How to Analyze Your Bluesky Posts

Post analysis on Bluesky has one structural quirk you have to design around: there is no impressions number. The AT Protocol does not record how many feeds a post landed in, so the usual "engagement rate = engagements ÷ impressions" calculation has no denominator.

That is less limiting than it sounds, provided you read the four counters you do have correctly. They are not interchangeable.

What each metric actually tells you

Likes are the cheapest signal. A like costs the reader nothing and commits them to nothing. High likes with nothing else usually means the post was agreeable rather than interesting — people nodded and moved on.

Reposts are the distribution metric. A repost puts your post into somebody else's followers' feeds, so reposts are the closest thing to a reach multiplier that Bluesky exposes. If you care about growth, this is the number to watch.

Replies are the conversation metric, and the most commonly misread. Replies indicate the post gave people something to say — which can mean it was genuinely engaging or that it was contentious. Read the replies before celebrating the count.

Quotes sit between reposts and replies. A quote both distributes the post and adds commentary, which makes it the strongest signal that a post landed with real force. Quotes are also where criticism tends to live.

The ratios worth watching

Without impressions, ratios between the four counters carry the information:

• Reposts ÷ likes — how shareable a post was, independent of how much people liked it. A high ratio means the post travelled.
• Replies ÷ likes — how much conversation it started per unit of approval. Unusually high often means the post was divisive.
• Quotes ÷ reposts — whether people wanted to add something. A high ratio means your post became a prompt.

Track these against your own baseline, not against anybody else's. Absolute numbers are meaningless across accounts of different sizes; ratios are comparable across your own posting history.

Finding the posts that worked

The mechanical problem is that Bluesky shows you post metrics one post at a time. Answering "which of my last hundred posts performed best" means opening a hundred posts and writing numbers down.

Dopplersky records post metrics on a schedule and sorts them for you, which turns that into a question you can actually ask. It also captures something a manual check cannot: how engagement accumulated over time. A post that collected 200 likes in an hour and one that collected 200 over three days are different posts, and only a tool taking repeated snapshots can tell them apart.

A method that works

1. Pick a window — a month is usually enough to have signal without drowning in noise.
2. Sort by reposts, not likes. Reposts correlate with growth; likes correlate with approval.
3. Look at the top five and the bottom five together. The contrast is more informative than either list alone.
4. Ask what the top posts share — format, time of day, topic, length, whether they contained a link or an image.
5. Check the ratios on your best performers to work out why they worked.
6. Repeat next month. One month is an anecdote.

What not to conclude

A post with low engagement was not necessarily a bad post. Without impressions you cannot distinguish "few people saw it" from "many people saw it and ignored it" — and on a chronological-ish feed, timing does a lot of work. This is why you look at patterns across many posts rather than drawing conclusions from any single one.