A founder we work with published a post about hiring. Not a bad post. Well written, genuinely useful, drawn from something that had actually happened that week. It did about a third of his normal reach.
He shrugged it off. One flat post out of twelve that month. Then the next post — squarely back in his lane, a mechanic he'd covered profitably four times before — also underperformed. So did the one after it.
By the fourth post he was back at baseline and had spent a week and a half wondering whether the algorithm had changed.
It hadn't. The off-lane post didn't cost him the off-lane post. It cost him the next three. That's the part almost nobody prices, and it's the reason "just post about whatever's on your mind sometimes" is more expensive advice than it sounds.
We've written before about building topic authority — the case for narrowing, staying in one subject, letting the platform learn what you're for. This is the other half of that ledger: what happens when you break it, how the cost is actually charged, and how long it takes to get back.
The mechanic: distribution is topic-shaped, and your account carries a profile
Here's the version we're confident in, without the invented percentages.
LinkedIn does not decide who sees your post by asking "who is connected to this person." It decides largely by asking "who is likely to find this relevant." Since the move to an LLM-based ranking layer, that relevance judgement is semantic — the system reads what your post is about and matches it against what it believes individual readers care about.
That has a second-order consequence most founders never think through. To do that matching efficiently, the system needs a working idea of what you publish about. Your account accumulates an inferred topic profile from everything you've posted, and that profile is what determines the composition of the initial test pool your next post is shown to.
So when you publish inside your lane, the first room your post walks into is a room the platform has spent months assembling out of people who reliably engage with that subject. When you publish outside it, the platform still uses that room. It has nowhere else to put you.
Your hiring post got shown to a couple of thousand people who follow you for Amazon creative. They didn't engage — not because it was bad, but because it wasn't the thing they raised their hand for. The system read that as a weak signal and stopped expanding distribution.
That's the first cost, and it's the small one.
The second cost is the one that lands on your next post
The weak result doesn't evaporate. It becomes an input.
The platform is continuously updating two things: its picture of what you write about, and its picture of how your content tends to perform with the audience it has built for you. An off-lane post feeds a poor result into both. Your topic profile gets slightly muddier, and your recent performance signal gets slightly worse.
So your next post — the good one, the on-lane one, the one you were confident about — starts from a marginally worse position. Smaller or less well-matched test pool, less confident placement. It underperforms slightly. That result feeds back in too.
This is why the recovery isn't instant and why it doesn't feel like a penalty. Nothing was penalised. There is no strike, no flag, no reconsideration request, nothing to appeal. The system simply became less certain about what you are, and certainty is what buys you a good starting room.
The practical shape, from what we see across client accounts and what's broadly reported by people who track this seriously: posts start landing back at baseline within about two to four weeks of consistent on-topic publishing. Building genuine topic authority from a standing start is closer to 60–90 days.
We'd treat both of those as directional rather than precise. Nobody outside LinkedIn has published a study on off-topic recovery, and if someone hands you a specific percentage for what an off-lane post costs, they made it up or repeated somebody who did. What isn't in dispute is the shape: the cost is not paid where it's incurred, and it takes multiple posts to clear.
Why this hits ecommerce founders harder than most
Three reasons, and none of them is a discipline problem.
Your lane is unusually narrow, which is the whole point. A generalist business account posting about leadership has a topic profile so broad it can absorb almost anything. A founder who writes specifically about Amazon creative, or 3PL selection, or DTC unit economics, has a tight profile — and a tight profile is precisely what produces the good test pool. The asset and the fragility are the same asset.
Your genuinely interesting weeks are frequently off-lane. You raised. You changed 3PLs. You hired a head of ops. You went to a trade show. These are the weeks with the most to say and the least of it inside the subject your audience subscribed to.
Broad posts feel like they're working. This is the one that keeps the mistake alive. An off-lane post about hiring, or burnout, or what running a business teaches you, is agreeable. It gets pleasant comments from peers. On some accounts it will outperform your normal reach on raw engagement while doing nothing for pipeline, because the room it recruits is a room of people who will never buy anything from you. You feel the applause and never see the bill, which arrives three posts later as a number you attribute to the algorithm.
The exception: an off-lane post that carries an on-lane payload
We are not arguing for never writing about anything else. We're arguing that the subject and the material are two different dials, and most founders move both at once.
The move that works is to keep the lane and change what you're using it for. The trade show becomes a post about what four conversations with buyers told you about how your category is being merchandised. The 3PL migration becomes a post about the returns mechanic you discovered in the middle of it. The hire becomes a post about the specific gap in the account that made you realise you needed one.
Same week, same raw material, completely different topic signal. The platform reads the subject, and the subject is still yours.
The version that costs you is the one where the raw material is the post — "here's what I learned about hiring" — because the subject has genuinely changed and there's nothing in the text tying it to the thing your audience is there for.
What to do if you've already done it
Nothing dramatic, and specifically do not try to fix it with a bigger swing. The instinct after two flat posts is to reach for something punchier — a hotter take, a broader hook, something engineered to travel. That is a second off-lane post with better packaging, and it makes the recovery longer.
The recovery is boring on purpose:
Publish three or four consecutive posts squarely inside your subject. Not adjacent, not "related." The most obviously-you material you have. This is the fastest way to re-confirm something the system half-remembers rather than asking it to learn something new.
Hold your normal cadence through the flat ones. The flat posts are the recovery, not evidence the recovery isn't working. Cutting cadence because the numbers look bad extends the exact problem you're trying to close.
Don't grade individual posts during the window. Grade the run. If posts five and six aren't back near baseline, the problem isn't the off-lane post — it's something else, and you now know that too.
Put the off-lane material somewhere it doesn't cost distribution. A comment on somebody else's post, a newsletter, a reply in a DM thread, a section on your About page. The instinct to publish it as a post is usually an instinct to say it, and there are cheaper places to say things.
How to decide before you publish, in about ten seconds
We keep this deliberately blunt for clients, because a rule that requires deliberation gets skipped at 11pm.
Ask one question: would a stranger reading only this post be able to tell what I do?
Not "does it mention my industry." Not "is it business-related." Would somebody landing cold on this single post come away knowing what you're for. If the answer is no, it's off-lane, whatever else it is, and it will be priced accordingly.
The second question, for the borderline cases: is there a version of this that keeps the story and changes the subject? There almost always is, and it's almost always the stronger post, because the specific operating detail is the part nobody else could have written.
FAQ
Is one off-lane post a month actually a problem? Probably not, if the rest of your cadence is tight and the off-lane post is genuinely occasional. The failure mode we see isn't one deliberate exception, it's drift — a founder who is bored, or busy, or off the tools, publishing progressively broader material until the profile is muddy and nobody can say what the account is about. That's not one post, that's a quarter.
What about announcements — funding, hires, awards? Same mechanic, plus a second cost we've written about separately: slots are the scarce resource. Publish them if they matter, keep them rare, and where you can, attach an operational lesson so the post has a subject beyond the news itself.
Does a post that gets high engagement but off-lane still hurt? Our honest answer is we don't know, and be suspicious of anyone who says they do. What we'd say is that engagement from a room of peers who found your broad take agreeable is not the same input as engagement from the audience your lane was built to reach, and the account-level effects we can observe track lane consistency more closely than they track raw reaction counts.
How do I know whether my reach dropped because of this or because of something else? Look at whether the drop is shaped. An off-lane cost shows as a soft patch of three to five posts that recovers on its own if you hold the lane and the cadence. A structural problem — a stale profile, a lane that's drifted broad over months, a cadence gap — doesn't recover on its own, and you'll still be flat at post eight.
If your content is technically fine and your reach keeps sagging for reasons nobody can name, the problem is usually structural rather than creative — and it's usually visible in twenty minutes of looking at the last thirty posts as a body rather than as individual posts. That's the first thing we do with a new client. Get in touch if you'd like us to run it on yours.