How to go viral on Threads

The short answer

• Video led every format. Single-video posts had a median of 116 engagements, 38% higher than text-only posts (84).
• Questions trade likes for replies. Posts that ended with a question had a median of 53, 44% lower than posts with no question (95), but they got 2.1 times as many replies.
• Links are rare in top posts. Only 45 of 22,869 (0.2%) had one, and their median was 36% lower than the rest.
• Fewer lines did better. One-line posts had a median of 94, 34% higher than posts with 7 or more lines (70).
• Length is not a lever. Posts under 50 characters led overall (119), but in trending-feed posts only, 400 to 500 characters led (88).
• Hashtags made little difference: 83 with one, 84 without.

Every number here is a median of likes, replies, reposts, and quotes per post. The typical post in the study earned 84. Want to know when to post instead? See our study on the best time to post on Threads.

How long should a Threads post be?

Posts under 50 characters had the highest median in the full set, 119 engagements, 42% higher than the typical post (84), but there is no single best length.

That lead comes from a small group. 541 of those 1,214 short posts, about 45%, did not come from the trending feed. We collected them by following 23 high-volume accounts directly. Keep only trending-feed posts and the order flips. Posts of 400 to 500 characters lead at 88, and posts under 50 characters drop to 63.

So don't cut a good post short to chase a number. Write as long as the idea needs. Threads gives you 500 characters per post, and our guide to the Threads character limit covers the rest. To see how a long post will look before you publish it, try the free Threads post previewer and character counter.

All 22,869 posts

Trending-feed posts only

How many lines should a Threads post have?

Fewer lines did better: posts written as a single line had a median of 94 engagements, 34% higher than posts with 7 or more lines (70).

Each step up in lines went with a little less engagement. The gap is smaller in the trending-only view, 77 for one line against 69 for 7 or more, but it points the same way. One strong line is easy to take in while someone scrolls. If your idea needs more room, that is fine. Just make the first line count. Our guide on how to grow on Threads has more on writing hooks.

Do images get more engagement on Threads?

Yes: single-image posts had a median of 92 engagements, 10% higher than text-only posts (84), and video did best of all at 116 (38% higher).

Carousels landed in between at 101. The pattern is even stronger in the trending-only view, where video reached 124 against 70 for text. Still, 87% of these top posts were plain text, so text clearly works. But if you have a good photo or a short clip, add it. ElvenFlow can schedule image, carousel, and video posts to Threads.

Do links hurt reach on Threads?

Top posts almost never had a link: only 45 of 22,869 posts (0.2%) had one in the text, and their median was 54 engagements, 36% lower than posts without one (84).

45 posts is a small group, so treat the exact gap with care. In the trending-only view there were just 29 posts with a link, too few to report. The clearer signal is how rare links are. Out of thousands of the best posts on Threads, almost none sent people somewhere else. If you need to share a link, make the post worth reading on its own. We did not measure links placed in replies, so we can't tell you if that works better.

Should you ask questions in Threads posts?

Posts that ended with a question had a median of 53 engagements, 44% lower than posts with no question mark (95), so ask when you want replies, not likes.

But look at the kind of engagement they got. Those posts drew a median of 25 replies, 2.1 times the 12 for posts with no question. Their likes fell from 72 to 22. Since likes make up 83.8% of all engagement, the total goes down even as the conversation goes up.

Posts with a question somewhere in the middle landed in between, at 64. So a question is a tool, not a trick. Use one when you really want answers, and expect fewer likes.

Medians per post. "Ends with a question" means the last character is a question mark, not counting emojis or spaces after it.

Do emojis help Threads posts?

Not really: posts with an emoji had a median of 71 engagements, 20% lower than posts without one (89).

The gap shrinks a lot in the trending-only view, though: 72 with an emoji and 76 without, just 5% lower. So emojis are not doing much for top posts either way. Use them when they fit your voice, and skip them when they don't.

Do hashtags work on Threads?

Hashtags made little difference: posts with a hashtag in the text had a median of 83 engagements, about the same as posts without one (84).

In the trending-only view, posts with a hashtag did better, 84 against 74 (14% higher). So a tag does not seem to hurt. One caveat: we can only see tags typed into the text. A topic tag added with the Threads tag picker does not show up there, so some posts in the "no hashtag" group may still have had one. For how topic tags work and how to pick yours, read do hashtags work on Threads.

What counts as engagement on Threads?

Mostly likes: they made up 83.8% of all the engagement on these top posts. Replies were 10.7%, reposts 5.5%, and quotes 0.1%.

Quote counts were only saved for 4,371 of the posts, so quotes are undercounted here. Even among those posts, quotes were only 0.7% of the total. Because likes dominate, anything that trades likes for replies, like a closing question, will lower the total even when it starts a real conversation. Keep that in mind when you compare your own posts.

How we did this

We analyzed 22,869 of the best-performing Threads posts from 14,009 creators, published from June 14, 2025 to June 13, 2026, drawn from the trending posts ElvenFlow tracks across Threads. For each one we looked at how the post was written and its engagement: the likes, replies, reposts, and quotes it earned. We took the median for every group, so a few viral posts could not skew the result. Since these are already strong posts, the study shows what good content tends to look like, not what makes a post take off. It is the same median-engagement method as our best time to post on Threads study.

We kept one copy of each post text, so a post saved twice counts once. That leaves 22,869 posts, close to the 22,799 in our timing study but not the same, since our post records have changed a little since that study ran. When a post continues as a thread, we measured the first post only. Length is the number of characters in that first post. Lines count every line with text in it. A link is a web address or a domain typed into the text. Groups with fewer than 30 posts are left out.

About a fifth of the posts did not come from the trending feed. To check that they did not drive the results, we reran every number on trending-feed posts only: we dropped the 4,354 posts we collected by following 23 high-volume accounts directly. That left 18,515 posts from 14,006 creators. Posts by those accounts that also showed up in the trending feed stayed in, so the cut removes posts, not creators. Most patterns held. Length did not, which is why we say length is not a lever.

Every result in one table

Median likes, replies, reposts, and quotes per post. "Trending only" means trending-feed posts only: we dropped the 4,354 posts we collected by following 23 high-volume accounts directly. Format uses the 18,669 posts collected since March 16, 2026.

Common questions