The LinkedIn Algorithm
Teasing out the rules, do's and don'ts, and how to verify facts of the LinkedIn algorithm
Audience: Online marketers, entrepreneurs, sales professionals and leadership
Contents
Executive Summary
Introduction and Methodology
· Introduction
· Purpose
· Methodology
· Justification
· The reality on the ground
· What readers can expect to find here
Updates: Just Connecting Data 2025 EOY Algorithm Study
· Caveats: Imponderables and unmeasurables
· Summary of Just Connecting’s study
· Our view on Just Connecting’s results
Updates: Recent Algorithm Changes
LinkedIn Conventional Wisdom
· Conventional Methodology
· Necessity: LinkedIn is a Monopoly
· Background
· The Overlogix POV
· Why LinkedIn?
Update: Followers Are Real People, Not Algorithms
Questions About LinkedIn
General Observations From Multiple Authors Regarding Marketing on LinkedIn
Purported Rules to Use LinkedIn for Marketing
· Purported LinkedIn Do’s
· Purported LinkedIn Don'ts
Our Assumptions, Expectations and Observations about LinkedIn’s Algorithm
A Few Caveats About Content Linked Here
A Collection of Video Channels and Articles About LinkedIn and its Algorithms
· Video Channels
· Articles
· LinkedIn Official
Fact-Checking with Google Gemini
References
Footnotes
Executive Summary
We gather the rough, big-picture conclusions here on use of LinkedIn for business marketing, specifically B2B business marketing. Job hunters and freelancers seeking work will probably find more value reading the following articles: EDITORIAL: Why Hiring and Getting a Job is So Difficult Now, Business: Hiring in a Noisy Employment Market and Proactive Talent Acquisition: A Strategic Imperative.
It is imperative for those seeking work to understand what hiring business are experiencing, and what they as work seekers are up against, so they can craft individual strategies and tactics that might work. Mass applications for jobs, even with carefully tuned resumes and AI writing assistance, appear to us and other writers to be increasingly futile, and not recommended.
For marketers, correctly seeing that LinkedIn must be included in any serious marketing plan, we have the following big-picture conclusions from our own research and experiments:
LinkedIn has been slowly, but surely, morphing into an advertising platform in recent years, mostly out of necessity. While business networking and job placement are still part of the picture, the money for, and hence attention of, LinkedIn as a company increasingly focuses on advertising sales.
Conventional wisdom on use of LinkedIn for marketing, as espoused on countless web sites and videos, and detailed below, is nearly worthless for marketing purposes. There are still interesting open avenues for successful marketing, few in number and increasingly difficult to use; however, LinkedIn changes their algorithm as they discover them, specifically to frustrate and prevent successful marketing without paying LinkedIn.
Development of alternative content platforms, most notably, but not limited to, Substack, is now an imperative for marketers and content creators. Substack in particular offers a relatively rapid pathway to monetization of content, and there are newer platforms much more conducive to business networking, a topic we will cover in detail in an upcoming article.
We prepared a detailed article on the remaining and potentially successful open opportunities for successful LinkedIn marketing, and link it here now that it is completed and published. Teaser: it’s a lot of work, time-consuming and far from easy. Supporting subscribers can also consult Floccinaucinihilipilificating LinkedIn, where we playfully beat on this dead horse.
Introduction and Methodology
Introduction
On this page, we’ll collect what we know, or think we know, or what other people think they know about LinkedIn’s algorithm, important to the fraction of members with something to sell on LinkedIn, including their own services, or simply as candidate employees looking for a job. This subset of members use LinkedIn for marketing, and have a very different view of LinkedIn than more casual users who aren’t doing marketing.
For better or worse, LinkedIn is now, and will be for the foreseeable future, a inescapable and necessary part of at least the early stages of marketing a business online, particularly a B2B business. Discovery of the rules and behavior of the algorithm, as they apply to a given business, is both difficult and fundamental to eventual success.
Doing so while minimizing cost is challenging, but still possible. We feel it is inadvisable to spend money, for example on advertisements, as LinkedIn so often urges business users to do, without a thorough grounding in advertising and marketing fundamentals, and how they apply to the specific business at hand. The green marketeer should know, to enough precision, what (s)he is getting for his(her) money, and refuse to spend unless there is enough value to justify the spending.
Purpose
Our purpose in writing and maintaining this document supports two general conclusions formulated after nine months of intense marketing on LinkedIn. The first is that the conventional wisdom gathered here should be known and understood by LinkedIn marketers, in order to thoroughly understand what not to do.
The second, and perhaps more important point, is that LinkedIn exists for the benefit of LinkedIn and Microsoft, not for the benefit of its users. Every business plan that includes use of LinkedIn as part of marketing should use and abuse LinkedIn wherever possible, and then get out, possibly gradually, but certainly permanently. LinkedIn is a dinosaur awaiting the asteroid that will destroy it.
Methodology
We arrived at the principles, rules and advice contained in this article via watching YouTube videos and reading many websites, mostly not LinkedIn.com, but off-site, to get as much information about how the algorithm works as possible. A sampling of such articles and videos is linked below, with occasional commentary.
The amount of “free” advice on use of LinkedIn is very large, more than any one person can absorb. Moreover, most of the so-called “conventional wisdom” regarding LinkedIn goes stale over time, as the algorithm is under constant change.
The material in this article is a distillation of principles, assumptions, do’s and don’ts, that tend to show up in multiple, independent sources. Few, if any, of the sources cite any sources themselves nor show any serious research to back up their conclusions. For this reason, we believe all of the content here amounts to Internet rumor.
Justification
The only reason we write about LinkedIn at all, instead of our preferred technical and business content, is that we have to use it. LinkedIn is a virtual monopoly. We don’t like that, but we can’t change it. The next best thing is to document what we have found, right here, update this page periodically, and continually experiment with new and hopefully effective marketing techniques.
We add hopefully helpful resources on LinkedIn at the bottom of this page as we find them, but caution readers that, more than likely, no one person nor group of people outside of LinkedIn knows for certain what’s in the algorithm nor what changes. LinkedIn certainly isn’t talking, and their existing documentation, linked below, is wildly insufficient for marketing decision-making.
Comments from readers are invited, especially comments with quality data, since one small group of individuals can know or research only limited knowledge. Together we are far stronger, and can, with time, greatly narrow down and characterize the specific behaviors of the algorithm, track its ever-changing nature, and blaze a pathway for new entrepreneurs to make successful use of the LinkedIn behemoth.
The important point to emphasize is that we are actually and actively using the information on this page ourselves, mostly to stay within the known guidelines and rules, and secondarily, to think up unconventional ways to market our business on LinkedIn without paying them a dime. When it comes to the use of LinkedIn for marketing, this page is our bible. We publish it, free of charge, in the hopes other entrepreneurs in similar situations find it useful, but cannot guarantee suitability for any particular business purpose. In that narrow sense, each start-up business begins “from scratch” and must find their own pathway to success by trial-and-error.
The reality on the ground
Posts to LinkedIn each receive a few impressions, metered mysteriously by the algorithm. This can vary from perhaps less than ten to a little over one hundred at the beginning. Posts that receive almost immediate engagement (likes and comments) get upgraded, slightly. There are millions of active users posting regularly, generating a lot of competition for impressions.
Conventional advice holds that the prospective poster should spend regular time liking and commenting on the posts of others. There are opportunities for our fledgling poster to slip in links to his own posts and articles in comments, and we have seen some improvement in results by so doing.
That this is a very scattershot way of getting the word out about one’s business should be obvious. The newbie marketer must spend an inordinate amount of time engaging with posters, and needs to be rather careful about both comments and posts, since there is a massive diversity of humans on the platform.
If a normal learning curve prevails, it is possible, with a very large amount of work, to build up a collection of followers, climb up the learning curve, and eventually plateau, so that no amount of additional work gets results worth having. However, LinkedIn’s algorithm is constantly changing, albeit in small ways, and so we’re not certain that the normal learning curve applies.
At any rate, the LinkedIn robots control impressions, and keep them small. The decision to follow someone on LinkedIn, however, is not controlled by the robots; it is a human decision. That means, for our purposes, that followers are much more important than impressions, and so, after consistency and longevity are achieved, the ladder to climb is to collect followers.
Once a poster has 150 followers (as of this writing), he can start publishing a newsletter on LinkedIn, which we have done (The Overlogix Sunday Times), and develop subscribers, one step up from followers. The marketer should be advised that LinkedIn does not offer any direct means of monetization.
Collecting followers and subscribers in LinkedIn merely garners attention, which must be re-directed somehow to a monetizable system, such as attracting customers to one’s business or subscribers to a platform such as Substack, which does offer monetization via paid subscriptions. Such a process is long and arduous, however.
We estimate that ten impressions, minimum, are required to get one view, in other words, a human actually reading a post or an article. Out of that, we estimate a minimum of ten views to get one follower. There is similar filtering when it comes to subscribers to a newsletter; we estimate that a little less than half of followers become subscribers to a newsletter. Conversion of these to a subscriber on Substack also causes further filtering, and paid subscribers are even more rare.
While there are various ways to improve these numbers, the point is hopefully very clear: it’s a very long way to making even enough money to pay the bills. The most important friction condition is the very large number of users on LinkedIn trying to “make it”. The overwhelming majority fail.
What readers can expect to find here
Below, the reader will find various rules, do’s and don’ts, currently all marked as purported, meaning that someone, supposedly but not necessarily with relevant experience marketing on LinkedIn, has claimed the purported rule is true. We haven’t tested these points yet to see if it is true for us. There may not be any obvious, or even possible, way to test the truth or falsehood of a particular proposition, and if we find that to be true, we’ll mark the rule, later, as undecidable.
We include a section of links at the bottom of the post. N.B: This post is a work in progress, a living document, and will be maintained from time to time as we collect and test new information. Writing it has clarified our understanding of marketing greatly, so that we now have a much better idea of the challenges we face and how to address those challenges.
Updates: Just Connecting Data 2025 EOY Algorithm Study
Richard van der Blom is the founder of Just Connecting and a renowned LinkedIn expert. We saw a post of his where he offered his 2024 end-of-year LinkedIn algorithm updates, including studies done with a large number of posts, so he brings data. We include a summary of the results below.
Caveats: Imponderables and unmeasurables
While these results are definitely data-driven and about as good as one can expect from external studies of LinkedIn algorithm behavior, they don’t take several factors into account, which we believe to be extremely important to success in marketing on LinkedIn.
Those points are, in order of likely importance:
Consistency and longevity in posting are supreme. LinkedIn is a noisy marketplace with far more posters and posts than “success slots”. In essence, posting on LinkedIn is like fighting a war of attrition with millions of other posters. One carries on for years, slowly improving posting skills, hoping that eventually, the competition gives up and quits. The LinkedIn marketer outlives their competition. Over time, one builds up a following that can be gradually monetized.
Audience acceptance: there is no formula we have been able to find that reliably guides a LinkedIn marketer to growing a business on the platform. We can see from the results found so far that some users like our content, but we have to market the hell out of every article to find them. Trial and error seems to be the only way for each individual marketer to find a formula that works for his / her business, and this must be maintained indefinitely.
Ever-changing fashions: LinkedIn is social media, and like other social media platforms, fashions come and go. Likewise, styles and topics of posting change regularly. One can perhaps catch a wave as it rises by very persistent reading of one’s feed, but, as with all activities related to LinkedIn, doing this is very time-consuming and initially unrewarding. Additionally, attempting to stay fashionable, purely to attract attention, comes at a considerable cost of credibility with more serious users.
Lurkers: We, of course, have no data whatsoever on how many people passively read our posts, nor what their interest level might be. We can see and count impressions, followers and subscribers, but not lurkers. We can merely assume they are there, perhaps evaluating us for possible later business or a more active level of engagement. It behooves a marketer on LinkedIn to assume people are watching and evaluating based on posts and comments and hopefully act accordingly, however we remain in the dark and will of course continue.
Summary of Just Connecting’s study
Introduction
The combination of text and a single image in a post remains the most favored format for publishing content on LinkedIn.
The choice of format significantly impacts how prominently your content appears in your network's feed, influencing reach and impressions.
Factors like text length, subject matter, and posting frequency may vary depending on the chosen format.
Understanding the difference between views and impressions is crucial:
Regular Post: Counts how often LinkedIn displays your post in your network's timeline.
Articles/Newsletters: Tracks the number of clicks on your article to open and read.
Native Video Content: Records the number of individuals who viewed your video for at least six seconds.
While impressions can indicate content success, the ultimate goal is to foster engagement and achieve desired conversions.
Key Insights
Text + Image Posts:
Optimal text length: 900 to 1,200 characters.
Shorter sentences (under 12 words) perform 20% better.
Promotional content may see up to a 75% decrease in performance.
Personalized images can increase engagement by 45%.
Vertical photos yield a 15% higher click-through rate.
Document (PDF) Posts:
Ideal number of slides: around 12.
Optimal text length: less than 500 characters.
Slides with 25 to 50 words perform best.
Content should be optimized for mobile viewing.
Vertical layouts have the greatest impact.
LinkedIn Polls:
Polls are top performers in terms of reach.
The most effective polls offer three answer choices.
Optimal duration: one week.
Including an "Other, see comments" option can increase engagement by 25%.
Text-Only Posts:
Optimal text length: between 1,800 and 2,100 characters.
Structuring posts into brief, readable paragraphs enhances readability and engagement.
Thoughtful formatting, including white space, improves performance.
Effective for external calls to action.
LinkedIn Video Posts:
Videos are getting more reach since September 2023.
Most engaging videos: between 1 and 2 minutes long.
High-quality visuals and clear audio can lead to a 50% increase in viewer interaction.
Vertical format videos gain up to 15% more reach.
Articles & Newsletters:
Articles are valuable for SEO and enhancing newsletter content.
Optimal word count: between 800 and 1,200 words.
Bimonthly publication yields the best results.
Articles with videos or trend summaries outperform others.
Additional Insights
External Links:
Direct inclusion in the original post may result in reduced reach.
Adding links in the comments can circumvent algorithm detection but may be hidden.
Editing the post to include the link after publishing is an option if less than 15% of the content is changed.
Content Strategy:
Matching content with the right LinkedIn formats is crucial for achieving the best engagement and impact.
Optimal posting times and frequency vary depending on the content format and audience.
“Pusturing” (Publishing & Nurturing):
Engage with your network and respond to comments to increase reach.
Strategic tagging and calls to action can also boost visibility.
Maximizing Reach:
Followers are favored over connections in terms of content reach.
Instant reposts and insightful comments can amplify the visibility of original posts.
Tagging:
Tagging up to 4 relevant profiles can increase reach.
Avoid over-tagging and ensure tagged individuals engage with the post.
Hashtags:
Contribute less to content categorization but are still useful for content discovery.
Content Types:
Free Content is the most effective strategy in terms of reach and authentic engagement.
Creator Mode:
Elevates influence by prioritizing follower growth and providing access to creator tools.
LinkedIn Audio Events:
Offer unique opportunities for engagement, lead generation, and discussions.
Ideal duration: 30-45 minutes.
Network Engagement:
Dwell time and universal reach are crucial for visibility.
Engagement metrics become more predictive for viral content.
Company Page Metrics:
Company Pages did not experience the same metric fluctuations as Individual Profiles in March and September.
Content diversification and employee advocacy are key strategies for Company Pages.
Popular Topics:
Finance & Business Economics, Educational Resources, and Technology & Engineering are among the most impactful topics for personal profiles.
Company Pages benefit from a diverse content strategy.
Additional Facts & Tips:
Unified algorithm approach for mobile and desktop.
Anatomy of a Top 5% Post: around 10 paragraphs and up to 15 sentences.
Posts in English outperform those in other languages.
Disabling link preview cards can improve reach.
Reading time and clarity are crucial for engagement.
Our view on Just Connecting’s results
While the above are very nice to have and we will implement them all, we are struck by the piecemeal nature of the results, in the sense that they don’t really offer a coherent marketing methodology. It’s all tactics, not strategy, and that makes the study only somewhat better than the conventional wisdom we gathered below. Beginning marketers desperately need a strategy they can memorize and internalize; absent that, they are very much like a Marine recruit in basic training: orders shouted at them from all sides, must comply immediately, no time for reflection nor thought.
We are gradually formulating a coherent strategy for Overlogix. Generalizing that, when mostly completed, so that other marketers can make use of our paradigm to create their own marketing strategy, is a herculean task. When we formulate it, we’ll certainly link it here. Readers can expect that will likely live behind our paywall. In the meantime, trial and error is the unfortunately order of the day.
Updates: Recent Algorithm Changes
Update 31.08.2024 We add a few additional observations recently thought through.
The amount of noise on LinkedIn, in the form of many users posting frequently to the platform, has increased so much in recent years that it virtually drowns out any signal. The overwhelming majority of posts are of SPAM-quality to low-quality posts. The deluge of low quality posts considerably devalues LinkedIn as a platform.
While it is possible to prune one’s LinkedIn feed, it has to be done one post at a time, and is time-consuming. There is also a delay between pruning actions and when the changes start happening. No data on how long, but it appears to be on the order of days. We have been doing this recently with fair, but not excellent results.
The large amount of noise, coupled with lack of official documentation, complicated and ever-changing algorithm rules and emphasis on rapid engagement effectively increase the amount of time that must be spent, constantly researching LinkedIn algorithm changes, fact-checking sources, performing difficult-to-design and even harder-to-implement direct experiments. Using LinkedIn effectively for marketing is getting increasingly expensive in terms of time required.
These added costs, particularly the strong emphasis on engagement, all point to a policy change intended to incentivize users to spend more time on LinkedIn. We are of the opinion that successful marketers will increasingly find that spending less time on LinkedIn, with the eventual goal of abandoning the platform entirely, is a preferred strategy.
Update 23.08.2024: A video from March 5, 2024 details the following recent changes to the LinkedIn algorithm:
New Connection Visibility: Every connection you add will see your content for two weeks, but if they don't engage within that time, they won't see your content again. This emphasizes the importance of creating engaging content to maintain visibility, while at the same time creating a real problem for content creators. It is difficult, exclusionary, infuriating, predatory and nearly impossible to implement. One has to literally message every new connection and ask them to engage before the cutoff date. Only a small fraction will. Two weeks is too soon in the relationship for such requests for action from other parties, virtual strangers, and likely to poison the relationship if attempted. We suspect this time period selection was deliberate, and this is yet another contradiction or conundrum regarding the algorithm limiting LinkedIn’s effectiveness. Busy people don’t have time to pay so much time nor attention to LinkedIn.
Video Length: Videos should be under 2 minutes long to maximize reach and engagement. Longer videos may see decreased visibility.
View Count Decline: While view counts may have decreased, don't worry. LinkedIn's AI is now better at targeting your content to your ideal audience.
Engagement Matters: The quality of engagement, such as likes, comments, and shares, is more important than the total number of views.
First and Last Lines: The first and last lines of your posts are crucial for grabbing attention and encouraging engagement.
Editing Posts: You can edit your posts up to 15% without affecting reach. However, significant edits may lead to decreased visibility.
Reposting Content: Reposting old content is discouraged. Instead, rewrite and repurpose your content to avoid penalties.
We feel the person who created this video a strong candidate for the class of LI users who breathlessly wait for every new algorithm change so they can obey it, and fail miserably. Successful marketers must think outside the box; conventional wisdom produces mediocrity, not excellence.
Update 18.8.2024: After considerable experience using LinkedIn as a marketing tool, we have arrived at a few conclusions:
The tips, rules, do’s and don’ts listed in this article are at best ineffective, due to reliance on somehow learning how to “game” the LI algorithm, coupled with the massive amount of noise content on LI. We have determined that gaming the algorithm is nearly impossible, and a huge waste of valuable time.
New entrepreneurs need to know the “conventional wisdom” listed here, mostly to question it, and probably to avoid it. The conventional wisdom is literally what everyone else is doing; so many are doing it that any new individual doing the same things will get lost in the enormous sea like a drop of rain.
There are limited times and opportunities when one can use the “known behavior” of the algorithm in one’s favor. However, management and developers at LinkedIn are constantly monkeying with, changing, and experimenting with the algorithm. Good advice from last week might be terrible advice next week. Trying to keep up with the changes to the algorithm is possible, but extremely time-consuming, directly competing with time spent actively marketing one’s business. Together, these add up to more than two full-time jobs. We are hunting for blogs and video channels that keep track of the changes to the algorithm, and will add them to our curated sources page when found. This is one of the built-in contradictions to the algorithm we believe will eventually kill off LinkedIn’s effectiveness.
Any marketing efforts that depend on the actions of others, whether human or AI, as opposed to actions by the marketer, and dependent only on himself/herself, are statistically bound to fail. Such efforts should be immediately abandoned in favor of far more effective choices. More on this topic to come.
Google searches for topics such as “how to market a B2B service business on LinkedIn” are very likely to contain advice nearly identical to what is listed here; a little extra research, limited to the LinkedIn website, will show that most, if not all, of that advice ultimately originates somewhere in the LinkedIn documentation.
LinkedIn’s purpose in changing their algorithm and giving such bad advice is almost certainly to sell advertising, mostly to newbie marketers who lack the experience, theory and market research to understand how to use advertising without nearly 100% loss. This practice is predatory, ethically questionable at best and downright unethical and possibly illegal at worst. We anticipate progressively more successful lawsuits against LinkedIn and parent Microsoft in the future as whistleblowers increasingly sound the alarm, name names, users complain on YouTube, then complain to lawyers and government, and the last get around to backing the effort.
Some users of LinkedIn have clued in to what LinkedIn is doing, and started doing it themselves, with varying results. Consequently, we believe that every bit of advice one reads anywhere with regard to LinkedIn marketing should be viewed with healthy skepticism. Potentially reasonable and well-presented methods might be candidates for testing to determine validity. Our feeling is that any such published and successful marketing methods have a short lifetime of effectiveness, and disappear statistically long before ordinary people can use them profitably. Possible proof of this assertion can be had by reading this history and ever-changing advice proffered by these “experts”; if they say completely different things every week, they are likely scammers.
We have a separate article here detailing the likely issues with using LinkedIn as a marketing platform. Reading that article is strongly recommended. As we find effective and persistent alternate, unconventional marketing paradigms, ones that will be difficult for LinkedIn to change, we will publish them here on Substack. Some of the more effective and persistent methods may well wind up behind the paywall. See our paywall criteria for details on how we’ll make such decisions.
LinkedIn Conventional Wisdom
Our feeling is that beginning marketers must know the conventional wisdom regarding their activities thoroughly, so they can avoid doing what everyone else is doing and prevent their getting lost in a sea of noise. Only then, when they know the basic rules, can they start thinking creatively and find the little nooks and crannies ignored by the LinkedIn machine, that allow them to get their business off the ground.
Conventional Methodology
We develop a list of general principles, specific do’s and don’ts, and perhaps workable testing methodologies, to accelerate the marketing process. As a starting point, we use the principles and advice frequently presented online regarding LinkedIn, by authors we deem credible as of this writing. Clearly, the lists presented here are subject to change as a result of testing.
We’ll design tests for each point, run them, evaluate the results, and publish them, so that each point, in turn, can be proven, disproven, or declared undecidable. Repeated testing, using different tests to look at a particular point or principle from several different directions, may be employed when the results of an initial test are insufficiently decisive.
Necessity: LinkedIn is a Monopoly
At the beginning of the marketing journey, we have to include and use LinkedIn. There is no other choice. Later, we hope to progressively replace LinkedIn, even if gradually, with a diverse set of methods more closely aligned with our own business priorities and more under our direct control. More than likely, this will involve building our own robot. Other businesses will undoubtedly come to similar conclusions, and we invite collaboration in the form of commentary. More, later.
Background
Beginning marketing of any startup business is an incredibly inefficient process, filled with blind alleys and opportunities to waste time. Trial and error is the only game in town: the successful keep trying until they learn what to do and what to avoid. While there are ways to apply systematic methodology to improve efficiency, those ways are not easy to apply, and fraught with opportunities for error.
When we refer to the LinkedIn algorithm, we are talking about a very large amount of programming designed to run the LinkedIn website, with more than a billion users, 24 hours a day, 365 days a year, with as little human intervention as possible. The details of the exact algorithm are not published. LinkedIn supplies a lot of documentation, advice and suggestions, but no authoritative, detailed and definitive description of the algorithm. From outside LinkedIn, it is a black box with no windows.
For new entrepreneurs, especially B2B marketers, detailed knowledge of the behavior of this algorithm is mission-critical. We need to know what positive steps we need to do to win business, and what things to avoid, so that we have some chance of piercing through the noise to attract customers we need to grow our business. It is fair to postulate that the LinkedIn algorithm, as of this writing, is one of the main obstacles to success.
The reader should be advised at the outset that, while this post offers some critical views of LinkedIn and its algorithm, the big-brush, well-known behaviors of the algorithm are there to enhance and protect multiple income streams arising from the operation of the LinkedIn website. The programming and empirical behavior of the algorithm are therefore understandable from LinkedIn’s point of view.
The ability to understand the counterparty’s incentives and motivations is fundamental to successful marketing and business. Without this empathy, progress is nearly impossible.
The Overlogix POV
We are looking at LinkedIn from the standpoint of a start-up attempting to use LinkedIn to market our business long enough to generate income from it; in short, we’re bootstrapping the business (see the Harvard Business Review link at the bottom of the page). When income is generated, we can then justify paying LinkedIn, and others, for advertising, not before. Doing otherwise invites financial disaster, as we describe below.
The LinkedIn algorithm presents the beginning Internet marketer with some formidable challenges. The last two months of data after a lot of “spray and pray” marketing attempts are not encouraging, and, in the course of putting this article together, we estimate that the algorithm is tuned specifically to frustrate attempts to market business on LinkedIn free of charge.
This may turn out to be the key insight resulting from efforts up to now. Proving the truth of this assertion is another matter entirely. We are reminded of Deming’s dictum: “In God we trust. All others must bring data.”
Why LinkedIn?
The main reason to consider LinkedIn as a marketing platform is its reach. LinkedIn is the largest professional business website in the world, claims around one billion (1,000,000,000) users, and so, as a professional platform, it has no current (Feb. 2024), significant competition.
That being said, we wonder about several questions, which we will attempt to measure and report here. Stay tuned, as this particular post will be maintained with new updates from time to time.
Update: Followers Are Real People, Not Algorithms
10.6.2024 We abandoned all attempts to “game” the LinkedIn bot to produce more impressions after observing a slow, but steady increase in our number of followers. The light bulb moment amounted to realizing we are much more interested in the decisions of real people (as in, to follow us) than in what the robot has been programmed to do: collect money for LinkedIn. The writeup can be found here.
Questions About LinkedIn
What fraction of members with an individual LinkedIn account actually use it, and what does using LinkedIn mean? We’ll place individual users, not companies, into a few obvious categories, described next. Companies, on the other hand, can be expected to produce content that is close to 100% marketing.
The first major variable to consider when discussing individual users is whether they are active users or passive users. This property of various users can be observed and measured externally. Since this is expensive in terms of time, it should be reserved for important cases only.
The fraction of active users either post themselves or comment on the posts of others on an ongoing basis, regardless of frequency, hence practice engagement. We’ll count folks who post jobs to LinkedIn as active, even though job posting behaves differently than regular posts or articles.
Members who use LinkedIn by responding to a job posting can apply for a given job or not, but cannot comment on the job, so posting a job does not appear to allow engagement. Active use of LinkedIn, in the form of likes, posts or comments, is observable from the outside. Job applications are not observable from the outside.
The fraction of passive users just read their feeds and perhaps apply for jobs listed on LinkedIn, but neither post themselves nor comment on other peoples’ posts. They get some of their news about the professional world from LinkedIn, news that may be largely innocent of truth filters.
Passive use of LinkedIn is nearly invisible from the outside, and must be surmised (read: assumed) by looking at an individual’s posting and comment history. If their posting and comment activity is scant or blank, one can safely assume they are passive users.
The second major variable, occasionally observable from the outside, is frequency of usage of the LinkedIn platform.
Frequency of use. We count this property of users separately since it appears (statistically) independent of active vs passive usage. This is another assumption that may yield to testing. User frequency can be classified simply as:
Almost never, by which we mean at most a handful of times a year. We needn’t pick a cutoff number for exactly what we mean by “a handful”. The number appears unlikely to have value.
Rarely, more than almost never, up to regular monthly usage. Again, value is a concern with this group.
Regular monthly usage, meaning a member will at least get on LinkedIn almost every calendar month, between one and three times in a month. This subset of users forms the lower boundary of the value category.
Regular weekly usage, meaning that a user will get on LinkedIn, actively or passively, at least one weekday every week.
Regular daily usage, meaning that a user will visit LinkedIn at least three weekdays a week, and between 45 to 52 weeks per year (we exclude vacation, sick time, bereavement and crazy busy times). We are members of this category.
The frequency categories may be adjusted later after studying usage statistics; the above ten categories give us a likely home for almost all users.
We make few, if any, assumptions about the numbers in the above categories, beyond the following: passive users significantly outnumber active users, and less frequent users outnumber more frequent users. Both assumptions appear reasonable, and can be tested in the case of active users by examining their posting and comment history.
Also, a given user can, and likely does, cross the boundaries described above, becoming more or less active at different times, and more or less frequent at different periods of time. The categories implied above can overlap, but some categories are disjoint, with almost no overlap.
We suspect the important category of decision makers are concentrated in the segment of regular monthly users of the LinkedIn user base.What fraction of individual LinkedIn users are decision makers, by which we mean that they are empowered to spend money (hiring, firing, buying goods and services) either for themselves, or for the companies that employ them?
What is the experience of the average job seeker on LinkedIn?
How many applications, on average, must they submit to get an interview?
How many applications must they submit, on average, to get hired?
How often do they get ghosted after applying for a job?
General Observations From Multiple Authors Regarding Marketing on LinkedIn
Start small, with as highly specialized and specific, targeted goals as possible.
Help one or a few people, for free, from the beginning.
Engagement with people, in other words eliciting comments and online conversations, is fundamental to growth.
Attraction of attention via consistent, quality and novel content is much better than playing a numbers game, hoping to find a few customers out of the billion or so members with mass marketing. Blatant marketing is almost certain to be ignored.
Trust and credibility of content (and the content creator!) is extremely important, since LinkedIn is a very noisy, SPAM-my environment. The overwhelming bulk of posts to LinkedIn are simply ignored, since there is so much of it.
Some writers suggest not trying to game the algorithm. We disagree - at least some understanding of the algorithm is absolutely vital to prevent massive amounts of wasted labor and get at least some positive marketing traction. Gaming is probably too strong a term. It’s better to say we seek to live within the rules of the algorithm.
Only a small numbers of customers are needed at first. The goals then, in order, are:
(a) finding customers in the market for what we sell;
(b) gaining their trust by posting credible, well researched, relevant and novel material;
(c) finding the pain points for each customer, always strongly individualized and particular for each different customer;
(d) creating solutions for said pain points that are affordable by the customer;
(e) persuading the customer of the efficacy and value of the solutions;
(f) initial sale of (probably) one solution to one customer; and
(g) ongoing, meaningful, effective and prompt support of existing customers.
Repeat as needed.
Purported Rules to Use LinkedIn for Marketing
People posting about LinkedIn’s algorithm say many different things, and rarely present data to back up their claims. When data is given, it is most often anecdotal and difficult to use. Isolated statistics, often too general, given without context, are unlikely to be usable for making business decisions.
A few points are repeated often by many different authors, and we include them here, along with a few of our own observations. We are attempting to test these notions, and finding it challenging.
Post no more frequently than once every three hours. We presume this applies to an individual account. We have reason to believe that if a given user has both an individual account and a business / company account, posts to either are counted and treated separately. No proof, or relevant documentation, has been found on this fact as yet.
Hashtags in a post should be few in number, most writers recommending three: one rather general, say #technology, another specific, such as #artificialintelligence, and a third even more specific, like #artificialintelligenceforbusiness.
Posting should be consistent. The oft-repeated, over-generalized meaning of “consistent” needs sharpening and likely depends on market segment and niche. This could mean anything from once a month to several times a day. Again, statistics matter to make the notion of consistency more clear.
Off-site linking is discouraged, meaning the algorithm punishes posts with links to non-LinkedIn websites. We suspect this is true.
LinkedIn posts have an average lifetime of about 30 - 36 hours, after which they disappear from the feed. Since we have read our own posts from years ago, we believe this means the algorithm stops serving up the post after that amount of time.
The magic time for a given post is the first hour after it is published. If a post receives engagement from other users during this time, the algorithm assumes it has higher popularity and serves it up more, until it fizzles out.
SPAM-my behavior is discouraged. Posts with requests for likes or connections / following are downgraded (served less frequently). Posting too frequently (more than every three hours) causes a reset to the newest post, wasting the previous post. We emphasize that “SPAM-miness” is a subjective, not a measurable, objective property, and so the definition of SPAM-miness is arbitrary and subject to change.
Deliberately creating content with no other purpose than getting impressions, likes, engagement (do-nothing polls, repeated or over-frequent commenting, sensationalism, too-frequent posting with irrelevant, un-businesslike content) is downgraded.
The behavior of the algorithm is unpredictable, and individual posts make it through the SPAM filters with blatant violations, occasionally even of the LinkedIn terms of service. Posts that follow all known rules are often hardly served up at all. Our feed, for example, is around 80 to 90% (unwanted) marketing, emotional posts, opinion pieces, political or polemic rants. It’s expensive to spend time training the robot not to do that, so we ignore the vast majority of posts we see. So does everyone else.
The algorithm is constantly changing in small ways, making it difficult to depend on any given specific behavior, and difficult to learn the behavior of the algorithm in general.
The algorithm, since June 2023, is tuned specifically to prevent viral posts. This essentially forces all growth to be organic, meaning that posts must be promoted by “word-of-mouth” means: other LinkedIn members voluntarily (and usually, un-asked) re-posting to their own networks or groups, or sharing links. It is incredibly difficult to make this happen deliberately, and even more difficult to sustain.
Purported LinkedIn Do’s
A / B testing is recommended, including by LinkedIn, where only one property of a post is varied at a time. Doing meaningful, true A / B testing through posts on LinkedIn is extremely difficult.
Regular posting does seem to make a difference, although we have some evidence to believe that user fatigue from repeated posts, or posts with essentially the same content, sets in fairly quickly, after about one month of regular posting, sometimes faster. At the time of this writing, we cannot distinguish between genuine user fatigue and the behavior of the algorithm. Finding data to make this very important distinction is challenging.
Long-form articles appear to do better than image posting (with minimal text) by roughly a factor of 5X to 10X. So far, results have been poor, with the best performing posts showing around 100 impressions (see below for definition), around 10 views or less.
To get more engagement, ask questions in the post that readers are invited to answer.
Purported LinkedIn Don’ts
Avoid links, even limited to LinkedIn, in posts. It has been recommended that adding a comment after publishing, with links in the comment, will not trigger downgrading of the post.
Avoid anything asking explicitly, or even implying, requests for likes, connections or followers. The robot stomps on these posts, immediately classifying them as spam.
“Spray and pray” posting is mostly a waste of time. Possible exceptions include establishing brand name recognition and associating that brand name with a few key characteristics.
Our Assumptions, Expectations and Observations about LinkedIn’s Algorithm
The only advertising variables LinkedIn’s algorithm has any control over is the number of times a particular post is served, and the dates and times when it serves up the posts.
The algorithm counts impressions for a post, which is defined as a post appearing on a user’s screen for at least 300 milliseconds (30% or 0.3 of 1 second).
Paying LinkedIn (“boosting” a post) will increase the number of times a post is served. From the messages we see, included by the algorithm, this can increase servings from roughly 100X to 1000X. This is likely the single most important variable in advertising success: paying LinkedIn.
We believe the amount of boost depends on the amount of money spent to boost a particular post. To the best of our knowledge, LinkedIn has not supplied specifics on the number of times a post is served versus money spent.
We have not yet tried spending money to boost posts, since we believe we should have some idea of where, what and how scarce dollars should be spent prior to trying it. Getting good (statistically significant) data on how to do that appears to be systematically discouraged.
A quick survey of recent results shows that un-boosted image ads with minimal text yield very low impressions, in the low teens, and are almost worthless for business promotion.
The same survey showed average view rates, where users actually read posts, hovers around 10% or less compared with impressions, so that it takes ten impression to yield one view.
A fraction of image + text posts, probably less than 10%, get any engagement at all, and those that do get only likes, and less than ten of those.
A negligible number of posts, out of hundreds posted so far, received comments.
The algorithm’s rules are strongly biased against beginners who don’t pay LI. Finding a bootstrap pathway to success, even when religiously following the commonly expressed rules, without paying LinkedIn ruinous amounts of money, appears very unlikely.
A Few Caveats About Content Linked Here
The vast majority of articles on the subject of LinkedIn marketing, outside of LinkedIn’s website, are there to hawk some software or website, usually at a cost, implying the software will somehow magically solve the bootstrap problem. * We are going to need a lot of quality data and a lot of persuasion before buying into any such software.
We doubt such software can help much, and believe spending the time with LinkedIn’s site (admittedly a not-so-great UI/UX) and lots of googling, is necessary to know how to find available information on the site and about the algorithm.
There are (again, poor quality) analytics available, and every effort to get data, information and knowledge for free should be exhausted prior to paying for anything.
NB: You have to know what you are doing in order to spend ad money effectively, and a few principles, tested with a lot of direct experiments, is the only known good way to go. The alternative is paid “spray and pray” marketing, with extremely high failure rates. No, thank you, LinkedIn, at least not yet.
Individual tidbits of knowledge can be gleaned from many different websites (but remember the first bullet point in this section), perhaps some of which can be tested and verified.
Professional writers are not professional marketers. They go to marketing websites for research, talk with marketers, or echo-chamber what other professional writers write. Many of the sources below are merely professional writers. They rarely know what they are talking about, and rarely conduct direct research themselves. †
The best source on LinkedIn’s algorithm is LinkedIn itself, their official posts and permanent help pages, also linked below. For the most part, they (LinkedIn) aren’t talking, instead offering difficult-to-follow generalities. Only direct research, such as repeated, extremely careful, A / B testing, can eventually home in on the (more or less) exact rules used by the algorithm, and then, only as it relates to the specific business conducting the tests. We are attempting to design quality A / B testing, and finding it difficult.
A Collection of Video Channels and Articles About LinkedIn and its Algorithms
This set of links should be viewed as starters for a much more in-depth, personalized and extremely specific marketing plan, to be formulated and gradually, step-wise refined, typically with loads of trials and testing. One should always read with caution, and be attentive and sensitive to understanding the agendas each writer holds and why they are writing such promotional material.
For example, ask whether the writer of the article is simply a professional writer, not a marketer. Some output of purely professional writers is linked below, so that readers can find some of the commonly believed notions regarding the LinkedIn algorithm. These have been marked with a dagger (†).
We assume everyone creating videos or writing publicly about LinkedIn has an axe to grind. Readers should, too. Our axe is simple and explicit: we’re selling Overlogix as a service business and a source of knowledge about our specialties.
It should also be pointed out again that LinkedIn’s algorithm is a moving target, that they don’t supply much in the way of details about how it works, and active users of LinkedIn, such as us, only find out about algorithm changes that have already happened, and after a time lag. LinkedIn is always changing.
We recommend marketers develop written strategies for themselves that work with the business they are conducting, ones that depend as little as possible on aspects of the algorithm that can change. We also recommend regular, ongoing research into the algorithm and how it changes. Some resources we have found useful are detailed below. As always, use your own independent judgement, exercise caution, do your homework and keep trying.
Video Channels
Your LinkedIn Business Strategist Salina Yeung’s YouTube channel, mostly devoted to marketing on LinkedIn, and in looking at the algorithm in particular. She gives numbers, multiples, etc., and so we like this.
Brooke Miles She’s a LinkedIn strategist, and makes lots of videos on LinkedIn. We find her style a bit hard to watch, but there is a lot of practical advice, which so far fits in with what we have seen elsewhere. She also advises against attempting to game the algorithm. Since the LinkedIn algorithm changes rapidly enough to defy writing a definitive rule book (including this article you are now reading), we advise at least some caution. This kind of advice works until it doesn’t, and knowing when to abandon a particular principle or piece of advice is difficult, usually understood only after repeated failure due to an algorithm change. We only know LinkedIn’s rules in our rearview mirror; a huge reason to seek alternatives as soon as possible.
Mark Firth This savvy and intelligent fellow does a lot of videos about LinkedIn, and there are plenty of clues regarding the algorithm mixed in with his current advice about working around it. His stuff is good, but watching enough of his videos reminds people of just how rapidly LinkedIn’s algorithm changes. Our central point in the above article should always be kept in mind. Additionally, Mr. Firth’s content goes very fast and at a very high level; using it requires some fairly serious study and instantiation to one’s own business. It is also obvious that he has a business axe of his own to grind, one that takes precedence over helping his viewers. Caveat Emptor!
Articles
Jack Shepherd: 41 Essential LinkedIn Statistics You Need to Know in 2024
Oleksii Bondar: Important LinkedIn Statistics Data & Trends
Courtney Johnson: The Linkedin Algorithm Explained + How to Make it Work for You [UPDATED 2023] †
Hootsuite: How Does the LinkedIn Algorithm Work? [2024 Changes Explained] *
Tamilore Oladipo: How LinkedIn’s Algorithm Works, According to the LinkedIn Team †
Grace Collyer: How the LinkedIn algorithm works in 2023
Brent Barnhart: How the LinkedIn algorithm works in 2024 *
John Espirian: How to promote multiple businesses on LinkedIn. Overlogix: Espirian’s articles tend to be fairly good, and can help the narrowing-down process if followed. This article doesn’t actually mention the algorithm, but does offer good advice to effectively harmonize with it.
Ryan Smith (Overlogix: Harvard Business Review, about startups, not LinkedIn, this is an important article!): Why Every Startup Should Bootstrap
Karen Tisdell: Understanding the LinkedIn Algorithm This article contains a copy of the LinkedIn algorithm content filtering diagram. We’ll link the actual LinkedIn page with this diagram when we find it. It’s another item LinkedIn marketers need to know cold. She gives good, if conventional, advice.
LinkedIn Official
LinkedIn Professional Community Policies This is as close of a rulebook as we are going to get out of LinkedIn. All active users of LinkedIn should read this page several times and re-read it periodically to catch changes. N.B.: This is not enough detail here, by a long shot, to understand how to use LinkedIn successfully to market a business, but these policies give a list of don’ts. We’re especially interested in the SPAM section. We quote:
We don't allow untargeted, irrelevant, obviously unwanted, unauthorized, in appropriately commercial or promotional, or gratuitously repetitive messages or similar content. Do not use our invitation feature to send promotional messages to people you don't know or to otherwise spam people.
There’s a lot of subjectivity in those two sentences, and new content of any kind gets processed immediately by part of the algorithm, which dumps the content into one of three buckets: SPAM, low quality content, and high quality content. To get any reach at all, marketers need to understand these three categories as understood by LinkedIn.
LinkedIn HQ, Sunnyvale, California Just in case readers haven’t noticed, California tends to march to a different drummer than the rest of the country, in fact to the rest of the world. How comfortable are readers with denizens of this strange, mad place deciding, for them and without invitation nor permission, the meaning of SPAM and quality of content?
LinkedIn Help Pages Notice how utterly sparse this page appears, and for a website with more than one billion users. Amazing!
LinkedIn Help: Basics Examine this page carefully, for the real story is what is not there.
LinkedIn Official Blog Last entry: Feb 22, 2023. Hmmm.
LinkedIn official: Mythbusting the Feed: Helping our members better understand LinkedIn (Part 1) This article, from 2022, is written by LinkedIn’s VP of engineering and contains videos explaining “What kind of conversations are welcomed on LinkedIn?” and “What does it mean to be professional when it comes to content on LinkedIn?”
LinkedIn official: Mythbusting the Feed: How the Algorithm Works (Part 2) This article, from 2022, explains, with videos from the LinkedIn VP of engineering, how the LinkedIn feed works. Note that major changes to the algorithm happened in June 2023, so this article counts as background material.
LinkedIn official: Mythbusting the Feed: How We Work to Address Bias (Part 3) Another article from 2022, this one largely about LinkedIn’s efforts to implement DEI. It would be fascinating to find out what LinkedIn thinks about this, now that DEI is increasingly being called into question.
An example of an algorithm change, from December 2023. Carousels forbidden, except for paid accounts. Why?
Interestingly enough, we have been unable, despite multiple attempts, to find official LinkedIn documentation or commentary regarding their own algorithm later than the above three, and absolutely nothing since the large-scale algorithm changes of June 2023. We’ll continue searching.
Plenty of people outside LinkedIn have published content on the subject; absent official endorsement of any of these by LinkedIn, we place those in one of the two sub-categories (Video Channels, Articles) above. We’re not interested in putting words into LinkedIn’s mouths, we want to know how the algorithm works and why.
LinkedIn official: Marketing solutions Read with extreme caution. The purpose of this site is to sell you their stuff.
LinkedIn official: Get started with LinkedIn Ads Read with extreme caution. The purpose of this site is to sell you their stuff.
LinkedIn official: LinkedIn Marketing Partners: Read with extreme caution. The purpose of this site is to sell you their stuff.
LinkedIn official: Presenting LinkedIn Collective – A Community Built for B2B Marketers Read with extreme caution. The purpose of this site is to sell you their stuff.
LinkedIn official: A/B Testing Overlogix: We will have much more to say about this topic in the near future, since there is no doubt that A / B testing (or simple extensions of the idea) remains a crucial way to narrow down an individual’s business choices from nearly infinity to the tiny list every marketer really needs. So far, from pondering the rules presented in this article, real A / B testing appears very challenging to do properly.
Fact-Checking with Google Gemini
Since Gemini is more or less continuously updated, we decided to run this article through it for a fact check. Now, as persistent and frequent users of AI, we’re painfully aware of its faults, and have written about them previously. However, we think enough of its typical output to include it here. We were looking for any glaring falsehoods that might have slipped past our own, relatively vigilant BS filters. Here are the results, make of it what you will.
Overlogix: … I didn't need summaries yet. We'll get to that, soon. First, fact-check the article, and list any points that are questionable or false. Ignore true statements in the article. Include evaluations of questionable claims as (1) likely opinion (2) possibly false (3) probably false (4) definitely false.
Gemini: Fact-Checking the LinkedIn Algorithm Article
Note: While this analysis is based on the information provided in the article and general knowledge of LinkedIn, it's important to remember that the LinkedIn algorithm is constantly evolving, and specific details may change over time.
Questionable or False Statements:
"LinkedIn has been slowly, but surely, morphing into an advertising platform" (likely opinion): While it's true that LinkedIn has increased its focus on advertising in recent years, it's still primarily a professional networking platform. The extent to which it has "morphed" into an advertising platform is debatable.
"Conventional wisdom on use of LinkedIn for marketing...is nearly worthless" (possibly false): While some conventional wisdom may be outdated or ineffective, there are still valuable strategies that can help you succeed on LinkedIn.
"LinkedIn is a dinosaur awaiting the asteroid that will destroy it" (likely opinion): While LinkedIn has faced challenges, it remains a dominant platform for professional networking. Its future is uncertain, but it's unlikely to disappear completely.
"The algorithm is tuned specifically to prevent viral posts" (possibly false): While LinkedIn may prioritize quality content and engagement over virality, it's unlikely that the algorithm is specifically designed to prevent viral posts.
Additional Considerations:
The article's claim that "LinkedIn's algorithm is a moving target" is generally accurate. Algorithms are constantly updated and improved, so strategies that work today may not work tomorrow.
The article's emphasis on the importance of "engagement" and "quality content" is consistent with current understanding of the LinkedIn algorithm.
The article's advice on avoiding spammy practices and focusing on building relationships is sound.
Overall, the article provides a valuable overview of the LinkedIn algorithm and offers practical advice for marketers. However, it's important to remember that the algorithm is constantly evolving, and what works today may not work tomorrow.
Overlogix: OK, we’ve read statements and watched videos from LinkedIn staff about preventing viral videos. When we find them again, we’ll add them to the references above. We’re rather confident this statement is true.
Points 1. and 2. are admittedly opinions, based now on nine months of increasingly targeted marketing, but tend to match the more perceptive of outside commentators; see the channel of Mark Firth above, who is fairly adamant that conventional wisdom with respect to LinkedIn is a road to nowhere.
Point 3. was delivered with tongue firmly in cheek. Long range predictions of something so complicated are, of course, the bread and butter of pundits, and, like all prophesy, very hard to verify. If we’re being something of a canary in a coal mine here, we stand by it, and strongly emphasize that beginning entrepreneurs and small business marketers should develop newer and better platforms for their business marketing, and in the meantime use LinkedIn where they can, until the effectiveness disappears. Of this last, we remain confident: LinkedIn is increasingly focused on paying customers. Everyone else is product.
References
In-Depth Analysis of LinkedIn Content Performance
LinkedIn Ads benchmarks for 2024 This article covers performance of LinkedIn advertising for various regions, including costs and average results. Of interest to companies paying for LI advertising only, not for bootstrappers.
Why is LinkedIn consistently slow to load pages? This Quora chat covers LinkedIn’s persistent slow page loading and performance problems. We’re experiencing the same thing, and it is maddening.
Understanding the Linkedin Algorithm in 2024
How the LinkedIn algorithm works in 2024 This in-depth article is one of the best we’ve found so far, and is a recommended read, especially for B2B marketers.
Footnotes
Thank you for reading this article!
More information about Overlogix can be found at Welcome to Overlogix!
We currently publish on both LinkedIn (general interest articles, summaries, TL;DR’s: easier and faster to read) and Substack (in-depth articles, how-to’s, technical studies and new approaches to business).
LinkedIn:
· Introduction: Welcome to Overlogix!
· The Overlogix Sunday Times Our newsletter, with occasional specials, published roughly every two weeks.
· Master Index All our articles can be found from here in two clicks.
· The Overlogix Table of Context All Overlogix articles in reverse chronological order
· Applied Artificial Intelligence: Index of Articles One of our specialties is Applied AI. This index lists all relevant articles on the topic, in reverse chronological order.
· Applied AI: Stories in the News Our semi-permanent, curated listing of interesting and important news from the world of artificial intelligence, from many different sources.
· Index: Getting a Job Up until recently, getting a job, much less a good job, has been a nightmare for most job seekers. We publish articles on how and why this is so, and what job hunters can do to find the perfect job for them. We also supply credible external resources, so people can consider their alternatives.
· Starting a B2B Business For everyone who can, we heartily recommend starting your own business. The tools are there, and there has never been a better time to do it.
· Building Our Own Robot We’re automating Overlogix from the start, and this series of articles tells exactly how we are doing it.
· Rebuilding the Linux Server: Index of Articles Running AI on your own machine (recommended) requires a modern, up-to-date operating system, and often a lot of additional software infrastructure. This series, dedicated to exactly that sort of system administration, details what we have done to build a powerful server that runs both databases and artificial intelligence, locally.
· The Gospel According to ChatGPT Conversations with various AIs and additional articles on the various challenges associated with actually making profitable use of artificial intelligence.
· TL;DR: Index of Fast Reads Brief, fast reads on various topics in artificial intelligence. If you are a beginner at AI, or a busy human needing fast and factual explanations of complicated technical topics, this is the place to start.
· TL;DR: Overlogix Artificial Intelligence Mini-Wiki Same Fast Reads as previous but arranged in a mini-wiki format some folks may like better.
Substack:
· Welcome to the Overlogix Substack
· Overlogix: Table of Context Index to our Substack articles arranged by topics.
· Criteria for Paid Content Rules for what goes behind our paywall.
· Curated IT and AI Sources Annotated links to sites and YouTube channels we think are valuable.