Is Bluesky Worth It in 2026? I Sampled 7,200 Random Posts to Find Out
Bluesky is worth it if you'll do the slow work of getting a few hundred people who care about your topic to follow you, and it isn't if you want reach without that work. In a random sample of top-level posts from September 24, 2026, leaving out accounts that post 20+ times a day, 47.2% had zero likes a day later. Under 100 followers it was 64.9%. Between 1,000 and 9,999 followers it was 23.7%.
The "is Bluesky worth it" reviews I read for this question answer with impressions rather than numbers. One honest-review post says the platform "now hosts around 85 million active users globally," with no source. That's about double Bluesky's registered accounts. None of them measure the thing you'd want to know before you put time in: if I post here, does anyone respond?
(Fair's fair, since I'm about to pick holes in other reviews: I make Agent Sky, a Bluesky growth tool, so I have an interest in the answer being yes. More on it at the end.)
So I counted. On September 11, 2026 I looked at engagement rate for accounts sampled from public feeds, and only 5.4% of their posts got zero likes. That's the Bluesky you see when you scroll. This time I wanted the Bluesky you'd actually be posting into, so I took posts at random from the whole network.
How I Measured It
I replayed Bluesky's firehose through Jetstream, Bluesky's own JSON feed of every new post on the network. I took 300 consecutive new posts every half hour, in 24 batches from 04:33 to 16:03 UTC on September 24, 2026. That's 7,200 posts from whoever happened to be posting at the time, with no algorithm or follow graph choosing them.
At 16:07 UTC on September 25 I read each post's likes, reposts, replies and quotes from the public API, along with each author's follower count and account age. Every post had 24 to 36 hours to collect engagement. A few more likes may come in later, but a post nobody had liked after a full day isn't about to take off.
I dropped 283 posts in all. 239 (3.3%) had been deleted or taken down by the time I checked. 46 had a timestamp more than an hour off from when they arrived, mostly link-sharing bots that stamp the article's publish time (two posts were in both groups). That leaves 6,917 posts: 3,970 top-level posts and 2,947 replies.
One more cut matters. About a third of top-level posts (34.2%) came from accounts that have averaged 20 or more posts a day over their whole life. News feeds, bots and cross-posters live there, along with some very chatty humans. They drag the numbers down, and their experience isn't yours, so the figures below leave them out unless I say otherwise.
How Many Bluesky Posts Get Zero Likes?
Almost half. Without the high-volume accounts, 47.2% of top-level posts got zero likes after a day, and the median post got one. With them it's 56.4%, and the median is zero. Follower count sets almost all of it:
Top-level Bluesky posts by author's follower count. Random firehose sample, September 24, 2026, counts read 24–36 hours later, excluding accounts averaging 20+ posts/day:
| Author's followers | Posts | Zero likes | At least 1 like | 10+ likes | Median likes |
|---|---|---|---|---|---|
| 1–99 | 1,009 | 64.9% | 35.1% | 0.8% | 0 |
| 100–999 | 1,029 | 42.4% | 57.6% | 10.2% | 1 |
| 1,000–9,999 | 481 | 23.7% | 76.3% | 25.4% | 3 |
| 10,000–99,999 | 66 | 12.1% | 87.9% | 40.9% | 6.5 |
| All (incl. 27 posts at 0 or 100k+) | 2,612 | 47.2% | 52.8% | 10.1% | 1 |
Things get better at every step up, and fastest between a few hundred and a thousand followers. Under 100, about two posts in three get no likes at all, and almost none reach ten. Split more finely, the zero-like rate goes from 67.4% at 1–49 followers to 43.3% at 300–499, then 31.1% at 500–999 and 15.3% at 5,000–9,999. At 300–999 followers, 63.2% of posts get at least one like. That's roughly where posting starts to feel like a conversation. It's also further than most people posting have got: the median top-level post in the full sample came from someone with 173 followers.
The likes that exist are extremely concentrated. Among top-level posts from the non-high-volume accounts, the top 1% collected 40.8% of all the likes and the top 10% collected 80.7%. The viral posts you see in Discover are the far end of that curve.
Buffer's much larger dataset points the same way. Its analysis of 1.7 million posts across X, Threads and Bluesky found that half of all posts got four or fewer engagements in 2024, and that Bluesky's median had dropped to three by February 2025. The useful question is where on that curve you sit.
English-language posts did a bit better than the sample as a whole (39.7% zero likes). Newer accounts did worse: 54.3% of posts from accounts under a year old got zero likes, against 45.9% for older ones.
Is Bluesky Worth It for Artists (and Which Posts Get Noticed)?
Image posts did best of the main formats. Replies were almost as likely to get a like, though they rarely got many. Same sample, same exclusions:
Top-level posts and replies by format. Bluesky random firehose sample, September 24, 2026, excluding accounts averaging 20+ posts/day:
| Format | Posts | Zero likes | 10+ likes | Median likes |
|---|---|---|---|---|
| Image post | 657 | 34.4% | 18.9% | 1 |
| Text only | 1,007 | 51.5% | 5.4% | 0 |
| Link card | 499 | 63.3% | 3.8% | 0 |
| Reply | 2,251 | 37.3% | 2.9% | 1 |
(Video, quote posts and a few rarer formats are in the "All" total above but too thin to break out.)
For artists, this is encouraging, with one catch. An image post was three and a half times as likely as a text post to reach ten likes. The catch: image posters had more followers (median 218 against 92 for text-only), so some of that gap is audience size. But images still won inside the follower groups. Under 100 followers, 54.4% of image posts got zero likes against 64.2% of text posts. At 100–999 it was 27.6% against 43.9%. Above 1,000 the two were about even. That cuts against the review I quoted above, which says to skip Bluesky if "visual content is your primary format." One likely reason images do well is that art has active custom feeds, which show posts to people who don't follow you yet. My data can't prove that part.
Link cards did worst, even with the high-volume accounts removed. 63.3% got nothing, although image and link posters had almost the same median follower count (218 and 212). Most of those posts weren't bare links, either: the median one had 126 characters of text. I'm not claiming Bluesky penalizes links, because I have no evidence it does, and I can't tell from this data why they lag.
Replies are where small accounts get noticed. Only 25.4% of replies got no engagement of any kind, against 41.3% of top-level posts. A reply lands in front of people who are already reading that thread. A top-level post from a 60-follower account only reaches the few of those 60 who happen to be scrolling. If you're under 100 followers, spend more of your time in other people's threads than on your own posts.
Who Is Bluesky Worth It For, and Who Should Skip It?
It's worth it for people who want an audience in a specific topic and will build one, and a poor fit for anyone who needs reach on day one. Here's my read, partly from the sample and partly from how the app works:
- Writers, journalists, researchers, artists and developers. These communities have their own feeds and starter packs on Bluesky, and the data says image posts do well. Once you're past a few hundred followers, most of your posts get a response.
- People leaving X who mainly want their old conversations back. The Following feed is reverse-chronological, so followers who read it see your posts in order, with no ranking in between.
- Small brands and creators, with patience. Expect weeks of replying and following people in your niche before your own posts get consistent responses. The first table shows why.
- Probably skip if you need big numbers fast. Bluesky had 41.4 million registered accounts at the end of 2025 (from its own transparency report), and far fewer use it daily (the full numbers are here, including whether Bluesky is dying). It isn't X-sized.
- Skip if your plan is auto-posting links. Link cards got zero likes 63.3% of the time, the worst of any format.
What Are the Downsides of Bluesky?
No built-in way to earn, a smaller audience, and very little privacy about what you do. The first matters most if "worth it" means money. Bluesky has no ad revenue share, creator fund or tipping; Metricool's monetization guide lists all three as missing, and I haven't found anything since that changes it. Creators who earn from Bluesky do it off-platform, through newsletters, Patreon and commissions.
The privacy point cuts both ways. Everything on Bluesky is public, including your likes and who you follow. That openness is also why I could run this study at all without logging in or asking anyone's permission: every like count above is visible to anyone.
How Do You Get Past the Point Where Nobody Replies?
Get to a few hundred relevant followers as fast as you honestly can. The first table gets better at every step up in followers, and nothing else in the data moves the numbers as much.
The work is dull, and it's what moves the numbers: reply in threads in your topic, follow people who are active there (not the biggest accounts, where one more follow gets lost in the noise), and unfollow the ones who never follow back so your following count doesn't balloon. To see where you stand right now, our free Bluesky analytics tool takes any handle and shows your engagement rate against the benchmark for your follower bucket, plus which posts and hours worked best.
The follow-and-prune part is the job I built Agent Sky to do. It follows accounts genuinely active in your niche on a paced schedule, then checks for a live follow-back before unfollowing anyone. $9 a month, free to start. It won't make a bad post good. What it saves you is the hours of finding, following and pruning by hand. (If you're stuck under 100 right now, here's why that usually happens.)
Honest limits on these numbers. This is one day's posts across 11½ hours, so US evenings are missing and Japanese-language posts, which were 22% of top-level posts, are probably overrepresented. Likes were read once, 24 to 36 hours after posting. And the 20-posts-a-day cutoff is a rough filter for automated accounts, not a bot detector. The follower-count pattern held with the high-volume accounts included and with a looser 50-a-day cutoff, which is why it's the finding I'd rely on.
Quick FAQ
Is Bluesky worth it in 2026? It is if you'll put in the work to reach a few hundred followers who care about your topic. It isn't if you want reach out of the box. In a random sample of top-level Bluesky posts from September 24, 2026 (leaving out accounts that average 20+ posts a day), 47.2% had zero likes 24 to 36 hours later. That was 64.9% for accounts under 100 followers and 23.7% for accounts with 1,000 to 9,999. Counting every account, including the high-volume ones, it was 56.4% of 3,970 posts.
Is Bluesky worth it for artists? It's one of the better cases. In my September 2026 sample, image posts did best of the main formats: 34.4% got zero likes, against 51.5% for text-only posts and 63.3% for posts with a link card. Image posts were three and a half times as likely as text posts to reach ten likes. Image posters tend to have more followers, but images still came out ahead within the under-100 and 100-to-999 follower groups.
Is it worth posting on Bluesky? Yes, but replies are the better bet while you're small. In my sample, 37.3% of replies got zero likes, against 47.2% of top-level posts, and only 25.4% of replies got no engagement of any kind. A top-level post from an account under 100 followers had a 64.9% chance of getting zero likes. Replying in threads where people already are is how small accounts get seen.
Is it worth making a Bluesky account? An account is free and takes a minute, so the real cost is the time you put into it. Newer accounts have a harder start: in my sample, 54.3% of posts from accounts less than a year old got zero likes, against 45.9% for older accounts. Expect a few weeks of building a following before posting feels like a conversation.