Most founders think about followers the wrong way round. They worry about how many they have, and almost never ask how LinkedIn decides who sees their name in a list of people worth following. Yet that list is one of the main ways a buyer who has never heard of you ends up following you.
In August 2026, LinkedIn's engineering team published a detailed account of how it rebuilt that system. It is technical, but the implications for a founder are practical and fairly simple. This guide translates it, adds what LinkedIn has said elsewhere about profiles and the feed, and turns it into a short test you can run on your own profile today. All LinkedIn sources were read on 18 November 2026. If you want the wider question of whether followers matter at all, our guide to how the LinkedIn algorithm works is a good companion.
What LinkedIn published about follow recommendations
The source is a LinkedIn Engineering Blog post dated 24 August 2026, "Rebuilding LinkedIn's Follows Recommendations with LLM-Based Semantic Retrieval and Ranking" (LinkedIn Engineering). Here is what it says, in plain English.
- What the system does. It powers suggestions in the MyNetwork tab and the Home Feed, and LinkedIn says it "has one job: match members with creators whose expertise, perspective, and content they're likely to find valuable" (LinkedIn Engineering).
- What was wrong before. The old system relied on "heuristic candidate generation and popularity-driven ranking". LinkedIn says that was "good enough for popular, obvious recommendations" but "overlooked strong matches that happened not to be popular" (LinkedIn Engineering).
- What changed. LinkedIn now writes each member's and creator's profile "out as a short piece of text", passes it through a large language model, and places similar profiles close together, so it can "find relevant creators and rank them by what they're about rather than by how popular they are" (LinkedIn Engineering).
- What it reads. The post lists profile fields "including bios, headlines, taglines, skills, and experiences", and in its conclusion says the prompts today "rely on profile fields like headline, title, and skills" (LinkedIn Engineering).
- Did it work? LinkedIn reports that in production A/B tests the new pipeline "delivered statistically significant lifts in follow rate across member segments, with the strongest gains for new members" (LinkedIn Engineering).
What this means in plain terms
Picture two lists. One is every member on LinkedIn, each described in a short paragraph built from their profile. The other is every creator who can be recommended, described the same way. The system looks for creators whose paragraph sits close to a member's paragraph, meaning they are about the same professional world, and suggests them.
For a founder, three things follow from that.
- Your profile text is your pitch to the recommendation system. If your headline says 'Founder and CEO' and nothing else, there is very little for the system to match on. If it says who you help and with what, there is.
- Your buyers' profiles matter as much as yours. Matching works between your profile and theirs. The closer your words are to the words your buyers use to describe themselves, the easier the match.
- Popularity is no longer the whole story. A founder with a modest following and a clear, specific profile can be suggested to exactly the right people. That is good news for founders in narrow B2B niches.
How it connects to the feed
Follow recommendations are not the only place LinkedIn matches people by professional context. Its feed work points the same way.
- Profile Relevance. In a 9 November 2026 post, LinkedIn's engineers describe a feed system that builds a "member prompt" from profile highlights ("headline, summary, and key skills"), work history and education, and matches it with post embeddings generated "from post text, titles, and transcripts" (LinkedIn Engineering, Profile Relevance).
- Beyond your network. LinkedIn's feed engineering post says the aim is "discovering content from authors you don't follow, on topics you care about" (LinkedIn Engineering).
- Not a popularity contest. In March 2026 LinkedIn said: "We want the Feed to be relevant to your interests, not a popularity contest" (LinkedIn Pressroom).
Put together, the message is consistent. LinkedIn is trying to match people on what they are about. Your profile tells it what you are about, and your posts confirm it. Our profile optimisation guide covers the field-by-field fixes; this guide is about why those fields matter for who gets suggested to you as a follower.
What LinkedIn didn't say
It is worth being careful here, because posts like this get over-read.
- It doesn't publish weights. LinkedIn does not say how much any single field counts, or how often a given creator is shown.
- It still uses other signals. The new retrieval "runs alongside our existing graph- and popularity-based generators", so your network and popularity still play a part (LinkedIn Engineering).
- Activity signals are a next step, not today. LinkedIn lists "recent content engagement, browsing activity, and connections" as signals it could add to the prompts in future, which means they aren't described as part of these prompts yet (LinkedIn Engineering).
- It doesn't give follower numbers. The results are reported as statistically significant lifts in follow rate, not as figures you can benchmark against.
The Narrative Profile Test
This is our own exercise, built on how LinkedIn describes the system. Since LinkedIn writes your profile out as a short piece of text, do the same yourself and read it the way a matching system would.
- Copy your headline, the first two lines of your About section, your current title and company, your top three skills, and the first line of your current role's description.
- Join them into one paragraph, in that order, without editing.
- Score the paragraph on three questions, 0 (no), 1 (partly) or 2 (clearly).
- Known for: would someone reading only this paragraph know the one problem you are known for solving? Five different topics score 0.
- Known to whom: does it name the type of buyer you help, by role or industry, in words those buyers would use?
- Consistent: do the headline, About, title, skills and role description all point at the same topic, or do they tell different stories?
A score of 5 or 6 means the system has something clear to match on. 3 or 4 means you are partly visible to the right people. 0 to 2 means you are probably being matched on job title and network alone.
Add the Buyer Mirror check
Now open the profiles of five people who are exactly the kind of buyer you want following you. Write down the words they use in their own headlines and job titles. Compare them with your paragraph. If they call themselves 'Head of Revenue Operations' and your profile only talks about 'go-to-market excellence', you are describing the same world in different words. Match their language where it is accurate to do so. Our content strategy guide shows how to find the same words in your buyers' questions.
Should a founder make Follow or Connect the main button?
Follow recommendations bring strangers to your profile. What they see when they land is the next decision.
- The difference. LinkedIn explains that "Following lets you learn from someone's content, while connecting builds a two-way relationship", and asks members to "Only send invitations to people you know and trust" (LinkedIn Help, follow and connect).
- The setting. You can choose who can follow you (everyone, or only your connections) and turn on Make follow primary to show Follow as the main action on your profile (LinkedIn Help, a528011).
- A trap to avoid. If you limit followers to your connections, LinkedIn says "followers outside your network will be removed", and "there will no longer be a Follow button next to your posts or articles" (LinkedIn Help, a528011). For a founder who wants buyers they haven't met to follow along, that setting works against you.
- Connect is still there. If Follow is primary, people can still connect via the More menu (LinkedIn Help, a541669). LinkedIn also notes that when Follow is your primary action, someone who sends you a connection request follows you while they wait for your response (LinkedIn Help, a5999182).
Our rule of thumb: if you post at least weekly and want buyers you have never met to keep up with you, make sure everyone can follow you and consider making Follow primary. If most of your LinkedIn activity is staying in touch with people you meet in person, Connect as the main button is perfectly reasonable. Either way, check that Who can follow you is set to everyone.
A worked example: one founder, before and after
This is an invented example, not a client. A founder runs a nine-person procurement consultancy that helps mid-sized manufacturers cut supplier costs. Her Narrative Profile Test paragraph reads: 'Founder and CEO. Passionate about people, growth and innovation. Founder at her firm. Skills: leadership, strategy, public speaking.'
- Known for: 0. Nothing says procurement or supplier costs.
- Known to whom: 0. No buyer is named.
- Consistent: 1. Everything is consistently vague.
- Buyer Mirror: her target buyers call themselves procurement directors, CFOs and heads of operations at manufacturing firms. None of those words appear in her profile.
She rewrites the headline to say she helps mid-sized manufacturers cut supplier costs, opens her About section with the problem in her buyers' words, updates her skills to procurement, supplier negotiation and cost reduction, and adds one plain sentence to her current role describing the work. Her new paragraph scores 6. Nothing about her follower count changed that day. What changed is that a matching system reading her profile now has a clear answer to the question it is trying to answer: who is she relevant to?
How to tell whether it's working
You can't see follow recommendations directly, but post analytics give you useful clues. LinkedIn reports "Followers gained from this post" and post viewer demographics, which "may include the job title, location, company, company size, industry, and seniority of your post viewers" (LinkedIn Help, post analytics).
- Are new followers in your buyers' roles and industries, or mostly peers and other creators?
- Is the share of viewers outside your network holding up as you post on your core topics?
- Are buyers you have never met mentioning your posts on calls?
Track these monthly rather than weekly. Our post performance table and our guide to measuring LinkedIn content ROI show how to log them without drowning in numbers.
What not to do
- Don't chase follows with engagement swaps. LinkedIn's policies say not to "artificially increase engagement" with your content (Professional Community Policies), and our explainer on engagement pods covers why they backfire.
- Don't use tools to follow or connect in bulk. LinkedIn does not permit software that automates activity on its website (LinkedIn Help, a1341387).
- Don't stuff your headline with keywords. The system reads meaning, not a list of terms. A clear sentence beats ten buzzwords.
- Don't change your positioning every month. Consistency between your profile and your posts is what makes the match clear.
Where Influx fits
Influx does this for founder-led B2B firms, done for you. We capture your expertise in conversation, tighten your profile so it says clearly what you are known for and for whom, write posts in your voice on the same core topics, and report on which buyers are engaging. Distribute. Getting each post in front of the buyers you want to reach, not just your existing network. Distribution is included, with no ad spend. See how the Influx Authority Engine works and who Influx is for. Our founder, Alex Alleyne, was named a LinkedIn Top Voice in Sales in 2019 and 2020. We guarantee 30 qualified sales calls in 90 days. If we don't hit it, we keep working at no cost until we do. We only guarantee what we control: the right buyers on your calendar. Closing is yours. (The Influx Guarantee Explained)
About Influx
Alex Alleyne is the founder of Influx, a two-time LinkedIn Top Voice with over 60,000 followers on LinkedIn. Influx is a done-for-you LinkedIn authority system for founder-led B2B firms that turns a founder's expertise into qualified sales calls, with content written in their voice and distributed to their buyers. Influx, the LinkedIn authority firm, is operated by Fetched L.L.C-FZ in Dubai, UAE, and is not affiliated with other companies called Influx.
Want a profile and a posting plan that put you in front of the right buyers, not just more of them? Book a call.

