Silent followers on X: how many never post, and does it hurt reach?

A silent follower follows you but gives us nothing to judge: no posts of their own, posts too thin to read anything into, or a private account. In the 15 TrustAudience reports on the current method, between 11 and 33 of the newest 100 followers were Silent on small accounts, against 54 and 72 on the two large ones.⁠^[1] X's published ranking code scores each post for each viewer and has no rule that marks an author down for quiet followers, so removing them is housekeeping; nothing X has published suggests it will lift your reach.⁠^[2]

What "Silent" means in a TrustAudience report

We read up to 10 recent public posts or replies from each of up to 100 of an account's newest followers (see the method). A follower is Silent for one of three reasons, and each report counts them separately:

• No posts of their own. We read their latest timeline page and it was empty, or held only reposts and other people's posts.
• Posts too thin to judge. They have posts, but too little for the model to say anything about them.
• Private. A protected account. We count it without reading its posts, and nothing about it goes to the model.

Silent scores 40 in the Audience score, between Mixed (60) and Repetitive promotion (25). It is drawn as a hollow person, never a robot, because staying quiet is not evidence of automation. Two other labels are easy to confuse with it. Unavailable means the account is suspended or deactivated (scored 25). Unclear means we could not read the follower at all, and they are left out of the score.

"Silent" is also not the same as "inactive". It describes posting, not reading. Someone who opens X every day, reads, and never posts is Silent in our reports, and we have no way to see how often they visit.

How many of the newest followers are Silent

The data is small and early: 15 public reports on method audience-context-2026-10-03-v3, all checked on 3 or 4 October 2026.⁠^[1] Every account on the board was checked by someone's choice, not drawn at random, and only the newest followers are read. So these are ranges for these accounts, not a rate for X.

No account between 1,000 and 100,000 followers has a current-method report yet, so we can't say where the change happens or whether size is the cause. Both large accounts are widely known names, which may draw follows from people who want the name on their list more than a conversation. The full spread, account by account, is in the newest-followers study.

No posts, thin posts or private: why followers were Silent

Pooled across the 15 reports, 394 of the 1,325 followers we counted were Silent.⁠^[1] Most had nothing of their own to read.

Why newest followers were Silent (pooled)

"No posts of their own" was the largest reason in 14 of the 15 reports. Private accounts were a small slice everywhere: 13 across the two large accounts' 200 newest followers, 15 across the ten small accounts' 1,000. None of the 1,325 was Unavailable, and none was Unclear.

Do silent followers hurt your reach on X?

The best public evidence is the code X publishes for the For You feed, in the xai-org/x-algorithm repository. It was created on 19 January 2026, its README lists updates on 13 and 14 August and 18 September 2026, and we read it on 4 October 2026.⁠^[2] Three points in it bear on quiet followers:

1. Ranking happens per viewer. For each post, the model predicts how likely that viewer is to like, reply, repost, click, dwell, follow, or mute, block or report. A weighted sum of those predictions becomes the post's score for that person. X's README stresses that the weights multiply your own predicted probabilities, which are "substantially driven by your own behavior".⁠^[2]
2. Followers are a source, not a grade. In-network posts come from a store of recent posts from the accounts a viewer follows. Out-of-network posts come from retrieval models. Both are ranked by the same model.⁠^[2]
3. Reading counts. Dwell, dwell time and clicks are among the predicted actions.⁠^[2] A follower who reads without posting is not invisible to the model, even though our reports call them Silent.

The README describes no step that looks at the share of an author's followers who post, and no follower-quality penalty in ranking. A follower who never engages simply adds nothing to your posts' engagement. A quiet following can make engagement look thin next to your follower count, but the published code ranks posts on predicted actions, not on that ratio.

What X hasn't published

Some of it is still unknown, so we don't promise anything either way:

• The README says some files are held back to limit gaming, including some spam-labelling rules and the prompts used by its content classifiers.⁠^[2]
• One published component, user-cred-v2, runs PageRank over the follow graph and engagement edges to give each account a score.⁠^[2] The README doesn't say how much a follower who never engages contributes to it. It lists this among the content-understanding systems whose scores visibility filtering reads.
• Parameter values change over time (the repository documents one such change), so the code shows the system, not every live setting on a given day.⁠^[2]

So the accurate answer is that no published rule penalises you for silent followers, and nobody outside X can rule out an unpublished one.

Should you remove inactive followers?

Usually not. The published code gives no reason to expect a reach gain, many of those followers may be reading, and X already handles one kind of dead weight itself: when an account is locked, X removes it from follower counts across profiles.⁠^[3]

Removing someone does make sense when:

• They spam you or fill your replies with the same promotional post. That is the Repetitive promotion or Repeated reply pattern label, not Silent.
• You show your follower count to sponsors and know a batch of follows wasn't genuine interest, for example from a follow-for-follow push.
• You simply don't want that account following you.

To remove one, open your followers list or their profile, choose the more icon, then Remove this follower. X says this works on the web only for now, and the account can follow you again later.⁠^[3] Third-party mass-removal apps need access to your account. Be wary of giving it to them for a benefit X's code doesn't support.

Check it by hand

You can estimate your own Silent share in about twenty minutes, without TrustAudience:

1. Open your followers list on the web and take the first 30 accounts. Our data provider documents its follower list as newest first; X doesn't document the order of its own list, so treat this as a rough stand-in for your newest followers.
2. Open each profile. Mark it no posts if the Posts tab is empty or holds only reposts. Mark it private if it shows the lock icon.
3. For the rest, read the Posts and Replies tabs. If one or two short posts are all there is, mark it too thin.
4. Add up the three marks and multiply by 100/30 to compare with the table above.

Keep in mind that this is a snapshot of your newest followers, not your whole following, and it tells you about posting, not reading. If you'd rather not do it by hand, a TrustAudience check reads up to 100 newest followers and shows the Silent split for you.