Three parts, in order
A post that works has a hook, a body and a payoff. Most founder posts have only a body: a paragraph of reasonable observation with no claim at the top and nothing at the bottom. It reads as thinking out loud, which is fine in a group chat and invisible in a feed.
| Part | Its job |
|---|---|
| Hook | Make a claim worth arguing with. This is the only line most people will read. |
| Body | Earn the claim. Evidence, mechanism, the specific thing you saw. |
| Payoff | Land the claim again, harder. Not a summary. |
The proportions matter less than the presence of all three. A post can be four lines and have every part.
The hook is a claim
The single most common failure is a first line that describes instead of claiming. "Some thoughts on how we approach hiring" tells a reader what is coming and gives them no reason to want it. "We stopped doing take-home tests and our hiring got better" is the same post with a claim on top.
The working test is the one we apply to launch copy: could somebody who knows the subject disagree with this line? If nobody can, the post produces no replies, and replies are a large part of what drives reach.
Ranking systems weight engagement per impression, so a post everyone nods at and nobody responds to is telling the system that people saw it and felt nothing. The mechanism is here.
Two constraints on this, both of which people get wrong in the same direction.
- Arguable is not the same as untrue. Writing something you cannot defend gets you attention for an afternoon and a reputation you keep. The claim has to be one you actually hold.
- You have to turn up. A contrarian line abandoned the moment somebody pushes back is worse than no line, because the retreat becomes the story. Budget some of the evening for the replies.
The strongest hooks tend to cost you something: a mistake you made, a thing your product is bad at, a position most of your peers disagree with. They read as credible precisely because a marketing team would have removed them.
What hedging costs
"I could be wrong about this, but." "Not sure if this is a hot take." "Just some thoughts."
Every one of those is a sentence spent apologising in the only line most people read. The instinct behind them is decent, and it is expensive: the hedge is there to soften the claim, which is the same as removing the reason to keep reading.
The odd part is that hedging does not even protect you. Being wrong in public with a clear claim is recoverable, and people respect the correction. Being vague in public just means nobody engaged either way.
Cut every clause that exists to lower the stakes. If the claim is genuinely uncertain, say what would change your mind, which is a stronger sentence than a hedge and invites exactly the replies you want.
The body is evidence
Once the hook has made a claim, the body has one job: make the claim survivable. Not to add related thoughts, not to cover the caveats, and not to reach a word count.
- One idea per line. Line breaks are how a reader decides whether to keep going. A wall of text is a decision to lose the skimmers, and most of your audience is skimming.
- Specifics over adjectives. A number, a name, a duration, a thing that happened. "Significantly faster" is a claim with nothing under it; "cut it from forty minutes to four" is evidence.
- Cut anything that does not serve the hook. Interesting adjacent points belong in another post. A post that makes two arguments makes neither.
- Add media if you have it honestly. Media of any kind ranks above plain text, and the lift comes from having motion rather than production value. A ten-second screen recording of the actual thing beats a graphic. More on that trade.
Where the payoff lands
The last line restates the hook on its own, after the evidence. It is not a summary of the post and it is not a call to action.
The reason it works is that a reader who arrived at the bottom has just been given the evidence, so the claim they were unsure about at the top is now a claim they can hold.
Restating it there is what they screenshot and what they quote when they repost it, which is the mechanism by which your framing rather than someone else's becomes the version people repeat.
Keep it shorter than the hook if you can. A payoff that runs to three lines has become a conclusion, and conclusions get scrolled past.
Where the next twelve posts come from
The structure above solves one problem and immediately creates another. You now know how to write a post and you are staring at a blank page on a Tuesday with nothing you actually want to claim, so you write something reasonable and general, which is the failure this whole page exists to prevent.
The fix used by most people who post consistently without it becoming their job is to stop generating posts one at a time.
Produce one substantial thing, then take it apart. An hour of recorded conversation about a decision you argued over internally.
A written piece on why you built something the harder way. A walkthrough of the thing that surprised your own team.
An hour of a real technical argument contains somewhere between ten and twenty claims a knowledgeable person could disagree with.
Each of those is a post, and it arrives with its evidence already attached, because you were explaining it to somebody when you said it. The cost of twelve posts becomes one hour of talking plus twelve short writing sessions, rather than twelve blank pages.
Two things make this work rather than turning into filler.
- The source has to be real. An hour of you saying marketing sentences decomposes into twelve marketing posts. The material has to be something you would have said anyway, to somebody who would have pushed back.
- Each extract needs its own hook. A clip or a pulled quote is a body looking for a claim to sit under, and that claim has to be written fresh each time using the test above.
The related move is quoting your own longer piece later with a different argument from inside it. Done properly this is a new post: a fresh claim on top, with the original as the evidence underneath it, which gives a reader who missed the first one a reason to go back. Done lazily it is the same text posted twice, and the audience can tell the difference immediately.
If the substantial thing is recorded rather than written, the clips are also the cheapest media you will ever have. Anything with motion ranks above plain text, and a set of extracts from one recording means every post that month carries media instead of one of them doing. The production economics are here.
What this will not do
None of this makes you a great poster. The accounts you admire are mostly run by people who spend hours a day on the platform, reading it, arguing in it, and rewriting posts before publishing. That is a real craft and it is a full-time one.
You are running a company. The correct ambition is competence, not mastery: get the structure right, publish, argue with the people who turn up, and get back to work. Forty minutes a day is enough to compound, and trying to do more usually means doing the company badly.
There is also a ceiling that better writing cannot lift. A well-built post from an account with a small following reaches the people who already follow you plus whoever the system decides to test it on.
That is a real number and it is not a large one. Getting past your own audience is a distribution problem rather than a writing problem, and it is the thing we actually do.
Write the post properly anyway. Distribution amplifies whatever it is given, and it is just as efficient at spreading a weak claim as a strong one.
Questions we get asked
What makes a good hook for a post?
A claim somebody could disagree with. Descriptions are safe and produce nothing, because a reader who agrees immediately has no reason to respond and replies are a large part of what drives reach. The working test is whether a knowledgeable person in your field could argue with the line, and whether you would still hold it in the replies that evening.
How long should an X post be?
Long enough to support the hook and no longer. Most founder posts are one long middle with no claim at the top and nothing at the bottom, which reads as thinking out loud. Three or four short blocks under a strong first line outperforms a paragraph almost every time.
How much time should a founder spend on this?
Twenty minutes to write, and some of the evening in the replies. Trying to become a full-time poster is the wrong trade for someone running a company. Do the structure well, publish, argue with the people who turn up, and get back to work.