How to Get More Likes on Bluesky: What 14,400 Random Posts Show
To get more likes on Bluesky, post images instead of links, tag image posts, and reply to people near your own size rather than to the giants. In a random sample of posts from September 27, 2026, image posts got zero likes 33% of the time against 66% for link cards, once you account for follower count. Follower count still sets the floor: with 1–99 followers, 62.5% of posts got no likes at all.
Several of the top "how to get more Bluesky likes" guides are written by sites that sell Bluesky likes (Mediamister's first tip is literally "Buy Bluesky Likes"). The rest repeat the usual list: ask questions, post at the right time, add hashtags. None of the ones I read check whether their tips work.
(Fair's fair, since I'm about to grade other people's advice: I make Agent Sky, a Bluesky growth tool. It doesn't touch likes, but I do have an interest in you growing here. More on it at the end.)
So I pulled a random sample of posts from the whole network and checked what the liked ones had in common. Some of the standard advice held up. Asking questions didn't.
How I Measured It
I replayed Bluesky's firehose through Jetstream, Bluesky's own public stream of every new post. I took 600 consecutive new posts every half hour, 24 times, from 04:33 to 16:03 UTC on September 27, 2026: 14,400 posts from whoever happened to be posting, with no feed or algorithm choosing them. At 16:08 UTC on September 28 I read each post's likes and replies and each author's follower count from the public API, so every post had 24 to 36 hours to collect likes.
I dropped 517 posts that were deleted, taken down or otherwise unavailable, and 66 with timestamps more than an hour off. Then, same as in my earlier "is Bluesky worth it" study, I left out accounts averaging 20 or more posts a day over their lifetime (news feeds, bots, cross-posters), which removed 29.8% of what was left. That leaves 9,696 posts: 5,320 top-level posts and 4,376 replies.
Bigger accounts get more likes and also post differently, so where a table says "adjusted", I've weighted the zero-like rate to the same follower mix (1–99, 100–999 and 1,000–9,999 followers; the few 10k+ accounts aren't in the adjusted figures). A format only looks good if it did better at each follower level.
The baseline: 45.0% of top-level posts got zero likes. By follower count it was 62.5% at 1–99, 38.2% at 100–999 and 20.0% at 1,000–9,999, close to what the earlier study found on a different day (64.9%, 42.4%, 23.7%). The follower pattern held on a second day.
What Kind of Bluesky Post Gets the Most Likes?
Image posts, and multi-image posts most of all. Link cards were the worst format by a wide margin.
Top-level Bluesky posts by format. Random firehose sample, September 27, 2026, likes read 24–36 hours later, excluding accounts averaging 20+ posts/day. "Adjusted" = weighted to the same follower mix:
| Format | Posts | Zero likes (raw) | Zero likes (adjusted) | 10+ likes |
|---|---|---|---|---|
| All image posts | 1,365 | 31.1% | 33.0% | 21.9% |
| ↳ 2–4 images | 253 | 24.5% | 25.1% | 22.1% |
| ↳ 1 image | 1,112 | 32.6% | 35.0% | 21.9% |
| Video | 216 | 35.2% | 39.9% | 25.0% |
| Quote post | 495 | 36.6% | 45.3% | 9.1% |
| Text only | 2,248 | 50.6% | 45.6% | 5.1% |
| Link card | 876 | 61.9% | 65.7% | 5.1% |
- Links are the one format I'd avoid if likes are the goal. About two in three link-card posts got nothing, and links came last at every follower level. Even at 1,000–9,999 followers, 39% of link posts got zero likes, against 10% of image posts. This is the second sample in a row where links came last. Why, I can't tell from this data. I've seen no evidence Bluesky deliberately demotes links.
- More than one image helps, and hardly anyone does it. Only 18.5% of image posts used more than one picture. With 1–99 followers, 34% of multi-image posts got zero likes, against 54% of single-image posts.
- Text-only posts land in the middle. Very short ones (under 40 characters) did worse, at 50.1% adjusted. Past 40 characters, length made little difference.
Do Hashtags Work on Bluesky?
On image posts, yes. On text posts, no. This one surprised me:
Zero-like rate (adjusted for follower mix) by number of hashtags, same sample:
| Hashtags | Image posts | Text-only posts |
|---|---|---|
| None | 35.7% (n=928) | 45.6% (n=2,089) |
| 1–2 | 33.6% (n=222) | 47.2% (n=131) |
| 3–5 | 22.0% (n=149) | too few (n=26) |
| 6 or more | 18.7% (n=66) | too few (n=2) |
An image post with three to five hashtags was about 40% less likely to get zero likes than one with none. One or two tags barely moved it. Six or more did as well again, but on only 66 posts, so I wouldn't read much into that row. The effect held with Japanese-language posts left out, and on single-image posts alone. On text posts, hashtags did nothing I could measure. (Link posts did poorly with or without them: 68.0% zero likes untagged, 57.8% with one or two.)
My read, which the data supports but doesn't prove: art, photography and similar communities run on tag-driven custom feeds and hashtag searches, so a tagged image reaches people who don't follow you yet. Text posts don't have the same tag-driven feeds behind them.
This cuts against my own earlier advice in how to use hashtags on Bluesky, where I said to cap it at two or three. For image posts, three to five did clearly better than one or two. Keep them specific to your community either way.
Do Questions, Alt Text or Posting Time Get More Likes?
Not in this sample. Three pieces of standard advice didn't show up in the numbers:
- Asking a question. Posts containing a question mark got zero likes more often (51.9% adjusted, against 45.2% for posts without one). But they were more likely to get a reply: 27.3% against 19.3%, and that gap held at every follower level. Questions buy replies rather than likes, which may still be the better trade if you want conversations.
- Alt text. Image posts with alt text on every image got zero likes 38.2% of the time (adjusted), against 32.2% without. On 233 posts that's within the noise, so the fair reading is "no measurable effect on likes." Only 17.1% of image posts had full alt text. Add it anyway: it's how blind and low-vision people read your post, and it takes a few seconds.
- Posting time. Across two-hour blocks from 04:33 to 16:03 UTC, the adjusted zero-like rate only moved between 43.3% and 47.0%. But that window runs from about 00:30 to noon Eastern and misses the 1–3 PM Eastern peak that scheduling-tool data points to, so I can't test that claim. Within the hours I did cover, timing barely moved the numbers.
Are Replies the Fastest Way to Get Likes on Bluesky?
If you're small, yes. But who you reply to matters a lot. Replies to other people got zero likes 33.3% of the time, against 45.4% for replies continuing your own thread. With 1–99 followers, replying to others (45.3% zero likes) beat posting top-level (62.5%) by a wide margin.
Then I split replies by how big the account being replied to was, compared with the person replying:
Replies to other people's posts, by the size of the account replied to relative to the replier. Same sample:
| Account replied to is… | Replies | Zero likes | Zero likes, repliers with 100–999 followers |
|---|---|---|---|
| Smaller than you | 1,035 | 29.2% | 31.6% |
| 1–10× your size | 1,362 | 28.3% | 28.8% |
| 10–100× your size | 669 | 37.2% | 34.1% |
| 100×+ your size | 421 | 53.0% | 50.8% |
Replying to huge accounts is the worst place to spend your replies. Half of those replies got nothing. My guess is they get buried under hundreds of others. Replies to accounts up to 10× your size were about half as likely to go unliked, probably because those threads are small enough that your reply gets read.
How Do You Get Free Bluesky Likes?
Free ways to get more likes on Bluesky, in order of how much they moved the numbers:
- Reply to accounts no more than about 10× your size. 28.3% of those replies got zero likes, against 53.0% for replies to accounts 100× bigger.
- Post two to four images instead of a link card. 25.1% zero likes against 65.7%, adjusted.
- Tag image posts with three to five hashtags your community uses. 22.0% zero likes against 35.7% with none.
- Write more than 40 characters. Very short text posts did worse (50.1%).
- Grow followers in your topic. The median top-level post got 0 likes at 1–99 followers, 1 at 100–999, and 4 at 1,000–9,999. Every format did better at every step up.
Should you use like-exchange sites? Sites like Upvote.club give you 13 points at signup, let you earn more by engaging with other members' posts (paid plans buy more), and spend those points on likes for yours. That's still fake engagement even when no money changes hands. Bluesky's community guidelines say: "Do not artificially manipulate features or social signals to gain unearned reach or mislead users, including engagement metrics, follower counts, or other measures of community interest." Every like on Bluesky is a public record of who liked it, so a post padded by strangers is easy to spot. And those strangers aren't in your topic, won't read your next post, and won't follow you.
The slow part is item 5, and it's what I built Agent Sky for. It follows accounts that are 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 and following people in your niche by hand.
To see how your own posts compare, our free Bluesky analytics tool takes any handle and shows your engagement rate against the benchmark for your follower bucket, which post types worked for you, and your best hours.
Honest limits on these numbers. This is one day's posts over 11½ hours, and Japanese-language posts were 32.9% of top-level posts in that window, which probably overrepresents them. Likes were read once, 24 to 36 hours after posting. The adjustment only covers follower count, so other differences between people who post images and people who post links (topic, for one) are still mixed in. These are patterns in what got liked, not proof that changing one thing will change your numbers. The format, link, hashtag and reply patterns held in English-only posts too, and the follower and link patterns matched my earlier sample from a different day.
Quick FAQ
How do you get more likes on Bluesky? Post images instead of links, tag image posts with three to five hashtags your community uses, and reply to accounts no more than about 10 times your size. In a random sample of Bluesky posts from September 27, 2026, image posts got zero likes 33.0% of the time against 65.7% for link cards (adjusted for follower count), and replies to accounts 1 to 10 times the replier's size got zero likes 28.3% of the time against 53.0% for replies to accounts 100 times bigger. Follower count still matters most: 62.5% of top-level posts from accounts with 1 to 99 followers got no likes.
How can I get free Bluesky likes? Reply to accounts no more than about 10 times your size (28.3% of those replies got zero likes, against 53.0% for replies to accounts 100 times bigger), post two to four images instead of a link card, and tag image posts. Like-exchange sites such as Upvote.club give you points to spend on likes from other members, but those likes come from strangers outside your niche, every like is a public record, and Bluesky's community guidelines forbid artificially manipulating engagement metrics.
Do hashtags work on Bluesky? On image posts, yes. In my September 2026 sample, image posts with three to five hashtags got zero likes 22.0% of the time, against 35.7% for image posts with none, adjusted for follower count. One or two tags made little difference (33.6%). On text-only posts hashtags made no measurable difference (45.6% with none, 47.2% with one or two).
Why do my Bluesky posts get no likes? Usually because too few people who care about your topic follow you yet. In a random sample of 5,320 top-level Bluesky posts, 62.5% from accounts with 1 to 99 followers got zero likes a day later, against 38.2% at 100 to 999 followers and 20.0% at 1,000 to 9,999. Link cards and very short text posts make it worse. Replying to people near your own size is the quickest way out.