Meta description: An informal count of 50 posts that spread on X, what the sample showed, what it can’t tell us, and how X really counts its 280-character limit.
X gives free accounts 280 characters per post, and almost nobody in my sample used all of them. This is an informal look, not a scientific study, and I’ve tried to be clear about what it can and can’t show.
How I counted
I collected [50] posts from my own X timeline between [dates] that had clearly spread, which I defined as [for example: more than X reposts or replies, or posts I saw quoted several times]. I counted each one by [your method: pasting the text into a character counter / using X’s own composer], and I [did / did not] treat links and emoji differently.
Posts are listed in [link to your spreadsheet or a table of post links and counts]. I link to the posts rather than copying their text.
What the sample showed
| Measure | Result |
|---|---|
| Posts counted | 50 |
| Average length | 94 characters, about a third of the limit |
| Shortest | 12 characters |
| Longest | 267 characters |
| Most common range | 60–140 characters, about 6 in 10 posts |
Only a handful went past 200 characters. Most were short enough to say out loud in one breath.
What this sample can’t show
- It isn’t a random sample. My timeline reflects who I follow and what X’s algorithm shows me, so these posts say more about my feed than about X as a whole.
- There is no comparison group. I didn’t count ordinary posts that didn’t spread. If most posts are also short, the length of the viral ones tells us little. A fair test would count a sample of posts that didn’t take off, then compare.
- It is small. With 50 posts, “6 in 10” means 30 posts, and a few posts would change the picture.
- Other things drive spread. The author’s audience, timing, topic and whether the post has an image all matter, and I didn’t control for any of them.
So I can say that in this sample, the posts that spread skewed short. I can’t say that shorter posts spread further.
Patterns I noticed (observations, not findings)
- One idea, no hedging. The shortest posts made a single claim and stopped, which may make them easy to react to.
- Line breaks instead of long sentences. Several of the longer posts were two or three short lines stacked, which read quickly even though the count was higher.
- Jokes seemed to sit under about 100 characters. That was my impression, and I didn’t measure it separately.
- Long posts tended to be explaining something, such as a news reaction or a correction, where the extra length did real work.
How X actually counts to 280
If you check your own posts, know that X doesn’t simply count keystrokes. According to X’s developer documentation, it uses weighted counting:
- Every link counts as 23 characters, however long the real address is.
- Most emoji count as 2 characters each.
- Chinese, Japanese and Korean characters count as 2 each.
- Basic Latin letters, digits and punctuation count as 1.
X Premium subscribers can post up to 25,000 characters, though the feed collapses long posts behind a “Show more” link. Everything above applies to free accounts.
A gut-check before you post
- Could you say it in under 100 characters without losing the point? If so, consider it.
- Are you hedging something that would land better stated plainly?
- If it’s long because you’re adding context, a correction or a source, that’s fine. Length that carries substance reads differently from length that is just throat-clearing.
You can paste a draft into our character counter and set the limit to 280 to see a live count, along with words, reading time and reading level. If your post has links or emoji, check X’s own composer for the final number, since X counts those differently.
Last updated: October 30, 2026. Weighted-counting rules are from X’s developer documentation.
About the author
Kent Williams is the founder and developer of Open Character Counter, a free browser-based tool for tracking character counts, word counts, and readability. He writes about writing constraints, platform limits, and SEO. Read more about Open Character Counter →
Before you publish, decide honestly which case you’re in
- If you really collected the data: keep the spreadsheet, fill in every bracket, and publish the method. A smaller, honest sample is more valuable than a polished claim. Consider adding a comparison sample of ordinary posts, which would turn it into a real finding.
- If you didn’t: don’t publish the numbers. Retitle it as a guide, something like “How Many Characters Should an X Post Be?”, and drop the first-person study entirely. I can write that version for you.
- Check your own tool. The post tells readers your counter shows a live count against X’s limit. If it counts raw characters, it will be wrong for posts with links or emoji. Either add X’s weighted counting (X publishes an open-source library for it) or keep the caveat I added.
Leave a Reply