Impressions, views and retweets on X: what's the difference
Direct answer: on X, an impression is the number of times the post appeared on someone's screen, and it isn't deduplicated — the platform's own documentation states that the same person seeing it twice counts as two impressions. A video view is a different counter, one that lives on the media rather than the post, and it's summed across every post carrying that same video.
The short version
- The post's public counter is impressions: the number of times it appeared on screen, without discounting repeats.
- The video has its own, aggregated counter — the same video across multiple posts adds up to a single count.
- Playback quartiles count unique users, but only the post's owner can see them.
- Reposts and quote posts are counted separately: the repost count doesn't include people who quoted with a comment.
- X doesn't publish how much weight each action carries in the algorithm.
- X's developer policy bans services that pay to generate actions on the platform.
What's the difference between an impression and a view on X?
This mix-up costs money: people buy the wrong metric thinking they bought another one. X's metrics documentation separates what belongs to the post from what belongs to the media — the post has impressions, likes, replies, reposts, quotes and bookmarks; the video has a view counter and playback quartiles. When someone says "my post got 40k views," the number they're reading is actually impressions: screen appearances, not people.
| Metric | What X's documentation says | Counts unique people? | Who sees it |
|---|---|---|---|
| Post impression | Times the post appeared on someone's screen | No — two views by the same person count as two | Public |
| Video view | Aggregated across every post that contains that video; a reposted video has one running total | Not disclosed | Public |
| Playback quartile | Unique users who played each percentage of the video | Yes, unique | Only the post's owner |
| Repost | Repost count; doesn't include quote posts | Not disclosed | Public |
Notice what's not written there. In its publicly accessible documentation, X doesn't disclose a minimum watch time for a video view to count — unlike Spotify, which publishes the 30-second threshold. We'd rather write "not disclosed" than guess at a number that circulates on marketing blogs.
Why does my post get a lot of impressions and few likes?
Because impressions are cheap and likes are expensive. An impression happens when the post crosses someone's screen while they're scrolling the timeline for another reason; a like requires stopping, reading, and deciding. The ratio between the two is the most useful number on the dashboard, and it's the ratio — not the raw count — that tells you whether the copy worked. A post with 100K impressions and 30 likes didn't get good reach: it got accidental distribution — and padding the denominator only makes that reading worse. Anyone who still needs the full counter can find the option at impressions for X posts, knowing it's the easiest metric to raise and the least informative.
What does a retweet do that a like doesn't?
There's a structural difference that has nothing to do with any algorithm: a repost republishes the content inside another timeline and can generate impressions that would never happen otherwise. A like doesn't move the post anywhere — it just adds to a counter on the original post.
The documentation adds a detail almost nobody knows: the repost count doesn't include people who quoted the post with a comment, because quotes are a separate field. A post that went viral through quotes can show a modest repost count while still having had huge distribution.
As for weight in the algorithm, the honest answer is short: X doesn't publish it, and the "1 retweet = X likes" tables floating around have no source. Anyone who wants to work each signal separately can find reposts for a specific post and likes on X posts as separate services.
Which metric should you buy for each goal?
The right question isn't "which one is strongest," it's "what does the person judging me look at first."
- Empty profile: the bottleneck is the bio and the count of followers on your X profile, not the post.
- Post with traction but little social proof: likes and reposts.
- Published video: the visible counter is the video's own, addressed by X video views.
- Report that calls for appearances: impressions.
The public price list is open with no sign-up, if you want to compare first.
What this does NOT fix
A number isn't an audience. None of the five metrics brings you a customer, a reply in conversation, or repeat business — they only move what the platform displays, and nothing else.
- It doesn't fix future distribution. A post with a high count today doesn't make your next post reach more people.
- It doesn't replace frequency. A profile that posts once a month doesn't grow because of a counter.
- It doesn't protect you from platform policy: the X Developer Policy has a "Pay to engage" section, which bans paying people to perform actions and bans selling or receiving payment for posts, follows, reposts, likes, and replies. No service eliminates that risk.
- It doesn't fix a profile with no clear pitch. If your bio doesn't say what you do, a bigger counter only makes it stranger.
It's the same reasoning as our piece on the real risk of buying followers: the product has a use, and that use is narrow.
Frequently asked questions
Are impressions and views the same thing on X?
In terms of the public counter, yes: the number that shows up under the post is a count of how many times it appeared on someone's screen. X's documentation calls this field impression_count and notes that it isn't deduplicated — the same person seeing it twice counts as two. Video is the one with a separate counter, on the media itself, not on the post.
Does X publish how many seconds someone needs to watch for it to count as a video view?
Not in the public documentation we were able to read. X documents that a video view is aggregated across every post containing that video, and that there are playback quartiles (25%, 50%, 75% and 100%) with unique users. A minimum time threshold, like the 30-second one Spotify publishes for music, isn't stated.
Is a retweet worth more than a like for reach?
X doesn't publish how much weight each action carries in the algorithm, so anyone quoting a number is guessing. What is structural and observable: a repost puts the post on another audience's timeline and can generate new impressions, while a like just adds to a counter on the original post. The X metrics documentation also notes that reposts and quotes are counted separately.
Does buying impressions actually make the post show up for more people?
Raising a counter isn't the same as distributing the content. The number goes up, the social proof changes, but nothing guarantees that people interested in the topic will actually start seeing the post. If the goal is real, recurring reach, the math is different: content, frequency and engaging in the conversation.
Does X ban buying engagement?
X's developer policy has a section called Pay to engage that states a service must not compensate people to perform actions on the platform, because that creates inauthentic engagement, and it bans selling or receiving compensation for posts, follows, reposts, likes and replies. It's X's rule, not ours, and it exists — go in knowing that.
The next step
Open the analytics panel on your own post and compare impressions with likes. If the ratio is bad, the problem is the copy, and no counter fixes copy. If it's good and the post has stalled for lack of social proof, then it's worth checking what's on the X services page and reviewing the refill window listed for each service, because it varies.
Read also: Guides by social network
Other articles in this series, each covering a different metric or platform.