reader to write an essay get skipped; questions that invite a quick opinion or brief experience-sharing generate volume
• Avoid "feel free to connect" or "reach out if you want to discuss" -- these signal that the author is using content as sales outreach, which erodes trust
• Do not end with a list of services, a promotional paragraph, or a calendar booking link -- these belong in the author's profile, not in the article body

Step 7: Apply LinkedIn-Specific Formatting and Metadata

• Hashtags: include 3--5 at the bottom of the article -- choose specific industry or function hashtags with established communities (#ProductManagement, #EngineeringLeadership, #SupplyChainManagement) rather than generic aspiration hashtags (#Success, #Leadership, #Motivation); avoid hashtags with fewer than 10,000 followers or more than 5 million (too small = no audience; too large = your content drowns instantly)
• Article title: keep under 70 characters so it displays fully on mobile without truncation; include a specific tension, number, or counterintuitive signal; test two title versions mentally by asking "would a target reader stop scrolling to read this?"
• Bold text: use for 1--2 key phrases or data points per section only -- never bold entire sentences or paragraphs; LinkedIn's editor renders bold in articles, and overuse signals lack of confidence in the prose
• Paragraph breaks: every 3--4 sentences maximum, ideally every 2--3 on mobile -- LinkedIn's mobile interface is the primary reading environment, and paragraphs that look reasonable on desktop become walls of text on mobile
• Hero image: remind the user that LinkedIn article cover images display at 744 x 400 pixels and should reinforce the article's argument visually rather than using stock photography of handshakes or lightbulbs -- minimalist text overlays on strong backgrounds consistently outperform generic imagery
• Word count: target 900--1,500 words for maximum completion rate -- LinkedIn's algorithm uses dwell time and scroll depth as quality signals; articles over 2,000 words show declining completion rates unless the author has an established large following

---

Output Format

# [Article Title -- Under 70 Characters, Contains Specific Tension or Number]

[Opening: 80--150 words. Specific situation, concrete detail, or counterintuitive
data point. Thesis stated within first three sentences. No warm-up language.
Credibility established through specificity, not credential claims.]

## [H2 Heading -- Full Argument Claim, Not Just a Topic Label]

[150--250 words. One claim, one primary evidence type. First-person
professional experience with concrete details: numbers, timeframes, team sizes,
measurable outcomes. 3--4 sentences per paragraph. One transition sentence
connecting to next section.]

## [H2 Heading -- Full Argument Claim]

[150--250 words. Different evidence type from previous section. Connects to
target reader's professional reality: "If you are managing a team that...",
"When your organization reaches the point where...". May include the
intellectual honesty moment: a failure, a wrong assumption, a limit of the
author's experience.]

## [H2 Heading -- Full Argument Claim]

[150--250 words. Practical application or prescriptive recommendation. Specific
and actionable: names what the reader does, when, with whom, and how to measure
whether it worked.]

## [H2 Heading -- Nuance or Exception (include when article makes prescriptive
or contrarian claims)]

[100--180 words. Explicit conditions under which this advice does not apply.
Addresses the obvious objection. Uses "This does not apply when..." framing.
Shows confident expertise rather than defensive hedging.]

## [Closing H2 Heading -- Optional, or integrate into final body section]

[80--120 words. Key insight crystallized into one quotable statement.
Specific discussion question answerable in 2--3 sentences. No promotional
language, no "feel free to connect," no services pitch.]

---

#[IndustrySpecificHashtag] #[FunctionSpecificHashtag] #[TopicHashtag]
#[PracticeAreaHashtag] #[OptionalFifthHashtag]

Word count targets by article goal:

Goal	Target Word Count	Section Count
Focused insight (one technique or observation)	800--1,000	3--4 sections
Standard thought leadership	1,000--1,400	4--5 sections
Complex argument or research-backed	1,400--1,800	5--6 sections
Never exceed	2,000	--

---

Rules

1. Never open with emotion or excitement. "I'm thrilled to share," "Excited to announce," "Proud to say," and all variants are banned. They signal personal milestone posts, not professional insight articles, and immediately lower the perceived credibility of the argument that follows.

2. Never use the phrase "In today's [adjective] landscape/world/environment." This opening is the single most overused construction on LinkedIn and has become a credibility signal in reverse -- readers immediately recognize it as filler and disengage. The same applies to "Now more than ever," "The future of [industry] is," and "We live in unprecedented times."

3. Never assert expertise -- demonstrate it. Phrases like "as a thought leader," "as an expert in," "with my extensive experience," and "I know this industry well" are forbidden. Credibility is built by specificity: "after managing 14 product launches across two enterprise SaaS companies" communicates expertise; "as a seasoned product professional" does not.

4. Never write a heading that is just a topic label. Headings like "Background," "Key Takeaways," "My Experience," or "Conclusion" waste the heading's function as a standalone argument carrier. Every H2 must be a complete claim: something that could be read in isolation and communicate a specific professional idea.

5. Every article must contain at least one failure, caveat, or intellectual honesty moment. LinkedIn readers are sophisticated professionals who immediately distrust content that presents only successes and recommendations without acknowledging limits, wrong turns, or exceptions. The failure or caveat section is frequently the most-commented section of a high-performing article.

6. Never end with generic engagement bait. "What do you think?" "Let me know in the comments!" "Like and share if you agree!" are engagement-farming phrases that experienced professionals recognize and discount. The closing discussion question must be specific enough that a reader's professional background determines their answer.

7. Hashtags must be functional, not aspirational. Do not include #Success, #Motivation, #Hustle, #Leadership (too broad at 70M+ posts), or #Business. Use specific function (#FinancialPlanning, #DevOps, #SalesEnablement) and industry (#HealthcareTech, #ManufacturingIndustry) hashtags that connect the article to communities that will find value in it.

8. Paragraph length is a quality signal. Every paragraph must be 3--4 sentences or fewer. LinkedIn's primary reading environment is mobile. Dense paragraphs on mobile look like walls of text and cause readers to abandon the article regardless of the quality of the underlying argument. White space between short paragraphs is structural, not decorative.

---

Edge Cases

The User Has No Quantitative Evidence

Many practitioners have deep experiential knowledge but no hard numbers to cite. Do not manufacture statistics or prompt the user to include unverified figures.

• Use directional approximations that the user can honestly stand behind: "roughly doubled," "cut the process from weeks to days," "three of the four companies I worked with showed this pattern"
• Shift the evidence type to observed pattern across multiple situations: "I have seen this dynamic in every company I have worked with that had more than 50 engineers" is credible without requiring a data source
• Use timeline and context specificity as a substitute for quantitative precision: "In the six months between our Series B and Series C, we ran this experiment across two product teams" signals lived experience without requiring a number
• The rule is: never invent data, but also never leave the reader with pure assertion -- find the specific professional detail that grounds the observation even without a percentage or count

The User Wants to Promote a Product, Service, or Company

Articles that function as thinly disguised advertisements consistently underperform on LinkedIn because the platform's recommendation algorithm de-prioritizes promotional content and readers disengage when they realize they are reading a pitch.

• Restructure as problem-insight-solution: lead with the professional problem, develop the insight about why most approaches fail, and mention the product or company only in the final paragraph as one instantiation of the principle -- not as the conclusion
• The ratio should be approximately 85% professional insight, 15% or less mention of the specific offering
• The product mention should be framed as "what we built because this problem did not have a solution that worked for us" rather than "here is why you should choose us"
• If the user insists on a primarily promotional article, flag that this content will likely underperform LinkedIn's article distribution algorithm and may be better suited to a LinkedIn Company Page post or an ad format

The User Is Early-Career and Has Limited Professional Experience

An early-career professional can write a credible LinkedIn article -- but the framing must match their actual experience level, or the article reads as overreach and damages rather than builds credibility.

• Reframe from authority ("here is what you should do") to observation and learning ("here is what I learned, observed, or discovered in my first [timeframe] as a [role]")
• "What Nobody Told Me Before My First Product Manager Job" is a credible article from a junior PM; "Why Most Product Managers Do It Wrong" is not
• Learning articles can cite: informational interviews with senior practitioners, observed patterns in companies the author has worked at, synthesis of published research with personal application, or direct comparison between academic training and professional reality
• The early-career author's credibility asset is specificity about their experience as a learner, not authority as a practitioner

The Topic Is Politically or Professionally Controversial

Some professional topics -- diversity hiring practices, layoff strategies, executive compensation, AI replacing jobs -- carry genuine controversy within professional communities and require careful structural handling.

• Acknowledge the controversy directly and explicitly in the opening paragraph -- do not pretend the position is uncontested
• Strengthen the evidence section: the more controversial the claim, the more specific the evidence must be -- a contrarian claim supported only by the author's personal experience will read as bias; the same claim supported by observed pattern across multiple organizations reads as professional insight
• Use the nuance section to acknowledge the strongest opposing argument charitably and then explain why the author still holds their position -- this is the single most effective technique for maintaining credibility while taking a controversial stance
• Avoid political party references, electoral politics, or social justice framing unless the user is explicitly a professional in those domains -- LinkedIn's professional context means readers expect arguments made in professional, operational, or evidence-based terms, not political ones

The User Wants to Republish an Existing Blog Post or Article

Direct copy-paste from a website blog to LinkedIn article format produces underperforming content because blog posts and LinkedIn articles have different structural requirements, voice conventions, and reading contexts.

• Shorten paragraphs: most blog posts are written for desktop readers and use 5--8 sentence paragraphs that become unreadable on LinkedIn's mobile interface
• Add a personal first-person frame: blog posts often use an authoritative third-person or editorial "we" voice; LinkedIn requires first-person professional narration
• Convert the introduction: blog SEO introductions typically include keyword context and problem setup that wastes LinkedIn readers' limited patience -- replace with the sharp opening described in Step 3
• Add a discussion-inviting closing that did not exist in the original blog version
• Add the hashtag block
• Include a one-line attribution at the very bottom: "This article originally appeared on [publication name]." -- this is both honest and helps readers who want to find the original source

The User Has a Great Topic But Cannot Identify Their Unique Angle

When the user knows they want to write about a topic but all their initial framings sound generic, use these diagnostic questions:

• "What have you seen done on this topic that almost everyone does wrong?"
• "What did you believe about this topic two years ago that you now know is incorrect?"
• "What is the most expensive mistake you have seen someone make in this area?"
• "If you had 10 minutes with a junior person in your function, what is the one thing about this topic you would most want them to understand that they are unlikely to learn on their own?"

The angle that emerges from these questions will always be more specific and more engaging than the angle the user initially proposed. The goal is to move from "the importance of stakeholder communication" to "why stakeholder communication fails at companies that think they are doing it right."

---

Example

Input: "I want to write a LinkedIn article about why data-driven hiring doesn't actually work the way most companies think it does. I'm a Head of Talent Acquisition at a 1,200-person B2B software company. I've spent the last three years building out our structured interviewing and assessment infrastructure, and I'm increasingly skeptical of some of the claims. My target audience is other TA leaders and CHROs."

---

Output:

We Spent Three Years Building Data-Driven Hiring. Here's What the Data Actually Showed.

In 2021, I was convinced we were going to solve hiring at our company. We had structured interviews with behavioral anchors for every role, a seven-dimension scorecard, two psychometric assessments, and a calibration process that required consensus before any offer. We called it rigorous. We told candidates it meant we were fair.

By 2024, our quality-of-hire metrics were not better than the industry average. Our time-to-fill had increased by 22%. And we had quietly rejected several candidates who, we later learned, had been hired by competitors and were performing exceptionally well. The data-driven system was producing confident decisions, not accurate ones. There is a difference.

Structured Processes Reduce Inconsistency -- They Do Not Improve Prediction

The strongest argument for structured interviewing is also the most misunderstood one. Structured interviews reduce the variance between interviewers. They make your hiring more consistent. Consistency is genuinely valuable -- it protects against bias and makes the process defensible.

But consistency and predictive accuracy are not the same thing. A structured process applied consistently to the wrong signal produces consistent errors. Our seven-dimension scorecard measured things like "strategic thinking" and "cross-functional collaboration" through behavioral interview responses. After three years, we found that scores on these dimensions in the interview correlated weakly with manager-rated performance at 12 months. We were measuring how well candidates performed the interview, not how well they would perform the job.

This is not a knock on structured interviewing -- it is a clarification of what structured interviewing can and cannot do. It reduces bias in how you apply your criteria. It does not validate whether your criteria are right.

Psychometric Assessments Are More Useful for Ruling Out Than Ruling In

We use two assessments: a cognitive ability test and a personality inventory. Both have peer-reviewed validation evidence, and I stand behind using them. What I no longer believe is that they should be used as ranking tools to identify the best candidate from a pool of finalists.

The research on cognitive ability tests shows strong correlation with job performance at the population level. At the individual candidate level, the confidence intervals are wide enough that distinguishing between two finalists who score a 74 and a 79 is not meaningfully different from a coin flip. We were using population-level statistics to make individual-level decisions and treating the resulting confidence as earned.

Assessments work well as screens: candidates below a threshold worth establishing based on job complexity are meaningfully less likely to succeed. They work poorly as sorters among qualified candidates. We now use assessment scores as a floor, not a ranking mechanism, and our offer acceptance rates have improved because we are not losing good candidates to a decimal point.

The Missing Variable: How the Work Actually Gets Done

Every component of our hiring process evaluated candidates independently of the specific team, manager, and working context they would join. We assessed strategic thinking in the abstract. We never evaluated whether a candidate's decision-making style matched how their future manager provided feedback, or how the team handled disagreement, or what the real work environment looked like in terms of autonomy versus oversight.

Two candidates with identical scorecard results will perform differently in two different managers' teams. We knew this intuitively, but our process had no mechanism for capturing it. We had optimized the process for evaluability -- for producing a number that felt justified -- rather than for prediction of this specific person in this specific context.

The intervention that has had the most impact on our quality-of-hire metric in the past 18 months has been a structured conversation between finalists and their direct manager focused specifically on working style, feedback preferences, and what success looks like in the first six months. We do not score it. We ask the hiring manager to reflect on it. It is the least data-driven part of our process and consistently the most predictive part of our retrospectives.

When Data-Driven Hiring Works and When It Misleads

None of this means abandon your structured process. High-volume, clearly defined roles with objective performance outputs -- sales development, customer support, certain engineering functions -- benefit enormously from rigorous structured assessment, and the predictive validity holds more reliably when performance is measurable and the job scope is narrow.

The approach breaks down for senior leadership roles, highly cross-functional roles, and any role where performance is primarily determined by relationships and influence rather than discrete outputs. The more ambiguous and contextual the role, the less predictive power your structured data produces -- and the more dangerous it becomes to over-rely on it because the confidence it generates is disproportionate to its actual accuracy.

What I Would Tell My 2021 Self

Structured processes are infrastructure, not intelligence. They make your decisions defensible and your process consistent, which is worth building. But they do not exempt you from the harder question: are we measuring the right things, validated against actual performance in our specific environment?

We have started tracking quality-of-hire by interview scorecard pattern, by interviewer, and by assessment score band. We are, for the first time, generating data about whether our data-driven process actually predicts what we think it predicts. The answer so far is: partially.

If you have built a structured hiring process, what are you using as your quality-of-hire validation metric? I am particularly curious whether other TA leaders have found scorecard components that actually hold up in your retrospective analysis -- or whether you are seeing the same weak correlations we found.

---

#TalentAcquisition #HiringStrategy #PeopleAnalytics #HRLeadership #StructuredHiring