How to build a LinkedIn audience from zero

To build a LinkedIn audience from zero, decide who you are writing for in one sentence, find the 20 to 30 accounts that person already reads, and spend your first two months commenting there rather than waiting for your own posts to travel. Publish two or three times a week at the same time so the profile a stranger lands on has something on it. The cold-start problem on LinkedIn is not that your writing is bad, it is that distribution is network-shaped: posts move through people who already follow you, and at zero followers there is nothing for the system to move them through. Borrowing someone else's audience is the only mechanism that works before you have one.

What is a LinkedIn audience, and how is it different from followers?

A follower count is a number. An audience is a group of people who would notice if you stopped. Those two things drift apart constantly, and the drift is the reason a lot of accounts with impressive numbers produce nothing at all.

LinkedIn makes the confusion easy. Every connection also follows you by default, so the counter starts filling with colleagues from an old job, people you met at a conference in 2019, and recruiters who were hiring for a role you did not take. None of them are reading you. They inflate the denominator on every metric you look at, which is why so many people conclude their reach has collapsed when what actually happened is that their audience was never the same shape as their follower list.

A more useful definition. Your audience is the set of people for whom the following is true:

• They recognise your name when it appears above a post.
• They have some idea what you write about before they read the first line.
• They would think of you if the subject came up in a meeting.

At the start that group is small enough to name. Fifteen people is a real audience. Forty is a very good one. This is not a consolation prize framing: fifteen people who hold the job title you want to reach will do more for you than four thousand strangers, because the second group produces impressions and the first group produces conversations. We work through what happens after that in the piece on LinkedIn lead generation.

The follower milestone is a separate project with separate arithmetic, and we cover it directly in how to get your first 1,000 followers. This article is about the system underneath the number: who you are writing for, how you find them, and how you get in front of them before you have any reach of your own.

Who exactly are you writing for?

Most people cannot answer this, and the symptom is obvious in their posts. The writing addresses everybody, which means it addresses nobody, and it reads like a company blog that escaped.

The exercise takes ten minutes. Finish this sentence with one type of person, not a market:

I am writing for [specific role] at [specific kind of organisation] who is currently dealing with [specific problem].

Good: a head of finance at a 200-person software company who is closing the books manually and knows it is going to break. Bad: B2B decision makers. Bad: anyone interested in leadership. The difference is that the first version tells you what to write on a Tuesday morning and the second does not.

Two things people get wrong at this step.

They narrow the subject instead of the reader. Picking one reader does not mean writing about one topic forever. If you write for that head of finance, you can write about hiring a first analyst, choosing tools, what a bad month-end feels like, the argument you lost with your CEO, and what you got wrong in your first year. Five different subjects, one reader. Accounts that narrow the topic instead run out of material in six weeks and start recycling.

They pick the buyer instead of the reader. Sometimes the person who signs is not the person who reads. If your buyer is a CFO but the person who feels the pain daily is a controller, write for the controller. They are the ones who will send your post to the CFO with a line on top, and an internal recommendation beats anything you can publish directly.

Write the sentence down and keep it somewhere you will see it. Every future decision about what to publish is really a question of whether this specific person would stop scrolling for it.

Why does LinkedIn have a cold-start problem?

Because distribution is multiplicative and you are starting with a multiplier near zero. A post is first shown to a slice of the people already connected to you. What that slice does determines whether it travels further. With 90 connections, the first slice is a handful of people, and a handful of people is not a large enough sample for anything to compound. There is nothing being suppressed. There is nothing to amplify.

This produces a genuinely circular problem. You need an audience to get reach, and reach is how most people expect to build an audience. The circle only breaks in one place: attention that someone else has already assembled. In practice that means comment sections, and it is why the advice in this article looks so different from the advice given to accounts that already have 20,000 followers.

Worth knowing about the mechanics: LinkedIn retired the creator mode toggle in 2024. In its help documentation on the change, LinkedIn says the on/off toggle was removed and that members can choose whether Follow or Connect is the primary button on their profile. That is a small setting with a real effect on audience building. Follow as the primary action lets a stranger subscribe to you without asking your permission, which is exactly what you want someone to do after reading one good comment.

How do you find the people you want in your audience?

You do not find them one by one. You find the rooms they are already in and stand in those rooms.

Practically, that means building a list of 20 to 30 accounts your reader already follows. Not 200. The list has to be small enough that you can read it properly every morning. How to assemble it:

1. Start with three people you know your reader reads. Ask two customers or two peers who they follow. The answers are usually the same handful of names.
2. Read the comment sections of those three. The people leaving good comments are either your reader or your competition for their attention. Both are useful. Add the good commenters to the list.
3. Mix the sizes deliberately. A few very large accounts for reach, and more mid-sized ones where your comment is one of nine rather than one of four hundred. On a 5,000-follower account the author will actually read what you wrote, which is how the relationship part of this works.
4. Add the practitioners, not just the influencers. The people doing the job your reader does are often better rooms than the people commentating on it, because the comment sections contain fewer professional networkers.
5. Cut anyone whose posts you have nothing to say about. If you cannot add anything to a post, that account is not a room for you, however large it is.

Then turn on notifications for the smaller half of the list. Being early in a comment section matters more than being clever in it, because the replies that arrive first are the ones everyone else reads.

Why is commenting the cold-start mechanism?

Because it is the only free way to appear in front of people who have never heard of you, in a context where they are already paying attention.

Think about what a good comment actually is. It is a short post, published inside someone else's distribution, to an audience that self-selected into the topic, with your name and headline attached. If you wrote the same 60 words as your own post at 200 followers, a few dozen people would see it. Written as a comment on a busy post, it can be read by a meaningful slice of a much larger group.

The part people miss is that a comment is an audition, not a greeting. Three things make one worth leaving:

• It adds information the post did not have. A counter-example, a number, a case where the advice breaks, the thing that happened when you tried it. Agreement adds nothing and is invisible.
• It stands alone. Someone should be able to read your comment without reading the post and still get something. That is what makes it quotable and what makes people tap your name.
• It is specific enough to be wrong. Vague comments are safe and forgettable. A comment that takes a position invites a reply, and a reply is what puts you back at the top of the thread.

Length is not the variable. Two sharp sentences beat six padded ones. What matters is whether a stranger reading it concludes that you know something they do not. The practical version of this, including how many to leave and how to keep it from eating your day, is in how to use comments to grow on LinkedIn and how many comments to leave per day.

Our data is not a study of commenting, but it does say something about why comments are the currency worth trading in. Across 12,988 posts, the median top-decile post earned 72 comments against 5 for the median bottom-half post. Reactions differed by far less: 235 against 80. Comments are the scarce thing. Giving them away generously is therefore the cheapest valuable thing you can do for another creator, and the most likely to be noticed.

What should you post while your audience is still small?

Post the things that will be useful to a stranger reading your profile in month four, because that stranger is the actual audience for month one. Almost nobody sees these posts in the feed. They are read later, in a batch, by someone deciding whether to follow you.

That reframing changes what is worth writing. Four kinds of post do the job:

1. The thing you explain on calls every week. If you have said it out loud five times, it is a post. It is also the post most likely to make a reader think you do this for a living.
2. The opinion you would defend in a meeting. Something a reasonable person in your field could disagree with. Universally agreeable posts get read and forgotten.
3. The thing you got wrong. First-person openers appear about twice as often in our top decile as in the bottom half (19.6% against 10.1%), and a specific mistake is the easiest first-person post to write without sounding pleased with yourself.
4. The question you keep getting asked. If three clients asked it, three hundred readers are wondering it.

Two things to skip in the first hundred days. Reshares of other people's articles with a sentence of agreement on top, which over-index about 1.5x in the bottom half of our data, and personal milestones with nothing transferable in them. Both are easy to publish, which is exactly why the feed is full of them.

For the mechanics of the words themselves, the opening lines carry most of the weight, because a reader decides in about two seconds. That is covered in how to write LinkedIn hooks.

What does the data say about which subjects build an audience?

Our segment data holds subject matter constant, which is the useful way to read it: not “is leadership a good topic” but “what separates the best leadership posts from the weakest ones”.

In the leadership segment (2,117 posts from 53 accounts), that topic's own top decile (211 posts, 14 authors) says you somewhere in the visible hook 52.6% of the time, against 30.4% in its own bottom half (1,059 posts, 33 authors). Native video runs 36.5% against 7.9%. In career growth (1,389 posts from 50 accounts), the inner top decile (138 posts, 19 authors) uses second person in 64.5% of hooks against 34.4% in the inner bottom half (695 posts, 33 authors).

The pattern repeats across segments with enough authors to trust: the posts that build audiences are addressed to somebody. That is the same instruction as “pick a reader”, arriving from a different direction. When you have written the sentence about who you are writing for, saying you in the hook stops being a stylistic tic and becomes accurate.

One caution on reading segment tables. These compare the best posts in a topic against the weakest posts in the same topic, which controls for subject matter but not for author. Some of the gap is that better creators write better hooks and also happen to use second person. It is a correlation worth acting on, not a mechanism we can prove.

How do you know whether you are building the right audience?

Look at who shows up, not how many. LinkedIn gives you more of this than most people realise. Its post analytics documentation lists demographics of your viewers by job title, location, company, company size, industry and seniority, alongside impressions, members reached, profile viewers from the post and followers gained from the post. LinkedIn also states plainly that the numbers are estimates and that demographics only appear once there are enough unique viewers to protect privacy.

Use that panel as a compass rather than a scoreboard. Four questions it answers:

The last row has no dashboard and is the most informative of the four. Keep a note of who comments. When the same five or six names appear across consecutive posts, you have an audience rather than a series of unrelated reach events. That transition usually happens well before the follower count looks like anything.

The metric that misleads most is impressions. A post that goes wide outside your niche feels like progress and often makes things worse, because it fills your following with people who will never engage with your actual subject and drags your future reach toward them. We unpack that distinction in impressions versus reach.

How much of this do you actually have to do each week?

Less than the advice usually implies, but more consistently. The shape that works is heavy commenting early, tapering as your own posts start to travel.

Buffer's analysis of more than 2 million LinkedIn posts from 94,000 accounts, published in August 2025, found that moving from one post a week to two to five was worth around 1,182 more impressions per post and 0.23 percentage points of engagement rate, with larger gains at higher frequencies. Two caveats before you read that as “post eleven times a week”. It measures accounts that already post, and higher-frequency weeks may reflect people with more to say rather than the frequency itself. We take the trade-off apart in how often to post on LinkedIn.

The commenting half is where most people quietly stop, because it is unglamorous and produces nothing visible for about three weeks. It is also the half that works when you have no audience, which is the entire situation this article is about.

What kills an audience before it forms?

• Changing who you write for every few weeks. Recognition requires repetition. An account that writes for founders in January and job seekers in March has restarted twice.
• Chasing the post that did well. A wide post outside your subject is the most dangerous thing that can happen to a small account, because it rewards you for reaching the wrong people and you will be tempted to do it again.
• Commenting for volume.Fifty “great post” replies do nothing, take an hour, and make you feel productive. Five real ones outperform them completely.
• Buying followers or joining pods. Both inflate the count and destroy the only signal you have about whether the writing works. We go through the mechanics in what engagement pods actually do.
• Mass connection requests. LinkedIn's help documentation states that someone who receives an excessive number of “I don't know this person” responses can be restricted from sending invitations. LinkedIn does not publish the threshold and does not show you how many you have received.
• Waiting until the profile is perfect. The profile matters, but it is two hours of work, not two months. Every comment you leave carries your headline, so do it in week one and move on. The field-by-field version is in LinkedIn profile optimization.

When does a LinkedIn audience start to compound?

At the point where a post gets engagement from people who are not connected to you within the first hour. That is the observable signal that your writing is being carried by your audience rather than by your network, and it usually arrives somewhere in months three to six for people doing the work above.

Three things change once it happens. Your own posts start out-earning your comments, so the ratio should shift. Strangers begin arriving in your inbox referencing specific posts, which is the first commercially useful output. And the topics you can get away with widen, because an audience that knows what you are about will follow you into adjacent subjects that would have confused a cold reader.

What does not change: the reader you picked. Almost every account that stalls after a promising first year stalls because it drifted into writing about LinkedIn, or productivity, or whatever gets engagement, instead of the thing it was originally known for. The audience noticed before the metrics did.

How to build a LinkedIn audience from zero, in order

1. Write the one-sentence description of your reader. Ten minutes.
2. Fix the headline and the first two lines of your About section, since every comment you leave advertises both. Two hours, once.
3. Build the list of 20 to 30 accounts your reader already follows. One hour.
4. Comment eight to ten times a day on that list for four weeks. Twenty to thirty minutes daily.
5. Publish twice a week from the start, aimed at the profile visitor rather than the feed.
6. At week four, check the viewer demographics on your three best posts. If the job titles are wrong, the account list is wrong.
7. At week eight, count the repeat names in your comments. That number is your audience. Everything else is a proxy for it.

The honest summary

Building a LinkedIn audience from zero is a distribution problem before it is a writing problem. Pick one reader and write the sentence down. Find the 20 to 30 places that reader already spends attention, and earn your first strangers inside those comment sections rather than waiting for the feed to discover you. Publish for the profile visitor while the feed is quiet. Then judge yourself on who shows up rather than how many, using the viewer demographics LinkedIn already gives you and the repeat names in your own comments. The patterns behind all of it are in our study of 34,000 LinkedIn posts, and the honest version of the timeline is three to six months of quiet before any of it looks like it is working.