On Tuesday morning an asset management company publishes a ten-slide carousel on its LinkedIn Page. It explains how a target maturity fund works, it went through compliance twice, and the standard mutual fund warning sits in small type on the last slide. By Thursday the marketing head is looking at a different post: the fund manager’s own three-paragraph note on rate cuts, typed on her phone, with a long thread of comments from treasury heads and distributors. The brand asks why the LinkedIn algorithm ignored the expensive post. The better question is what the feed measured in the first few seconds each reader spent on it.
What is the LinkedIn algorithm?
The LinkedIn algorithm is LinkedIn’s name for the ranking models that decide the order of posts in each member’s feed. There is no single formula. LinkedIn’s engineering team describes a pipeline: one system finds a pool of candidate posts for a member, a second scores and orders them. LinkedIn also says its feed filters out spam, low-quality and unsafe content.
LinkedIn’s own help pages say the default feed view, called Top, is personalised from your activity and interests. The other view, Recent, lists posts from people you follow in reverse chronological order without personalisation. Top is the default, and LinkedIn says the sort control sits on desktop, so for most readers the LinkedIn algorithm decides what a buyer at a bank, an insurer or a fintech sees when they open the app.
In March 2026 LinkedIn’s engineering blog described the newest version of the feed. Candidate posts are now retrieved by a single system built on large language model embeddings, replacing several separate sources such as a chronological network index and trending lists. Ranking runs on a sequence model LinkedIn calls a Generative Recommender, which reads a member’s history of more than a thousand past interactions in order. The same post says ranking uses professional signals and engagement patterns and never uses demographic attributes.
For a B2B marketer this changes what “reach” means. A post competes for a place in the feed of anyone whose recent behaviour suggests interest in the topic, follower or not.
How does the LinkedIn algorithm differ from LinkedIn ad delivery?
The LinkedIn algorithm ranks organic posts; ad delivery decides where paid content appears. They are separate systems with separate rules, and confusing them is the most common reason a monthly LinkedIn report says nothing useful.
When you pay for a Sponsored Content campaign, LinkedIn shows the ad to the audience you targeted, through an auction, and what you are charged for depends on your optimisation goal and bidding strategy. LinkedIn’s help pages say sponsored posts are shown to members who follow the sponsoring organisation and may also be shown to members who fit the advertiser’s audience criteria. The organic feed ranking described in LinkedIn’s March 2026 engineering post covers only organic posts and says nothing about ads.
The practical difference shows up in three places in a B2B plan:
- Organic reach is earned post by post, and a weak opening line costs you the whole post. Paid reach is bought, and a weak opening costs you money per impression.
- Organic performance depends on who posts. A specialist with a network of treasury heads behaves differently from a brand Page.
- Thought Leader Ads sit between the two. They put paid budget behind a post from an employee or another member who approves the request, so a post that has already proven itself organically can be extended to a target account list.
A useful rule: learn from organic results which ideas hold attention, then put paid budget behind the winners.
How does the LinkedIn algorithm decide what a buyer sees?
The LinkedIn algorithm works in two stages that matter to a brand: it retrieves candidate posts for each member, then it ranks them by predicted reactions.
Retrieval. According to LinkedIn’s March 2026 engineering post, a fine-tuned language model turns both the member and each post into embeddings, and the feed finds posts whose embeddings sit close to the member’s. Training used posts the member engaged with as positives, and posts shown to the member that drew no engagement as hard negatives. For a new member with little history, the model can infer likely interests from the headline and job title. That is why a clear, specific post about co-lending or UPI (Unified Payments Interface) settlement can reach a chief risk officer who has never followed your Page.
Ranking. The ranking model predicts several outcomes at once. LinkedIn groups them into passive tasks such as a click, a skip and a long dwell, and active tasks such as a like, a comment and a share. A 2021 LinkedIn engineering post described the same split: passive consumption objectives balanced against active contribution, with the company placing high value on comments and reshares because they keep a network engaged.
Dwell time. Dwell time is how long a member spends on a post. LinkedIn’s 2020 engineering post on dwell time measured it in two ways: time on the feed while at least half of an update is visible, and time spent on the content after a click. Updates viewed for less than a threshold were treated as skips, and a predicted skip lowered a post’s score.
Freshness. The 2026 post says member and post representations update continuously, so a post that starts drawing engagement can reach relevant members within minutes. Early reaction from the right readers matters more than the clock time you posted.
Quality filters. LinkedIn’s spam help page says it may remove or limit the distribution of content designed to inflate engagement artificially. Its examples include emoji or reaction polls used to boost engagement, chain-letter posts asking for likes and shares, and excessive or repetitive comments.
The feed rewards a post a treasury head reads to the end; compliance requires that she meets the warning on the way.
How does the LinkedIn algorithm meet compliance rules in India and the Gulf?
The LinkedIn algorithm does not know your regulator, but your regulator decides where in the post the reader must meet the required lines. Every format choice that improves dwell time has to be checked against that placement first. This section describes the rules generically; it is general guidance, not legal advice, and your compliance team decides.
SIGNALNewsletter by Yamm Labs
The next piece, straight to your inbox.
New writing on BFSI brands, B2B and influencer marketing, and AI search.
One click to unsubscribe. Privacy policy
Mutual funds and SEBI. The Securities and Exchange Board of India (SEBI) requires the standard mutual fund warning on advertisements, and in video and audio ads the warning, shown on screen and spoken, must be clear and understandable. SEBI gives five seconds only as an example of what can meet that test. On LinkedIn that shapes a design decision before filming: in a 15-second scheme video a five-second warning takes a third of the running time, so in our practice 30 seconds is the shortest paid cut worth making for a scheme ad. Muted autoplay is common in the feed, so the on-screen version of the warning carries most of the weight.
Insurance and IRDAI. The Insurance Regulatory and Development Authority of India (IRDAI) replaced its earlier advertisement rules with 2024 regulations on protecting policyholders’ interests. They require every insurer’s advertisement to show the insurer’s registered name prominently with its trade name or logo, and an advertisement issued by a distribution channel needs the insurer’s written approval before it goes out. On LinkedIn, the second rule touches employee advocacy at a bancassurance partner or a broker: a carousel about a term plan posted by a channel partner’s sales head can count as an advertisement that needs the insurer’s sign-off first.
Lending and RBI. The Reserve Bank of India (RBI) requires lenders to give borrowers a Key Facts Statement in a standard format, including the annual percentage rate, before a retail or MSME loan contract is signed. For commercial banks this rule now sits in RBI’s November 2025 Responsible Business Conduct Directions. A LinkedIn post from a non-banking financial company (NBFC) that quotes a headline rate for MSME (micro, small and medium enterprise) loans therefore needs a landing page that is consistent with what the borrower will later see in that statement. Plan the post and the landing page together.
Influencers and employees under ASCI. The Advertising Standards Council of India (ASCI) influencer guidelines list “Employee” among acceptable disclosure labels and say a disclosure must not be hidden at the end of a post or buried among hashtags. They also expect banking, financial services and insurance influencers to disclose qualifications such as SEBI registration or an IRDAI licence upfront. So when a fund manager posts about her own scheme, the label goes in the first lines, before the “see more” cut, where the reader actually starts reading.
The Gulf. In the United Arab Emirates the Central Bank of the UAE, the Dubai Financial Services Authority (DFSA) and the Capital Market Authority (called the Securities and Commodities Authority until 2026) each run their own rules on financial promotion and consumer protection, depending on where a firm is licensed. In Saudi Arabia the Saudi Central Bank (SAMA) and the Capital Market Authority set the equivalent rules. Yamm Labs works from Gurugram and serves Gulf clients remotely, so every Gulf LinkedIn plan starts with a question to the client’s own compliance team: which regulator licenses the entity that owns the Page, and which entity employs the people who will post.
The global contrast. The UK Financial Conduct Authority’s 2024 guidance on financial promotions on social media is the most specific public text on formats. It expects prescribed risk warnings to stay prominent and not be cut off by a platform’s design, and it treats a post that puts the benefits in the image and the risks only in the caption as giving risks too little prominence. For carousels, it expects the warning on every slide that carries the promotion. Indian and Gulf rules are worded differently, but the FCA text is a good test: if a carousel would fail it, rethink where your own disclaimer sits.
Where do you meet the LinkedIn algorithm in practice?
You meet the LinkedIn algorithm every time a post’s first few lines decide whether a reader stays. The table below maps each part of the system to the place a B2B or financial marketing team runs into it, and why it matters when compliance reads the same post.
| Term | Plain meaning | Where you meet it | Why it matters for a regulated brand |
|---|---|---|---|
| Top and Recent views | Top is the personalised, ranked feed; Recent lists followed accounts by time | The sort control above the feed on desktop | Top is the default, so ranking decides most organic reach |
| Retrieval | Finding a pool of posts a member is likely to care about | A post reaching people who do not follow your Page | Your disclosure must work for strangers who have never seen your brand |
| Dwell time | Time a member spends on a post, on the feed or after a click | Long carousels, document posts and detailed text posts | Long formats can carry disclosures without losing the reader, if the warning sits with the claim |
| Skip | A post scrolled past faster than a set threshold | Posts that open with a logo card or a slogan | A predicted skip lowers the score; a vague first line costs more than a disclaimer does |
| Click bounce | Opening a link and returning within seconds | Posts that send readers to a slow or generic landing page | The landing page must match the post, which also helps the lending and insurance rules above |
| Active engagement | Comments, shares and likes | Threads under an executive’s post | Replies from your staff about a product are communications too; brief them on what they may say |
| Artificial engagement | Polls, pods or chain posts built to inflate reactions | Engagement groups and “comment YES” posts | LinkedIn says it may limit such content, and pre-arranged praise for a scheme or plan invites a compliance question |
| Thought Leader Ads | Paid promotion of a post from an employee or other member | Boosting the fund manager’s note to a target account list | The paid version is an advertisement, so it gets the same approval as any ad |
Two habits follow. Write the first two lines of every post for a stranger, and keep a claims record for every post that names a product, a return figure or a rate.
Which LinkedIn algorithm claims are unproven?
Most LinkedIn algorithm advice online is inference from creators’ own accounts, not a statement from LinkedIn. Before you change a B2B content plan on the strength of a tip, check whether LinkedIn’s help pages or engineering blog say it.
Several popular claims have no support in LinkedIn’s public material that we could find:
- A fixed “golden hour” after posting that decides a post’s fate. LinkedIn’s 2026 post talks about freshness within minutes, which is a different and weaker claim.
- A precise ratio of how much a comment outweighs a like. LinkedIn does not publish the weights of individual actions.
- A blanket penalty on every post with an outside link. The dwell time post describes click bounces as a noisy signal, which points to judging the landing experience, not banning links.
- A best number of hashtags. None of the official pages we read mention hashtags as a ranking factor.
LinkedIn has said a few things plainly. Its November 2025 engineering post says its feed systems consider hundreds of signals, including profile details, network and activity, and do not use demographic information such as age, race or gender to decide visibility. Its 2022 “mythbusting” post on the member blog says widely read content tends to offer ideas, insights and inspiration, and that the feed filters out low-quality and unsafe content.
For a regulated brand this gap is useful. Build the plan on what LinkedIn has confirmed: hold attention, invite real professional replies and avoid manufactured engagement. Those are also the easiest rules to defend in front of compliance.
How should a B2B financial brand plan for the LinkedIn algorithm?
Plan for the LinkedIn algorithm by deciding who posts, what the first lines say and where the disclosure sits, before anyone designs a slide.
- Pick the voice. Decide which posts come from the Page and which from named people: a chief investment officer, a head of product, a head of partnerships. A personal post from a specialist tends to draw the professional comments the ranking model counts as active engagement.
- Write the opening for a stranger. State one specific idea in the first two lines. “Our new treasury platform” invites a skip. “Why most corporate treasurers still reconcile UPI collections by hand” gives a treasurer a reason to stop.
- Fix the required lines with compliance. Agree the standard warning, the label and the registered name for each format before design starts: text post, carousel, video and Thought Leader Ad.
- Choose a format that holds attention. A document post or carousel that teaches one thing well earns dwell time. Put the warning on every slide that carries a product claim, not only the last.
- Brief the people who will reply. Comments from your staff are part of the post’s reach and part of your communications. Give them a short note on what they may and may not say about products and returns.
- Measure what the feed measures. Track comments from target accounts, saves and reposts, and time spent where LinkedIn reports it. Report meetings with target accounts as the business result, never impressions alone.
Yamm Labs works from Gurugram with financial brands such as TATA AIA Life Insurance, HDFC Life, Axis Bank, Paytm Money and Aditya Birla Sun Life Mutual Fund; the full list is on our clients page. Our financial services social media work starts with this same order: voice, opening, required lines, format, replies and measurement.
What should you tell your compliance team about the LinkedIn algorithm?
Tell your compliance team that the LinkedIn algorithm rewards attention, not volume, and that the plan keeps every required line where a reader meets the claim.
A short note you could paste into an email: “We are shifting part of our LinkedIn effort from Page posts to posts by named specialists. Each post that mentions a product will carry the required warning or label in its first lines or on every relevant slide, never only in a comment or at the end. Employee posts will use the label our advertising code expects. Any post we put paid budget behind goes through the same approval as an advertisement. Approvals and the final text are kept in our claims record.” This is general guidance, not legal advice; your compliance team decides.
Related terms for the LinkedIn algorithm
These terms come up when a B2B team discusses LinkedIn reach with an agency or with compliance.
- Generative Recommender. LinkedIn’s name for the sequence model that ranks the feed by reading a member’s past interactions in order. It matters because recent behaviour shapes what a member sees next, so a buyer who has been reading about co-lending will see more of it this week.
- Long dwell. A prediction that a member will spend a long time on a post. It matters because detailed, useful posts can earn reach without chasing reactions.
- Employee advocacy. Staff sharing or writing posts about their employer’s work. It matters because those posts are often the ones a ranking model favours, and in financial services they can be advertisements.
- Claims record. A log of every product claim in a post, with its source and approval. It matters because it lets you answer a regulator’s or a client’s question months later.
For quick answers on finance content questions, see our finance social media questions page.
People also ask
Does the LinkedIn algorithm favour personal profiles over company Pages?
LinkedIn does not publish a rule that favours one over the other. Its ranking predicts comments, shares and long reads, and personal posts from specialists often earn more of those than brand posts. Plan both, with named people carrying the ideas and the Page carrying products and announcements.
Does LinkedIn penalise posts with external links?
LinkedIn’s public material does not state a blanket link penalty. Its 2020 engineering post describes “click bounces”, quick returns after a click, as a noisy signal, so a link to a slow or generic page can hurt. A link to a page that delivers what the post promised is a smaller risk.
How long should a LinkedIn post be for a B2B audience?
Long enough to teach one thing well. LinkedIn’s ranking predicts skips and long dwell, so a short post with a vague opening loses more than a detailed post with a specific one. For a financial brand, length also gives room to place the required warning beside the claim.
Can a mutual fund or insurer use LinkedIn polls to boost reach?
Polls that ask a real professional question are allowed, but LinkedIn says it may limit polls designed to inflate engagement. A poll that mentions a scheme or a plan is still an advertisement, so it needs the same approval and warning as any other post.
Want to see how your LinkedIn posts read to a buyer and to your compliance team?
Get a free Fintech Brand Audit across 8 dimensions, plus 3 fixes. No pitch. Run the free Brand Audit →