An impression on LinkedIn is recorded when your post is rendered on someone's screen. That is the whole definition, and it is a lower bar than most people assume. It does not mean the post was read, that the reader stopped, or even that they consciously registered it — a fast scroll past your post in a feed counts as an impression.
This is not LinkedIn being deceptive; it is the standard definition of the metric across the advertising industry, and LinkedIn documents it. But it does mean the number on your post is the most flattering one available, and treating it as an audience count will mislead you.
What the number does and does not include
| Counted as an impression | Not counted | Commonly confused with |
|---|---|---|
| Your post rendered in someone's feed | A post that loaded below the viewport and was never scrolled to | Reach, which counts unique members |
| The same person seeing it on two separate occasions | The same render counted twice in one session | Views, which on video means something different again |
| A view from someone outside your network via a reshare | Your own views of your own post | Followers, which is an audience size not a delivery count |
| A render inside search results or a hashtag feed | An email notification mentioning the post | Engagements, which require an action |
Why a big impression number can be bad news
LinkedIn's feed tests a post on a small slice of your network first, and expands distribution if early engagement justifies it. That means impressions are partly a measure of how far the algorithm decided to push the post — and a post pushed widely that nobody engaged with is a clear negative signal, not a win.
The practical read: 10,000 impressions with 12 reactions is a worse outcome than 900 impressions with 40 reactions and 6 comments. The first tells you the feed gave you a chance and the post did not take it. The second tells you a smaller audience found it genuinely worth responding to, which is what earns wider distribution next time.
- Track engagement rate — total engagements divided by impressions — rather than impressions alone. It is comparable across posts of wildly different sizes.
- Comments are worth more than reactions as a signal, both to the algorithm and as evidence the post landed.
- One viral post distorts a monthly average badly. Look at the median post, not the mean.
- Compare a post against your own recent posts, never against someone else's screenshot. Impression counts scale with follower count and say nothing on their own.