What a Telegram Post’s View Count Actually Does

A Telegram post that is under an hour old is nowhere near finished. We read the same posts again two and a half hours later, and again six hours after the first reading: across 9,976 pairs of readings of the same post, the ones under an hour old had gained a median of 149% by the time we came back — 113% two and a half hours later, 208% six hours later — and every single one of them had gone up. Four days later we read them once more, and on the 244 channels where those posts were still on show, posts that had been under a day old carried a median of 46% of the views they showed by then.

That is the answer to the most common question a channel owner asks about this number, and almost every published answer to it is somebody’s impression. This one is a measurement, taken on 430 public Telegram channels between 20:26:38Z on 9 September 2026 and 03:00Z the next morning, and read again on 13 September. Nothing was joined, no account was used, no private channel was touched: every page we read is one Telegram serves to anyone with a browser. 325 channels published everything the measurement needs; the other 105 are listed with their reason.

Views do not mean what the number of subscribers suggests

Start with the figure that surprises people most. Across the 325 channels, the median channel gets 0.17 views on a typical post for every subscriber it has. Not 1. Not 0.8. About a sixth.

And that number falls steeply as a channel grows. This is the strongest pattern in the data — a rank correlation of −0.67 between subscriber count and view ratio.

Broken into ten equal groups by subscriber count, the smallest tenth (30 to 62 subscribers) has a median ratio of 1.38 and the largest tenth (776,854 to 9,593,612 subscribers) has 0.02 — a seventy-fold difference across the range.

The practical consequence: a view ratio is only meaningful against channels of your own size. A 500-subscriber channel getting 0.6 views per subscriber and a 500,000-subscriber channel getting 0.6 are not in the same universe, and the second one would be extraordinary.

“More views than subscribers” is ordinary, and it is most ordinary when you are small

This is the question we found most often asked and least often answered. 64 of the 325 channels — 20% — have a median ratio above 1.0, and 94 of them had at least one post above 1.0. Among the smallest tenth of channels, 59% are above 1.0. Among channels over 50,000 subscribers, 2% are.

Nothing about that requires an explanation involving fraud. A Telegram post is visible to anyone who opens the channel or receives the message forwarded, whether or not they subscribe: forwards into other channels and groups, links shared outside Telegram, the public web preview this survey itself read, and search all produce views from non-subscribers. A subscriber count that has fallen since the post went up produces the same reading from the other direction.

So if your channel shows more views than subscribers, the data says you are in ordinary company — especially under a thousand subscribers, where it is closer to the norm than the exception.

How a post actually fills up

Here is the part no cross-sectional reading can give you. Comparing different posts of different ages tells you how posts vary; reading the same post again tells you how one post grew. We did the second thing.

The shape is clear and it has an obvious operational meaning. A post is still filling rapidly for its first few hours and is still gaining after a day. Past a week the movement is too slow to see in a few hours: only one post in six older than that moved at all between readings a few hours apart — though over days, as the next section shows, posts a week or two old were still gaining.

Several sentences on this page were corrected on 14 September 2026, after we read the channels a fourth time and checked the page against its own data again. It used to open by saying a one-hour-old post has “roughly a third of the views it will end up with”, which read a gain over the next few hours as though it were a share of a final count. It said a post “is most of the way there within a day” and “essentially finished after a week”; the four-day reading shows a day-old post had about half of what it showed four days later, and that posts one to two weeks old were still gaining. It said the 149% was reached “by the second reading”, when that figure pools the second and third readings. And it said the ordering by age holds “monotonically, every time”, which three of the five thresholds contradict.

One caveat we have to state because it bounds the numbers above: Telegram prints view counts to three significant figures, so a post showing “1.2K” cannot reveal a change smaller than about 50 views. We restricted the growth figures to pairs where the counter’s step is under 20% of the reading, which means the small gains in the older buckets are understated rather than overstated. The direction of that error is knowable and it runs against our own headline, which is why it is here rather than in a footnote.

Four days later: the same posts, read again

Readings hours apart show how fast a post fills. They cannot show whether a post that looks settled at a day old keeps moving, so on 13 September 2026, between 21:24:58Z and 21:55:30Z, we read all 430 channels once more — the same two pages per channel, the same rate limit, the same User-Agent — a median of 97 hours after the first reading. All 860 requests returned HTTP 200.

A post from the first night is still visible four days later only if it is still among the last twenty the preview shows, so this reading mostly keeps the channels that posted little in between. 244 of the 325 channels still showed at least one first-night post, which gives 3,396 pairs. Of the 81 that did not, 76 had posted enough to push every first-night post out of the preview, 4 could not be read on the fourth visit, and 1 had no usable pair. The selection is strong: the 244 posted a median of 0.43 times a day against 13.99 for the other 81, and had a median of about 1,000 subscribers against 58,808. The table below describes slower, smaller channels, and nothing in it is a statement about busy, large ones.

One check on whether the 244 fill differently, over the hours both groups can be compared — the first gap of about 2.5 hours, each post counted once: their youngest posts grew faster than the other channels’ (1 to 3 hours old: 24.9% against 17.7%), posts 6 to 24 hours old grew a little slower (3.1% against 4.3%, and 1.7% against 2.4%), and from a day old on the two groups were the same. Every age band is published in the data.

Four things follow, and the first two correct this page. A post a day old is not most of the way there. Posts that were 12 to 24 hours old at the first reading showed about half of the views they carried four days later, and posts one to two days old about two thirds. A week is not where the counting stops; a month is closer. Posts one to two weeks old gained a median of 11.2% over the four days, and posts between two weeks and a month old 4.4%. Only posts already older than a month did not move: about half of them showed exactly the same count, and 762 of those were on counters precise to the single view, so they are a real standstill and not growth the counter hid.

Every share in the table is an upper bound. It is measured against a count that was itself still rising — posts two to seven days old gained another 22.8% over the same four days — so it overstates how much of a post’s eventual total was there at the first reading. And the young rows are thin: five posts under an hour old, and eighteen between one and three hours, five of those from a single channel, because a post published shortly before the first reading had to survive four days of newer posts to still be listed. Read those two rows for direction, not size; the 6-to-12-hour and 12-to-24-hour rows, at 38 and 32 pairs, are too close to order.

The trap this sets, and why it fools almost everyone

Put the two findings together and a specific mistake becomes almost unavoidable. If you compare a post you published this morning against the posts already sitting in your channel, you are comparing a post that is a quarter full against posts that are finished.

We measured exactly that. Of the 45 channels whose newest listed post was under six hours old, that post sat at a median of 25% of the same channel’s settled level — and 95.6% of them would have looked like a drop of more than half if you judged them that way.

The same arithmetic explains something that reads as decline and is not: in 228 of the 277 channels we could compare — 82% — the oldest post still visible has more views than the newest one. That is not a channel in decline. That is a young post next to an old one.

How much a channel’s own posts vary anyway

Before reading anything into a single post, it helps to know the noise floor. Within a single channel, the spread between a high-performing recent post and a low-performing one — the 90th percentile over the 10th — has a median of 2.24×. And from one post to the next, 52.8% of consecutive posts had fewer views than the one before, with 14.9% down by more than a quarter.

So a post doing half of what the last one did is a normal event that happens constantly, in channels that are doing nothing wrong. It is not a signal until it is a run, and a run means several posts compared at the same age as each other.

What this measurement cannot tell you

Four limits, stated because each is a question a reader will reasonably have.

It is not the numbers in the app. We read the public web preview. Those counters may lag or differ from what Telegram shows a channel owner in the app, and we never saw the app. It is also a different thing from Telegram’s in-app statistics screen, which has its own unlock threshold and is not visible from a public page at all.

It is not a random sample of Telegram. The channels come from public directories and listings, so channels nobody lists are absent by construction. Every figure here is of the form “across the 325 channels we could read on this date”.

It is not history. The web preview shows roughly the last twenty posts — a few hours for a busy channel, a year for a quiet one. These are recent posts, not a sample of any channel’s past.

The exact percentage at a given hour is softer than the ordering. The cross-sectional version of the age curve depends on where you draw the line between “still filling” and “settled”, and we recomputed it at five different cut-offs to see how much that mattered. It matters: the value in the under-one-hour bucket moves between 3% and 8% depending on the choice. What barely moves is the broad ordering: views rise with age at every threshold, except that the 6-to-12-hour and 12-to-24-hour buckets swap places at the three longest cut-offs. That broad ordering is the robust claim, and the two same-post tables above are the ones that do not depend on the choice at all. All five thresholds are published in the data rather than the one that flatters the story.

What we did not measure, and would not claim

This survey says nothing about bought views. It cannot: reading a public counter tells you what the counter says, not how it got there, and a channel with an unusual ratio has a dozen ordinary explanations before it has a suspicious one. We deliberately do not publish a ranked list of channels, and the channel identifiers in the data are there so the numbers can be re-derived from the stored pages and for no other reason. A distribution is a fact about a market; it is not an accusation about anyone in it.

The data

Everything above is published under novamya.com/data/telegram-view-ratios/: every response as it was served across all four readings, a fetch log with a content hash per response, one row per post, one row per channel, one row per post seen more than once, every channel we excluded with its reason, and the aggregates. analyse.py never touches the network — run it against pages/ and it rebuilds every number on this page offline.

Released under Creative Commons Attribution 4.0. Cite it as:

Novamya (2026). Telegram view-to-subscriber survey, 9–13 September 2026. Dataset. https://novamya.com/data/telegram-view-ratios/

If a figure here is wrong we want to hear it, and it will be dated and recorded on the corrections page. The standing rules this was collected under are on the research and data page, and what a view is in the first place — what Telegram counts, and what it does not — is in Telegram post views explained. The wider guide is Telegram growth.