How LinkedIn Works

Corporate executive coaching is bought through people. In the ICF’s global study of coaching clients, 46% found their coach through a personal referral or word of mouth. Only 7% used a search engine, and LinkedIn didn’t appear on the list at all [91]. Organizations buy the same way. In the 2019 ICF and HCI study, 38% chose coaches through referrals from people they trust and 29% through consulting firms, while “targeted searches on the internet (i.e., Google, LinkedIn)” came to 9% [92]. An earlier survey of 40 major buying organizations found 70% relied on referrals, recommendations or coaches they already knew [93].

Newer research doesn’t change the picture. LinkedIn’s own buyer study (Ipsos, 900+ B2B buyers) says buyers “typically shortlist just 2-3 vendors” and that AI now shapes how they evaluate vendors “before any human contact takes place” [44]. A 2025 survey of 511 B2B buyers found 40% begin vendor research on a search engine, 19% with peer recommendations, 15% at events and 9% on LinkedIn or other social media [94]. When HR practitioners on Gartner’s Peer Community traded notes on finding executive coaches, they talked about referrals, vetted panels, coaching associations and RFPs. Nobody mentioned LinkedIn [95]. Training Industry’s 2025 guide for learning leaders names industry reports, directories, awards and testimonials as the tools for picking vendors [96]. No published study from 2023 to 2026 is known to show HR or L&D buyers using LinkedIn to find coaches.

So what is LinkedIn for? The moment that matters comes after the referral. A CHRO hears about Executive Speaking from a peer and looks up the profile. What turns up either confirms the referral or cools it: an active expert teaching something useful every week, or a quiet profile full of other people’s posts. Everything in this report is about winning that moment, and about staying visible to executives between referrals.

The first two stages are LinkedIn’s own description: “While retrieval determines which posts reach the ranking stage, ranking determines what a member actually sees in their Feed” [1]. The feed serves more than 1.3 billion professionals and ranks millions of posts at any moment [1].

What the Model Looks At

LinkedIn says The Help Center says the system weighs “hundreds of signals,” among them the context of a post (a helpful insight, a job opportunity, a career milestone) and signals from the reader’s profile, network and activity [2]. The author counts too: “the system that ranks and surfaces content factors in the author’s professional identity, such as experience, content shared on their profile, and skills” [4]. The March 2026 system also reads what readers put on their own profiles (industry, experience, skills, geography) and tracks how their interests change [3].

Two things follow. The author’s profile is a ranking input, so skills, experience and published work all count. And a post that matches a reader’s interests can travel well beyond the author’s follower count. LinkedIn says the new approach helps most with “cold-start users and sparse content niches” [1], and that “exceptional content may even be distributed broadly across LinkedIn to members who are interested in the type of content you post, even if they don’t follow you” [4]. Executive communication coaching is that kind of niche.

What ranking leaves out: LinkedIn says it never uses demographic information such as age, race or gender to decide who sees what [2].

The Network Still Comes First

In 2019 LinkedIn said more than 80% of feed updates came from its network-based system, and those updates produced more than 95% of members’ conversations [7]. The figure predates the 2026 changes, so it’s history. The direction holds, though. Most reach still starts with followers and connections, and reach beyond them is earned on top.

LinkedIn also gives fresh posts a nudge. It tuned candidate selection for freshness because “content creators do care about receiving prompt responses,” even though readers care less [7]. That’s the documented kernel of the “golden hour” idea. There’s no published stopwatch (see the myths section).

Under the Hood

LinkedIn researchers have described 360Brew, a 150-billion-parameter model trained on LinkedIn data for ranking and recommendations. The public paper was later withdrawn from arXiv, and there’s no public confirmation of how much of the live feed it runs [45]. Vendor claims that 360Brew now runs the whole feed are unverified. The confirmed change is the March 2026 rollout described above [1][3].

LinkedIn says In a May 2020 engineering post, LinkedIn described adding a “skip” prediction to ranking: the chance a reader spends less than a set number of seconds on a post. Every post’s score drops in proportion to that chance [5]. The clock starts once at least half the post is on screen [5]. An October 2024 post described the reverse, a model that predicts whether someone will spend longer on a post than on most comparable ones. It’s recalibrated daily and was built to limit “click bait” and “dwell bait” [6].

LinkedIn’s engineers also write that “a comment is much more valuable than a click” [10], and they treat a reader’s intent to comment, share or react as a ranking goal of its own [7]. What LinkedIn doesn’t publish is how much each signal weighs. Outside studies fill that gap, with vendor numbers.

Who comments matters as well. Van der Blom reports that comments from second-degree connections carry 2.6× the impact of comments from direct connections, and that comments arriving after the first 90 minutes carry 1.8× [67].

In practice, the best posts are ones people want to keep. A checklist, a short script, the exact words to use when a board member interrupts. Close with a genuine question.

LinkedIn says From the March 12, 2026 feed update [3]:

• Engagement bait such as “Comment ‘Yes’ if you agree.”
• “Recycled” thought-leadership posts “that don’t say much in terms of substance or insight.”
• Posts whose video “has nothing to do with the associated text post,” which “should no longer gain additional reach.”
• Repetitive, click-driven posts.
• Engagement pods and comment automation, “including the use of third party software or browser extensions, which are not allowed on LinkedIn.”

LinkedIn backs this up. It says it “actively limit[s] the reach of inauthentic activity, including engagement pods and automated comments,” has removed groups that behaved like pods, and has warned members about account restrictions and removal from programs such as Top Voices [12]. Comments it detects as automated may be dropped from the “Most relevant” view and sometimes hidden from everyone outside the commenter’s network [12]. And when LinkedIn spots inauthentic engagement around someone’s own posts, “it can affect how that content is distributed across the platform” [13].

The written rules match. The Professional Community Policies ask members not to “agree with others ahead of time to like or re-share each other’s content” [14]. The spam policy names emoji or reaction polls, chain-letter requests for likes and shares, and excessive or repetitive comments [15]. Penalties range from reduced visibility and labels up to account restriction [14].

Polls top the reach column and sit near the bottom on engagement, so their big numbers turn into very little conversation. Documents and images are strong on both counts. Images edge ahead on engagement; documents lead on reach.

Documents Are the Workhorse

Study For accounts with 20,000 to 50,000 followers, the bracket this account sits in, documents are the top reach format at 1.30×, climbing to 1.49× above 50,000 [55]. Socialinsider’s study of 1.3 million company-page posts puts native documents first on engagement at 7.00%, up 14% on the year [62][63]. Its explanation suits a coaching business: “LinkedIn is one of the few social platforms where audiences are genuinely looking for practical takeaways (reports, frameworks, templates etc). Document[s] work great because they’re downloadable” [63].

The craft is specific. Across 1.2 million posts, van der Blom found that 8 to 12 slides works best and that the first slide “builds 40% of the success” [64]. One caution: median reach for documents still fell about 25% in a year, so they beat the other formats without escaping the general decline [56].

Reshares Are the Weakest Format

AuthoredUp puts it bluntly: “If you want to reference someone else’s content, write your own post about it and tag them, don’t hit the repost button” [55]. MagicPost’s numbers agree. An original post earned a median of 28 likes, a repost with added thoughts 8, and a one-click repost 2 [74].

Video: LinkedIn and the Data Disagree

LinkedIn says videos get 1.4× more engagement than other formats [38] and are watched 36% more than a year earlier [39]. Study Independent reach data mostly points the other way. Socialinsider found video views down 36% year over year for pages of every size [62], and AuthoredUp measured typical video reach dropping 72% in a year, from 1,764 people to 477 [57]. One dataset disagrees: van der Blom’s August 2025 update puts video at 1.49× reach, ahead of documents at 1.30× [64]. On balance, video is a gamble rather than a reliable reach engine. When it’s used, LinkedIn’s own advice is 30 to 60 seconds [40], and van der Blom’s data favors vertical, captioned clips over horizontal ones [64]. A video paired with unrelated text is now demoted [3].

Polls and Articles

Polls reach widest and engage least. MagicPost’s study of 11,000 polls found a median of 6 likes, against 34 for an image post [75]. A big impressions number on a poll isn’t growth. LinkedIn articles reach about 31% fewer people than an average post [58], so the article works best as the permanent archive and the feed post as the version that travels.

Reach Is Down for Everyone

A falling number isn’t a personal penalty. AuthoredUp found the typical post reached 1,211 people in June 2024 and 636 by May 2025, a drop of more than 47% [57]. Its later guide puts average reach down 34% in 2025, with 98% of users affected [56]. Van der Blom, drawing on 1.3 million posts from 50,000 creators, says active creators “lost 60% over the last two years” of their reach, and calls the late-2025 move from a connection-based feed to an interest-based one “the biggest, biggest shift they have done in the last 10 years” [68].

LinkedIn shows about 210 characters on desktop and about 140 on mobile before it cuts a post off [59]. Line breaks and emojis count toward the limit, so preview each post before publishing. The hard limit is 3,000 characters [59]. Most reading happens on phones. AuthoredUp puts mobile at over 91% of browsing [56] and van der Blom at 72% [65]; either way, the small screen is the one that matters.

Study Only 7% of posts open with a story, although story openers beat every other style [60]. The most-recommended opener, a question, came last. Short first lines win too. Across 323,111 posts, first lines under 40 characters earned a 2.61% median engagement rate against 2.08% for first lines over 200, with every length in between falling in order [60][61]. This plays to Executive Speaking’s strength. The coaching already teaches executives that facts get edited and stories survive.

How Long

The newest data favors longer posts. Across 372,126 posts, those of 1,301 to 2,500 characters earned 27% more engagement than posts under 400 [59]. Across 702,562 posts, anything under 100 words was the weakest on every format, and posts that take 60 to 120 seconds to read beat 30-to-60-second posts by 15% to 43% [61]. An older AuthoredUp guide put the sweet spot at 800 to 1,000 characters [56], but the larger, newer datasets point higher. The target: 1,300 to 2,000 characters (roughly 200 to 300 words), with a first line that earns the click on “see more.”

Plain Words, Short Paragraphs

AuthoredUp reports that posts written above a 10th-grade reading level get over 35% less reach [56]. Van der Blom says formatted posts, with short paragraphs and line breaks, perform 2.1× better [67]. Neither shows its method in detail, but both fit LinkedIn’s own dwell-time model: a post that’s easy to read gets read for longer.

Study Buffer compared each of 94,000+ accounts against itself, busy weeks against quiet ones, across 2 million+ posts. Compared with posting once a week, 2 to 5 posts a week added about 1,182 impressions per post, and 6 to 10 added about 5,001 [69]. AuthoredUp’s personal-profile data peaks sooner: 4 to 5 posts a week had the best engagement rate (2.60%, with 870 impressions per post), and above 7 a week both numbers fell sharply, to 1.79% and 586 at 8 or more [55]. Flypost found that a second post within a day costs each post about 8% to 10% of its reactions, though the pair still beats a single post [73].

The two biggest datasets overlap at 4 to 5 original posts a week, and one of them turns negative above 7. That’s the recommended cadence, held week after week. Buffer’s own line: “It’s better to post three times a week every week than to post daily for two weeks and then go quiet” [69].

Time of Day

The big timing studies contradict each other. Buffer’s 2026 analysis of 4.8 million posts favors 3 to 8 p.m. on weekdays, with Wednesday at 4 p.m. strongest [70]. Hootsuite’s data favors 8 to 9 a.m. on Tuesdays and Wednesdays [72]. Buffer’s own best window moved from working hours in 2025 to the evening in 2026 [70]. No universal best time exists. Weekdays, while US executives are at work, make a sensible default; the account’s own analytics should settle the rest.

LinkedIn says “Every comment usually generates a Comment Viral Update that is distributed to the commenter’s connections” [9]. In LinkedIn’s 2017 data these were the most engaging updates in the feed, drawing 2.5× the interactions per view of connection updates and 1.8× those of like-based updates [9]. The same 2017 write-up says comments are ranked separately for every reader using close to 100 features, including the commenter’s standing (profile views, influencer status) and what the commenter and reader have in common [9]. The default “Most relevant” view hides some comments completely [17].

LinkedIn is paying more attention to comments. In August 2026 trade press reported that LinkedIn now ranks comments by relevance to each reader and plans to surface more discussions in the feed, after its Q2 report showed time spent in comments up 18% year over year [78].

What Isn’t Proven

No one has measured whether commenting grows followers. One marketer’s two-week test at 10 comments a day saw profile views fall 17% and followers rise 0.9% [80]. A reply-rate study on the subject came from an AI tool that posts comments automatically, the exact behavior LinkedIn bans [79]. The mechanism is documented; the payoff isn’t. It’s still worth doing, because it costs nothing and puts the author’s name and headline in front of people who read the same leaders the buyers read.

How to Do It

• Leave a few substantive comments a day on posts that HR, L&D and executive readers follow. Add a point, a short story or a practical tip. “Great post” does nothing.
• Reply to every real comment on the account’s own posts. Buffer reports that replying lifts a post’s engagement by about 30% [71].
• Write comments by hand. Van der Blom warns that AI-written comments can kill a post’s growth [67], and automated ones break LinkedIn’s rules [12].

LinkedIn says Follower growth feeds the newsletter. When someone follows the author, LinkedIn automatically invites them to subscribe (they still have to accept), and subscribing makes them follow the author too [19]. By default, the author’s connections and followers were invited when the first edition went out [21].

Delivery is the catch. Each edition triggers push, in-app and email notifications, but LinkedIn says “if you receive an in-app or push notification, you should not expect to also receive an email for the same notification” [19]. So some subscribers never get the email. Practitioners have published estimates of how many, but none of them can be verified.

The feed is the dependable route. When an edition goes out, it’s also shared as a feed post [21], and subscribers see updates in their feed [20]. That feed post is the main delivery, so the writing advice above applies to it: a strong first line, something worth saving, and a real question.

• LinkedIn’s own growth tips: share editions on other platforms and by email, and make Follow the main button on the profile [22].
• Keep an Off the Cuff item in the Featured section, as it is now, so visitors can find the newsletter in one step.
• Newsletter analytics show impressions, article views (including views from the email), new subscribers, and subscribers by job title, company, seniority and location. Editions older than two years drop out of the dashboard [23].
• Keep the weekly promise. The website says “Published weekly,” and anyone can check the dates.

Search

LinkedIn says People search has no fixed order. Results are tailored to each searcher, so one query can put a profile on page 2 for one person and page 5 for another [31]. Order depends on the profile and its activity, the searcher’s network, and how other members behaved on similar searches [32]. LinkedIn’s advice is to “add relevant keywords to your headline and summary” and keep experience and skills current [33], to use standard job titles because “they are what other members search for,” and to add skills, which are “among the most common queries” [31]. Overdoing it backfires: an overly optimized profile “may be impacted by spam detection systems” [32][34]. Test searches from a couple of accounts say little. LinkedIn points to profile views as the better measure [32].

Personal Profile Versus Company Page

Study Organic company-page posts make up about 2% of what members see, and company-page text posts reach 0.42× of a page’s typical post against 0.88× for personal profiles [56]. Socialinsider found that pages with 1,000 to 5,000 followers average between 235 and 525 impressions per post, depending on format [63]. The personal profile has about 22,800 followers [98]. The Executive Speaking page has about 1,900 [99].

No LinkedIn statement on record says company pages are treated differently, and ranking looks at the author’s professional identity rather than the account type [4]. No study has posted identical content to both and compared, so the exact gap is unknown. The direction is clear enough: publish from the personal profile, and let the company page share selected pieces.

Profile Setup

• Creator mode is gone as a switch (March 2024), but its tools and analytics remain. Follow is now the default main button, and the About section sits at the top of the profile [24].
• A custom profile button (such as “Book an appointment”) is limited to “a select group of LinkedIn Premium Business, Sales Navigator, and Recruiter Lite subscribers” [25]. The Featured section is free.
• People You May Know can pass over profiles that are missing current or past companies, positions or schools [37], so every field is worth filling in.

Groups

Posts written inside a group stay in that group’s feed and don’t appear in the author’s own feed [35]. LinkedIn caps membership at 100 groups and posting at 15 posts per group per week, and it may flag the same post shared across several groups as spam [36]. Groups are a side channel. A few well-chosen ones where HR and L&D leaders talk are worth joining, with posts written for each group.

Why It Doesn’t Help Directly

Google ranks pages partly on links from other websites, and a link from LinkedIn barely registers. On the public profile, links inside posts carry rel="nofollow", the tag that tells Google not to treat a link as an endorsement [98]. The Website link on the profile doesn’t point to executivespeaking.net at all. It points to a LinkedIn redirect page, and LinkedIn’s robots.txt file tells Google’s crawler to stay out of those pages (Disallow: /redir*) [48]. Since 2019, Google has treated nofollow as a hint rather than a hard rule, so these links aren’t worthless by definition [51][52]. Nothing suggests they move rankings.

Social activity isn’t a ranking factor either. Google’s Gary Illyes put it bluntly in 2017: social media helps “not because SEs will rank you better, that’s BS, but because you market your content,” adding that “PageRank wise most social media links count as much as a single drop in an ocean” [84]. Mentions of the name without a link don’t count as a signal. Asked in 2022, John Mueller said he “wouldn’t assume that there’s like some… I don’t know… SEO factor that is trying to figure out where someone is mentioning your website name” [85].

Where It Helps: The Name Search

When a referred buyer searches Anett Grant’s name, the LinkedIn profile is one of the pages most likely to appear. Jason Barnard of Kalicube, who tracks name and brand search results, says “LinkedIn dominates Brand SERPs, especially in B2B, but also personal SERPs” [88]. That ties back to section 01: the profile is part of the credibility check a buyer runs before any call. It only shows up if the public profile is switched on. LinkedIn lets members choose which sections appear “on search tools such as Google or Bing,” and changes can take “several weeks or months” to show there [49].

Where It Helps: Telling Google It’s the Same Person

Google’s Organization markup, a block of structured data on a website’s home page, accepts a sameAs property: “a URL to your organization’s profile page on a social media or review site.” Google says the markup helps it “disambiguate your organization in search results” and can shape the knowledge panel [53]. Kalicube reports that for professionals, Google draws knowledge-panel facts from databases “such as Crunchbase, LinkedIn,” and that when LinkedIn contradicts the website, the panel can “shrink, or even disappear altogether” [89]. Consistent titles, company details and descriptions across the site and LinkedIn are cheap insurance.

Where It Helps: AI Answers

Study AI search tools cite LinkedIn often. A Semrush study run with LinkedIn’s cooperation (325,000 prompts, January to February 2026) found LinkedIn in about 11% of AI responses: 14.3% on ChatGPT Search, 13.5% on Google AI Mode and 5.3% on Perplexity. It ranked second among cited domains [87]. What gets cited matches the rest of this report. Articles made up 50% to 66% of LinkedIn citations, 54% to 64% of cited pieces shared knowledge or practical advice, reshares made up about 5%, and about 75% of cited authors posted 5 or more times a month [87]. LinkedIn helped run the study, so read the headline figure with that in mind.

Access matters too. LinkedIn’s robots.txt lets the search crawlers from Google, Bing, OpenAI and Anthropic reach profiles, posts, articles and newsletters, while blocking crawlers that gather AI training data and blocking Perplexity’s crawler entirely [48].

Where It Helps: Articles and Posts in Google

Public LinkedIn articles are indexable web pages, and LinkedIn lets authors set an SEO title (cut off after 60 characters) and an SEO description (140 to 160 characters recommended) for search results [50]. The ground is shifting, though. An analysis of Ahrefs data by Foundation found Google traffic to LinkedIn articles peaked at 33 million monthly visits in March 2024 and fell to 3.6 million by March 2026, while traffic to regular LinkedIn posts grew from 3 million to 11 million monthly visits after October 2025 [86]. Those are third-party traffic estimates from one analysis, so they show a direction, not exact numbers.

Where It Helps: Links Worth Having

The links that do count come from other websites: a podcast’s show notes, a trade article, a conference speaker page. LinkedIn is one place the people behind those pages look. In Muck Rack’s State of Journalism 2026, 47% of journalists said they spent more time on LinkedIn over the past year, and one panelist described it as having “really become the professional news feed” [90]. Steady, useful posting makes an expert easier to find, and a found expert gets quoted and linked. The chain is real, but no study has measured it.

One Risk: Copying Blog Posts Into LinkedIn

Pasting a full blog post into a LinkedIn article puts a second copy of the same text on a very strong domain. Google’s guidance on syndicated content says the canonical tag “is not recommended” for this, and that “the most effective solution is for partners to block indexing of your content” [54]. The article settings LinkedIn documents are the SEO title and description only [50], so neither option is on offer there. Publishing on the website first, then posting a shorter LinkedIn take that adds something and links back, avoids the problem.

What to Do

• Keep the public profile visible to search engines, with the headline, About and Featured sections switched on [49].
• List the LinkedIn profile and company page as sameAs links in the website’s structured data [53], and keep titles and company details identical in both places.
• Publish long pieces on executivespeaking.net first. Use LinkedIn for a shorter version that links back.
• Fill in the SEO title and description on every LinkedIn article and newsletter edition [50].
• Judge LinkedIn’s SEO value by branded search, AI mentions and the links earned elsewhere, never by the links LinkedIn itself provides.

LinkedIn says Every member gets post and audience analytics at no cost [26][27]. The fields worth watching:

• In-network versus out-of-network impressions. The share of views from people who don’t follow the account. It shows when a post has escaped the existing audience [26].
• Engagements, including saves and sends. The signals that matter most, counted separately [26].
• Follower breakdown by job title, seniority, industry and company. Shows whether more of the audience is HR, L&D and executives. It exports to a spreadsheet [27].
• Profile views and search appearances. Full profile-view detail appears only once a profile has had three or more unique viewers in seven days [30]; search appearance data updates every Wednesday [33].

A few quirks. Impressions are estimates, and the account’s own views count toward its totals [28]. Most engagement counts are kept for 1,000 days, reach for 400 days, and demographics and video data for 180 days, so a monthly export preserves year-over-year comparisons [28]. On company-page posts, an impression counts once the post is at least half on screen for 300 milliseconds [29].

To tie LinkedIn to booked calls, add UTM tags to the Calendly links shared on LinkedIn. Calendly can filter bookings by those tags and export them [97].

Premium won’t raise organic reach. Premium Business and All-in-One include monthly credits to boost posts and invitations that ask engagers to follow the account. Both are paid promotion, and no Premium plan lists a change to how the feed ranks anyone [41]. Boosting a post is an ads product [42]. Sales Navigator, the prospecting tool, starts at $119.99 a month [43].

Paid LinkedIn can work for a high-ticket service if it’s set up well: Thought Leader Ads run from the personal profile, a retargeting layer, and a budget of at least about $50 a day. That evidence comes mostly from practitioners, and it’s a separate decision. Nothing in this plan depends on it.

LinkedIn doesn’t publish its ranking weights. The format and engagement multipliers in this report come from outside firms, and most of them sell LinkedIn tools. They agree on direction far more than on size.

Some sources disagree: video (LinkedIn’s growth figures against falling reach), best posting times, older versus newer advice on length, and whether posting more than 5 times a week keeps helping.

A few things nobody has measured. No study has posted the same content to a profile and a company page and compared the two. Nobody has published a verified share of newsletter subscribers who get the email. The follower payoff from commenting is unknown, and so is organic reach for LinkedIn Live.