LinkedIn Content Strategy: The Algorithm Rewards the Comment, Not the Post

You publish every week and still feel invisible on LinkedIn. The distribution you are missing is sitting in other people's comment threads, and nobody put it on your calendar. Which twenty accounts would you borrow an audience from tomorrow?

At 8:40 on a Tuesday you have a post queued for nine, a document open in another tab, and no plan for the twenty minutes after you hit publish. In that same window, an account with 800 followers leaves four comments in threads that reach the exact list you cannot buy.

Publishing has software, templates and a hundred thousand posts about hooks. The comment side gets a few paragraphs of encouragement, so the larger half of your distribution runs on leftover time.

LinkedIn Is a Commenting Network That Also Lets You Post

Standard LinkedIn content marketing advice treats the platform as a publishing channel. You pick themes, set a cadence, write hooks, publish. Waalaxy has the most complete taxonomy of content types. lemlist has the most complete profile checklist. All of it answers a question about publishing, and none of it answers the question a small account actually has.

LinkedIn's own engineering team described the feed in March 2026. In "Engineering the next generation of LinkedIn's Feed", Hristo Danchev writes that candidate retrieval runs on LLM embeddings rather than keyword matching, and that ranking reads a member's history as an ordered sequence of interactions instead of scoring each impression alone. The feed serves content "whether it comes from your immediate network or professionals you've never connected with before."

Both halves of the advice industry get that line wrong. A post is not fenced inside your followers, but it does start there. A comment starts somewhere else entirely, inside an audience someone spent years assembling.

The same engineering post separates passive tasks (click, skip, long dwell) from active tasks (like, comment, share). Comment is modeled as its own action, distinct from a like and distinct from a dwell. LinkedIn publishes no weight for it, and anyone who quotes you one is guessing. What the split does establish is narrower and more useful: a comment is a first-class signal in a system that ranks sequences.

So here is the structural difference. A post buys you one audience, the one you already have. A comment places you in a second audience you did not have to build. Nothing on page one schedules for that.

The First Hour Is a Habit, Not a Deadline

You have read that the first 60 minutes decide a post's fate. I have written that sentence myself. Here is what the sources say in 2026.

Hootsuite's LinkedIn algorithm guide now states that early engagement "gives it a nudge, but it's not the whole game anymore", and that a post "can show up in feeds days or weeks later". LinkedIn's engineering post describes the same thing from the inside: when an existing post keeps gaining engagement, its embedding is refreshed and distribution is re-evaluated rather than decided once.

If distribution gets re-scored, a hard 60-minute deadline cannot be the rule. Treat the first hour as the window where your presence is cheapest, not as an algorithm you have to beat.

Three things happen in that window, and none of them are algorithmic. The author is still reading the thread. The people who already commented are still around. And the comment count is low enough that your name is visible without scrolling.

The reply matters more than the comment. A reply that adds a second piece of evidence reopens the thread in front of everyone who engaged with it. A reply that says "thanks for sharing" closes it.

Comment Strategy Before Content Strategy

Avoma is the closest page on this search result to the idea, and it stops one step short. It tells SDRs to comment before posting and suggests ten connection requests, three meaningful comments and one post a day. Direction right. The three is an assertion, and the frame is prospecting, not distribution.

Selection: twenty accounts, four tests

Build a list of twenty. Not fifty, because a list you cannot service is a wish. Each account has to pass four tests.

1. They post at least once a week. An account that posts monthly has no inventory to place you in.
2. Their audience overlaps your buyer, and they do not sell what you sell. You are borrowing attention, not funding a competitor.
3. The author replies to comments. Check three recent posts. If the author never answers, your comment lands in a void.
4. Their posts collect at least five comments. A thread already moving has readers coming back to it.

Keep the list in a sheet with six columns: account, why them, posting day, your angle, last comment, author replied. The last two columns keep the list alive.

Derivation: where fifteen comes from

Twenty accounts posting one to three times a week produce twenty to sixty threads a week. That is your inventory, and it is what you plan against.

Fifteen comments a week is a reasonable start for one person, three a day on weekdays. At five minutes each including reading the post, that is seventy-five minutes a week. One decent post takes longer than that.

Fifteen out of twenty to sixty threads means you choose, every week, which part of the available conversation you show up in. That is a strategy. The alternative is one more post in front of the audience you already had.

Mechanics: what a comment has to do

A comment earns its place in one of three shapes.

The specific instance. You add a number or an example the poster did not have. "We ran this on a 40-person team and the drop came from the second step" is a comment. "Great insight" is not.

The missing constraint. You name the condition under which the advice stops working. It is the most useful thing a practitioner can do in a thread and the least common.

The extension. You ask the question the post did not answer, in a way the author wants to reply to.

Avoid the three moves that mark you as a distributor instead of a participant: restating the post back at its author, leaving a link in the first comment, and commenting on something you clearly did not read.

Engagement pods are the failure mode of this whole section, and the line is simple. Is the comment useful to the person who wrote the post? If the answer depends on who else is watching, it is a pod with better manners.

This is the one part of the cadence I would not automate. There are good social media AI tools for drafting and scheduling. A comment that could have been generated is not the asset described here.

Timing: the window, not the deadline

Comment on the day a post goes out, or the morning after. Not because an algorithm counts minutes, but because the author is still in the thread and readers are still returning. A comment on a two-week-old post is a note left in an empty room.

On the days you publish, reply inside the first hour. On the other days, comment in one block before your own workday starts, so it does not become the task that never happens.

One person owns the list, even when three people comment. Split the twenty accounts by who knows the subject, so the engineer takes the technical threads and the founder takes the positioning ones. A comment written by someone who has not read the account's last three posts reads exactly like what it is.

The swap, stated plainly

A post is a placement in an audience you own. A comment is a placement in an audience you do not own yet, at no media cost, in less time than the post took. That is a decision about where your hours go inside a channel you already chose. It is not the content distribution strategy question of which channels get a slot, which is settled before this starts.

Run it on your own list. If the median account you target has an audience of 3,000 and three of your fifteen weekly comments land on their posts, that is twelve threads a month in front of 3,000 readers, before anyone replies. You did not buy that. You showed up.

Three Pillars That Survive a Quarter

I am not going to define a pillar here; our content strategy template does that. What no template tells you is which three post types a B2B team can still produce in week eleven, when the launch is over and nobody has an idea.

Proof from client work. What you did, what moved, what you would do differently. It is in your delivery notes already, and it gets easier as you sell more.

The objection you heard on a sales call. Answer it in public, once a month. You will hear it again next month, and now you have a link.

The decision and its trade-off. Why you dropped a channel, raised a price, or turned down a project. Not a lesson. A decision.

Thought leadership is not on the list because it is not a topic. It is a tone, and you cannot brief a tone, which is why it produces the vaguest posts on the platform. A reader on Reddit made the same point about LinkedIn advice generally: too much thought leadership, too little that helps a team nurture and close.

Format mix is a different lever, and the one that pays on a visual-first network is in our Instagram content strategy. On LinkedIn the constraint is the idea. A text post carrying a specific decision will outrun a polished carousel carrying a slogan.

Company Page or Founder Profile: Whose Name Carries the Reach

AuthoredUp's September 2026 study, covering 189,909 company-page posts from July 2025 to June 2026, puts median impressions per post on a personal profile at 1,006. Median engagement rate is 2.59% on company pages against 2.36% on personal profiles.

Read those together and the conclusion is not that one is better. A company page holds a slightly higher engagement rate on far less reach. A personal profile carries the distribution.

So split the jobs. The founder profile carries reach and argument, which is where the proof posts and the objection posts go. A person arguing with a person is the format the network was built for.

The company page is the record: archive, hiring, product news, and the version of your work a buyer checks before a call. It also gives every post somewhere to be reposted to.

Pick one founder and one or two subject-matter people and let them be the bylines. Three real voices beat one brand account writing in the third person. Do not hand the team a script so they can comment on each other, because that is engagement manufacturing and readers can tell.

What to Measure When Reach Is Not the Goal

LinkedIn will not attribute a comment to a profile visit. You are measuring a system, not a click path. Three readings are enough.

Profile visits, by week. Read the number every Friday and compare a week with fifteen comments against a week with three. You want a direction, not a rate.

Follower quality. Take your last fifty new followers and count how many are in the segment you actually sell to. A thousand followers who will never buy is a vanity number with job titles.

Comment to impression. For each account, note whether the author replied, whether anyone else replied to you, and whether a new follower appeared within seven days. After a month this says which five accounts deserve double the time.

What you cannot measure: no verifiable published figure exists for what a comment produces in reach or pipeline. Every number you have seen quoted is a vendor assertion or somebody's own dashboard.

A Two-Week Operating Cadence

The plan in one table. Publish three times, comment three times a day on weekdays, reply the same day, review on Friday. The second week differs in one place only: the list gets pruned.

On frequency, AuthoredUp's dataset is unusually clear. Four to five posts a week is the peak, with 28% more impressions per post than posting once weekly, and beyond seven a week both impressions and engagement rate fall. For most B2B LinkedIn content, two or three posts plus the comment block is a better use of the same hours. If your calendar already runs five, do not cut it to prove a point about comments.

One row per account, updated every time you show up. If a row still says "no angle" after two weeks, delete the row.

Start With the List, Not the Calendar

If you take one thing, take the order. Twenty accounts, four tests, fifteen comments a week, three shapes of comment, and a Friday review that prunes the list. Leave the publishing calendar as it is for a month, then read your profile visits.

If they moved while your posting stayed flat, you found the layer nobody scheduled for. If they did not move, you will know that too, which is more than most LinkedIn strategies tell you. Either way, you will have a LinkedIn content strategy you can actually run on a Tuesday.