Watch what a typical founder does after publishing a post on X: nothing. The post goes up, they close the tab, and four replies arrive into silence. Two days later they check notifications, heart a couple of comments, and wonder why their posts never seem to travel.

Now watch the accounts that actually grew from zero in the last two years, the operators who turned posting into pipeline. Almost universally, they treat the minutes after posting as the job itself: every reply gets a reply, fast, and the comment section under their post runs like a hosted conversation. One founder who documented his run described his own behavior in a way that stuck with us: the overwhelming majority of his total replies were replies to comments on his own posts. Not riding other people's threads. Working his own.

Two independent operators we studied converged on this from completely different products, which is usually the tell that a mechanic is load-bearing. Here's why it works and how to run it.

The mechanics: why answering comments multiplies reach

Post 1 laid out the feed's physics: velocity, especially reply velocity, is what the algorithm reads as "this post is a conversation" and carries outward. Every reply you make to a commenter is another engagement event on your own post, arriving in exactly the window where events count most. A post with 10 comments and 10 author replies isn't twice the conversation of 10 unanswered comments; in the feed's arithmetic it's far more, because the back-and-forth chains (commenter replies again, others join the exchange) stack engagement the silent version never generates.

That's the algorithmic half. The human half is stronger: a founder who answers every comment is visibly present, and presence changes what people do next. Lurkers comment on posts where the author writes back, because the comment might start a conversation instead of vanishing. Your comment section develops regulars. The regulars bring their followers' attention. The post after this one starts warmer. It's the smallest compounding loop on the platform, and it's entirely under your control.

The window matters: the first hour or so after posting is when reply velocity shapes the post's trajectory. Which implies something about when to post that founders miss: post when you can stay, not when the scheduling tool says engagement peaks. A post published at the "optimal" hour and abandoned loses to a post published whenever you genuinely have forty minutes to host it.

The part founders really miss: repliers are leads

Here's the reframe that turns this from growth tactic into pipeline: everyone who replies to your post about the problem you solve just self-selected. They read a founder's post about cold outreach, or channel strategy, or whatever your space is, and chose to engage with it in public. That's the same self-selection a LinkedIn giveaway post manufactures, arriving for free under your own content.

Treat the comment section as a worked list, not an applause meter:

The substantive replier (shared their own experience, pushed back, asked a real question) is a conversation already started. Continue it in-thread like a person, and when it naturally exhausts the thread, the DM is warm (post 7's rules apply: the interaction earned it).
The drive-by agreer gets a real reply anyway, because your reply is also for the audience, and because this month's "so true" is sometimes next month's customer warming up.
The recurring name is your actual asset. Someone who's shown up on three posts is telling you something no ad platform can: sustained interest, on the record. Know who they are.

None of this requires the post to be about your product. It almost never should be. The post is about the problem space you know deeply (the same funnel logic every channel series keeps arriving at: teach for the audience, let the profile and the DMs convert the hand-raisers).

What to post so there's something to reply to

This series won't pretend to be a content masterclass, but the own-room engine has input requirements, and they're learnable:

Post things a practitioner can respond to. Opinions with a stake, numbers from your corner of the world, "here's what broke and what we did" stories. The test is whether a smart reader has something to say back, not just something to feel. Announcements fail the test; most launch posts are announcements, which is why they get 4 likes.
End open when you can. A post that closes every loop leaves repliers nothing to add. The versions that breed comment sections leave an edge visible: "we still haven't solved X," "curious if anyone's seen the opposite."
If you're pre-audience, build in public honestly. What to post at $0 MRR is its own art, but the short version: the work itself, the decisions, the numbers you can share. Practitioners reply to practice.
Frequency follows post 4's rules. A post a day at human rhythm beats three posts on Monday. The comment-hosting time budget is the real constraint anyway; better one post you host well than three you abandon.

The two engines, one machine

Put this post together with the last one and you have the whole reach system: borrowed rooms and your own room. Riding puts you in front of audiences you haven't earned; your own comment section converts the strangers who followed into a compounding conversation you host. Riding without an own-room is a leaky bucket (attention arrives at a profile whose posts are dead ends). Own-room without riding starves (nobody new ever arrives). A few rides and one hosted post a day is the whole visible engine, and it fits inside the daily half hour post 8 assembles.

Hosting is easier with a second pair of eyes: Thread Otter watches your own posts' repliers alongside everything else, drafts responses in your voice while the window is hot, and remembers which names keep coming back, so the hosted conversation and the lead list stop being things you reconstruct from notifications at midnight.

This is post 3 of the 8-part X for founders series. Next up: volume with a human rhythm, or how the channel that rewards showing up also bans accounts that show up like machines.