AI tools helped gather research and write the first draft of this article. I edited it and checked every definition and quote against the page it comes from, listed at the end. The cover illustration is AI-generated. Details at the end.
One article goes out on three platforms. A day later the post on X shows 2,000 impressions, the LinkedIn post 600, the Telegram post 300 views. The tempting conclusion is that X brought the most readers, and that the article was seen about 2,900 times.
Neither conclusion follows from those numbers. Each platform counts a different event under a similar name. One counts every time a post appears on a screen, including the same person scrolling past it twice. Another counts each person roughly once for several days, and adds the views of forwarded copies. A third calls its figure an estimate.
A view count is only comparable with another view count from the same platform, measured the same way. Across platforms, the numbers can be read side by side but not added, ranked or turned into one total. This article sets out what each platform's own documentation says it counts, why the gaps matter, and what an author can compare instead.
This is part 5 of the Platforms & Metrics series. Parts 1 to 4 were about adapting one article to LinkedIn, X and Telegram and about links; from here the series turns to measuring the result.
Five platforms, five definitions of "seen"
The definitions below are quoted from each platform's own help or developer pages, fetched for this article. The platforms use different words, and the words hide different rules.
| Platform | What the number is called | What one count means | Same person counted again? |
|---|---|---|---|
| X | Impressions | The post "appeared on a user's screen" [1] | Yes: "Not unique" [1] |
| Impressions | The post "was displayed on screen"; "an estimate" [2] | Yes; a separate figure, members reached, counts people once [2] | |
| Telegram | Views | A person viewed the post; forwarded copies included [3] | Not for about four days, then again [3] |
| YouTube | Views | Since 24 August 2026, a video "starts to play", all formats [5] | Each play or replay of a Short counts [6] |
| Views | A reel or video was played, or a photo or text post was on screen [7] | Yes: three looks at a photo in a day are three views [7] |
X: every appearance on a screen
X's developer documentation defines impressions as the "Count of times the post appeared on a user's screen." It adds, without qualification: "Not unique—the same user viewing twice counts as two impressions" [1].
So an X impression is an event, not a person. A reader who sees the post in their timeline, scrolls back and then opens the thread may produce several impressions. The documentation also lists impression_count among the public metrics of a post, while link clicks and profile clicks are private metrics that are "only available for posts created within the last 30 days" [1]. After a month, the most useful click data for an author is no longer available through the API.
X also has a help page about the view counter shown under posts. It returned an error when fetched for this article, so the definition used here is the one in the developer documentation.
LinkedIn: an estimate, and a separate count of people
LinkedIn's help page on combined post analytics gives two numbers. Impressions are "The number of times your post was displayed on screen. This number is an estimate and may not be precise." Members reached is the "Number of distinct members and pages who viewed your post. Repeat views aren't counted" [2].
Two details matter here. First, LinkedIn itself says its impression figure is an estimate. Second, LinkedIn's page also defines a separate count of distinct people next to impressions. That makes the LinkedIn post look smaller if you compare its members reached with X's impressions, and larger if you compare its impressions with Telegram's views. Neither comparison is fair.
The same page says the analytics graph will "populate up to yesterday", so a post checked on the day it went out shows less than it will later [2].
Telegram: roughly once per person, plus forwards
Telegram's channel FAQ describes a different rule. "Views from forwarded copies of your messages are also included in the total count." And: "Please note that these numbers are approximate" [3]. Telegram does not keep a permanent record of who has seen what: "After a short while (around 4 days), Telegram will forget that you've seen a post" and "will count you again if you navigate to it" [3]. The FAQ even notes that this is why a person alone in a private channel may see several views on a post.
In practice that makes a Telegram view much closer to a person than an X impression is, at least for the first few days. It also means a Telegram figure can include people who never subscribed, because a forwarded copy in another chat adds to the original post's count.
Telegram's channel statistics, described in its API, give more detail than the counter under the post: "views_per_post" is "total_viewcount/postcount" for the period, there is a "Views by source graph", and "Detailed statistics about number of views and shares of recently sent messages and stories" [4]. The counter under the post remains a single approximate number.
YouTube and Facebook: a play is a view
The two largest video platforms show how much a definition can move over time.
YouTube's help page on engagement metrics now says: "Beginning August 24, 2026, views are counted the moment a video starts to play across all formats" [5]. For Shorts the change came earlier. YouTube's Shorts help page says: "Starting 3/31/2025, we're updating how Shorts views are counted", and that views "will count the number of times a Short starts to play or replay, with no minimum watch time requirement." The earlier Shorts metric is "now called 'Engaged views,'" and shows "how many viewers chose to continue watching your video" [6].
YouTube also corrects its numbers after the fact. It says it wants metrics that "come from actual humans and not computer programs", may "temporarily slow down, freeze, or change your metric count, and discard low-quality playbacks", and that "It can take several days for verified metrics to fully update" [5]. A view count read on day one is provisional.
Facebook made its own change in November 2024. In the announcement, Views "tells you the number of times a reel or video was played or the number of times photo or text posts were on screen." On reels and videos, "Plays will now be called Views"; for photo and text posts, views replaced impressions, and "a person viewing a photo three separate times in the same day would count as three Views" for that photo [7]. Facebook kept a separate Reach metric and its "3-second views, 1-minute views" [7].
These platforms are not in Dapp.Expert's own distribution, but they make the general point plainly. The same word, "views", means a play on one service, a screen appearance on another, and, for YouTube's Shorts before March 2025, something that needed some watching.
When the definition changes, history changes with it
A definition change breaks comparisons on the same platform too. Views a YouTube Short collected before 31 March 2025 and views it collected afterwards were counted under different rules: the earlier ones needed some watching, the later ones count every start [6]. A total that spans that date mixes the two, so comparing it with a newer Short's total is not like with like. YouTube points to engaged views for comparing performance across Shorts [6].
The same applies to any platform that renames or redefines a metric. Before comparing this month's numbers with last year's, check whether the platform's help page announces a change between the two dates. If it does, compare only the periods on the same side of the change, or use a metric the platform says did not change.
The denominator problem
Engagement rates multiply the problem. An engagement rate is a count of reactions, comments, shares or clicks divided by something. The something varies: impressions, views, people reached or followers. The platforms' own pages used here do not settle it for you. LinkedIn's post analytics page, for example, defines reactions as "The number of times members or pages reacted on your posts" but does not define an engagement rate at all [2].
Divide one LinkedIn post's 30 reactions by its impressions and you get one rate; divide them by its members reached and you get a higher one, because the second denominator counts each person once. A rate quoted without its denominator, and without the platform's definition of that denominator, cannot be compared with anything.
A worked example
The numbers below are invented for this article and do not describe any real post.
Return to the article from the opening. One day after posting:
- X: 2,000 impressions, 14 link clicks.
- LinkedIn: 600 impressions, 410 members reached, 34 clicks.
- Telegram: 300 views, 9 shares.
- The site: 85 visits tagged as coming from social posts: 12 from X, 31 from LinkedIn, 42 from Telegram.
Read naively, X won by a wide margin. Read with the definitions:
- X's 2,000 is a count of screen appearances, with repeats [1]. It is not a number of people and should not be set against LinkedIn's 410 people.
- LinkedIn's 600 is an estimate [2], and the day's graph may not be complete until tomorrow.
- Telegram's 300 counts each person roughly once and includes forwarded copies [3]. The 9 shares mean part of that audience may be outside the channel.
- The three figures cannot be added. "2,900 views" counts some people several times and others once.
The only line in the example that uses one definition for all three platforms is the last one: visits on the site, counted by the site's own analytics with the same rules for every source. In this invented case it reverses the ranking. Telegram sent the most readers, X the fewest. That is not a general finding about the platforms; it shows why the per-platform counters are the wrong place to look for one.
What to compare instead
Three habits keep the numbers honest.
- Compare each platform with itself. A post on X against your earlier posts on X, read at the same age, uses one definition. That comparison is meaningful; X against LinkedIn is not.
- Use one external unit for cross-platform questions. Visits to the article, counted by the site's analytics with tagged links, are measured by the same rules whatever the source. Dapp.Expert's author posts tag the verified article address with utm_source set to the platform, utm_medium=social and utm_campaign set to the article's identifier. That still measures visits, not reading; whether a visit became a read is a separate question for the site's own analytics.
- Write the definition next to the number. "2,000 impressions on X (screen appearances, not people)" is a little longer than "2,000 views" and much harder to misuse in a report.
Read every number at a fixed age, and not on day one. YouTube says verified metrics can take several days [5]; LinkedIn's graph stops at yesterday [2]; X's API stops returning private metrics after 30 days [1]. A fixed check, for example after seven days, makes posts comparable with each other within one platform.
A short check before you compare
- Are both numbers from the same platform, under the same definition?
- Does the number count events (impressions, plays) or people (reach, members reached)?
- Did the platform change the definition between the two dates?
- Is the post the same age in both readings, and past the platform's correction window?
- If you quote a rate, what was divided by what?
- For a cross-platform question, are you using a unit you measure yourself, such as tagged visits?
What this article cannot tell you
The definitions here are the platforms' own descriptions; none of the sources publishes how many repeat impressions a typical post collects, so the size of the gap between impressions and people cannot be given. X's help page on the public view counter could not be opened, and whether that counter equals the API's impression_count was not confirmed. Instagram, TikTok and other platforms were not checked and are not covered. The worked example is invented. Nothing here reports figures from Dapp.Expert's own accounts.
How this article was made
| Task | What the AI did | What a person did | How it was checked |
|---|---|---|---|
| Research | Collected platform documentation into research notes | Chose the question and the sources to rely on | Definitions and quotes checked against the pages listed below [1]–[7] |
| Draft | Wrote the first draft from the notes | Edited the text, structure and worked example | A separate review compared definitions and claims with the sources |
| Images | Generated the cover photograph-style illustration | Chose the illustration; designed the explanatory diagrams | Labelled as AI-generated in the disclosure line and alt text |
| Not done by AI | — | The decision to publish and responsibility for the text | — |
Conflicts of interest: none. The author has no commercial relationship with X, LinkedIn, Telegram, Google or Meta.
Responsible for this article: Dmitriy Lastovka, author. Corrections: contact form.
Sources
All sources are official (the platform's own page), fetched on 16 November 2026.
- X Developer Platform, "Metrics" (post metrics: public, non-public, organic). https://docs.x.com/x-api/fundamentals/metrics Official.
- LinkedIn Help, "Combined post analytics". https://www.linkedin.com/help/linkedin/answer/a701208 Official.
- Telegram, "Channels FAQ". https://telegram.org/faq_channels Official.
- Telegram API, stats.broadcastStats (channel statistics). https://core.telegram.org/constructor/stats.broadcastStats Official.
- YouTube Help, "How engagement metrics are counted". https://support.google.com/youtube/answer/2991785 Official.
- YouTube Help, "Get started creating YouTube Shorts". https://support.google.com/youtube/answer/10059070 Official.
- Facebook for Creators, "Introducing Views to simplify content distribution metrics" (14 November 2024). https://creators.facebook.com/introducing-views-to-simplify-content Official.










