guides · replies

Reply strategy on X

"Reply guy" is an insult because most people do it badly. Done properly it is the fastest way to grow a small account on X, and it is not close. This is the playbook: why replies beat posts below a few thousand followers, how to choose what to reply to, and what separates a reply that earns followers from one that reads as noise.

Why replies beat posts when you are small

Your own posts are shown first to your own followers. If you have 400 followers, only a fraction of whom are online right now, your post starts life in a very small room and has to earn its way out. That is the ceiling, and no amount of rewriting removes it.

A reply starts somewhere else entirely. It lands inside someone else's post, in front of their audience, and that audience is made almost entirely of people who have never seen you. You are borrowing distribution you have not built yet. Replies are also posts in their own right, so they accumulate impressions, and a reply that lands well gets its own likes, its own profile clicks, and its own follows.

The arithmetic is stark. Twenty replies a day is twenty appearances in front of new audiences. Two posts a day is two appearances in front of the same audience. Both matter, but only one is capable of finding you people. This is why the guide on getting more impressions puts replies at the centre.

Picking targets: the size window

Almost all of the outcome is decided before you write a word, by what you chose to reply to. There are three tiers and only one of them works.

• Mega-accounts. Hundreds of thousands of followers. Every post gets hundreds of replies within minutes, most of them from people running the same play. Enormous reach, your share of it approximately zero, and the author will never see you. Occasionally you get lucky. Do not build a strategy on it.
• Small accounts. Below your own size. Nothing wrong with replying to friends, and you should, but understand it as relationship building, not reach. There is no audience to borrow.
• Mid-sized accounts in your niche. The window. Big enough that their posts collect real attention, small enough that ten thoughtful replies is a busy thread rather than four hundred. Your reply is visible for hours, not seconds. And crucially, the author reads their replies.

That last point is the multiplier people miss. When a mid-sized account replies to your reply, or likes it, their audience is pulled directly onto you and the thread stays alive longer. Do that repeatedly with the same twenty or thirty accounts and you stop being a stranger. You become a familiar name in a niche, which is the actual mechanism by which people decide to follow you.

Build a target list. Twenty to forty accounts in your niche, sized above you but not out of reach, who post regularly and whose audience overlaps with the people you want. Check them daily. The list is the asset.

Early beats clever

This is the part almost nobody internalises. Replies stack chronologically and by ranking, and most people who open a post read the first few replies and leave. If you are reply number four hundred, your reply exists in a technical sense and nobody will ever read it.

So the highest-value action on X is finding a post that is minutes old, from an account in your window, that is about to take off, and being in the first handful of replies. When that post goes on to collect fifty thousand impressions, your reply sits near the top of all of them. One well-placed early reply can out-reach a week of careful late ones.

You cannot know for certain which posts will take off, but you can read the early signal: is this post accruing engagement unusually fast relative to that account's normal pace, in its first few minutes? A post doing five times its author's usual early velocity is a post the algorithm is already testing on a wider audience. That is the wave to be standing on.

The practical problem is that this requires watching a moving timeline continuously, judging velocity against each author's baseline, and deciding in seconds. Doing it by hand means living in the app. This specific job - rank the live timeline by reach potential and how early you are - is what Argus was built to do.

What a reply worth writing looks like

Once you are early and in the right thread, the reply still has to earn the profile click. The test is simple and unforgiving: would this reply be worth reading if the original post were not above it? If it only makes sense as a reaction, it is noise.

What passes the test:

• A specific counter-example. "This did not hold for us - we tried it at 200 users and churn went up. Here is what we think happened." Disagreement with evidence is the highest-value reply there is.
• A real number from your own experience. Concrete beats abstract. Numbers are the fastest way to signal you have actually done the thing.
• The next step nobody wrote. The post makes a point; you add the part that follows from it, or the caveat that matters in practice.
• A genuine question that opens the thread. Not "thoughts?" - a question with enough context that answering it is interesting for the author.

What fails, always:

• "Great post!" and every variant. Zero value, zero reach, and read as spam at volume.
• Restating the post in your own words so it looks like agreement with effort. Readers see through it instantly.
• Pitching your product. The fastest way to be blocked by exactly the accounts you needed.
• Manufactured contrarianism. Being difficult is not the same as being interesting, and it burns the relationship.

On length: one to three sentences is usually right. Long enough to contain a real thought, short enough that people read it in the feed without expanding. If your reply needs five paragraphs, it wanted to be a post.

Volume, honestly

Fifteen to twenty five replies a day is the target. That is enough to compound and few enough that each one can be real. At that volume you are looking at roughly thirty to forty five minutes once you know your target list, and most of that time is finding the posts rather than writing.

Consistency matters more than the number. Twenty a day for a month beats two hundred in a burst followed by nothing, both because the algorithm reads sustained activity and because the humans on your target list only start recognising you around the fifth or sixth time they see your name.

Two warnings. First, do not automate the send. Automated replying is the fastest route to a restricted account, X blocked API-based reply automation in February 2026, and the account at risk is the one you are trying to grow. Draft with help if you like; press send yourself. Second, do not chase volume by dropping quality - a hundred empty replies a day is bot behaviour and gets treated as such.

Measuring whether it is working

Reply strategy shows up in a specific pattern: impressions rise first, then profile visits, then follows, with a week or two between each. If impressions and profile visits are climbing but follows are not, your replies are working and your profile is the leak - that is the one case where rewriting your bio is the right move. Check the shape with the funnel calculator, and read grow on X with replies for the follower-count side of the same story.

Common questions

How many replies a day should I write on X?

Fifteen to twenty five a day is sustainable and produces real reach, and it takes about forty minutes if you are not hunting for targets manually. Consistency beats intensity: twenty a day for a month is worth far more than a hundred in one afternoon and none for a fortnight.

Should I reply to big accounts or small ones?

Neither extreme. Under a mega-account you are one of a thousand replies and invisible. Under an account smaller than you there is no reach to borrow. Target mid-sized accounts in your niche whose posts are currently rising, where a good reply is visible and the author may actually respond.

Does being early to a post really matter that much?

Yes, more than anything else you control. Replies stack and most readers only see the top few. An early reply on a post that later takes off rides the whole wave. The same reply three hours later is buried and effectively invisible.

Do short agreement replies like "great post" do anything?

No. They earn nothing from readers, and at volume they read as bot behaviour and can cost you more than they earn. A reply has to be worth reading on its own - a specific counter-example, a real number, or a genuine question - or it is noise.

Where Argus fits

The hard part of this playbook is not the writing. It is finding, live, the posts that are early, rising, and the right size, twenty times a day, forever. Argus watches your timeline in your own browser, ranks the opportunities by reach and how early you are, and drafts the reply in your voice so you are editing rather than staring. It never sends anything. You always press the button. See also the reply grader.