Cost Per Telegram Subscriber: The Complete 2026 Guide to a Number Most People Get Wrong

Somebody asks what it costs you to get a subscriber.

You open Ads Manager, take the month's spend, open Telegram, take the member growth, and divide. Four thousand dollars, 1,650 new members, two dollars forty. You say it out loud with confidence and everybody nods.

That number is fiction, and you are about to set next month's budget with it.

Not because you did the division wrong. Because both inputs are measuring something other than what you think. The top number includes spend that produced nothing. The bottom number includes people who arrived on their own, and excludes the ones who joined and left before you counted.

I want to walk you through what a cost per telegram subscriber figure is actually made of, the three reasons the simple version misleads, and the one variant of it that is worth making a decision on.

What cost per telegram subscriber actually measures#

A cost per telegram subscriber figure is only as honest as the two numbers you divide. In the usual version, both are wrong.

The spend is real but unallocated. You know what the account spent. You do not know which part of it produced members, because Meta reports clicks and your channel reports members, and nothing connects the two.

The member number is a net figure. Telegram shows you where the count ended up, not what happened. A month with 400 joins and 250 leaves looks identical to a month with 150 joins and no churn. Those are wildly different months and one of them is a problem.

So the standard calculation answers a question nobody asked: what did we spend per unit of net growth, from all sources combined, over a period long enough to hide everything interesting?

Three reasons the simple cost per telegram subscriber misleads#

Organic joins are in your denominator. Forwards, screenshots, someone pasting your @username into a group chat, people who found you months ago and finally joined. Every one of those makes your paid acquisition look cheaper than it is. The better your organic reach, the more flattering · and more useless · the number becomes.

Churn is netted out invisibly. If 400 people join and 250 leave, your growth is 150 and your apparent cost per subscriber triples. But you did not fail to acquire 400 people. You acquired them and lost most of them, which is a completely different problem with a completely different fix. One is a targeting problem. The other is an offer or expectation problem.

It is a single number for the whole channel. Even if the first two were clean, a channel-wide average cannot tell you which ad set to scale. The average is where good and bad ad sets go to cancel each other out.

Here is the same month, calculated three ways:

The third row is the number, and almost nobody has it, because producing it needs joins matched back to ad sets and then checked again a week later.

The cost per telegram subscriber worth deciding on#

Run this on your own figures rather than trusting anybody's benchmark.

cost per member who stays = ad set spend ÷ (attributed joins from that ad set that are still in the channel after 7 days)

Take two ad sets that look like this on a joins-only report:

• Ad set A: $900 spent, 300 attributed joins. $3.00 per join.
• Ad set B: $900 spent, 150 attributed joins. $6.00 per join.

On that report you scale A and pause B. Now add the seven-day check:

• Ad set A: 90 of the 300 are still there. $10.00 per member who stayed.
• Ad set B: 120 of the 150 are still there. $7.50 per member who stayed.

B was the better ad set the whole time, and the joins-only report told you to kill it. This is not a rare edge case in this niche · cheap joins and fast churn travel together, because the same targeting that produces effortless joins produces people with no particular reason to stay.

Those figures are illustrative, not benchmarks. Use the shape, not the numbers.

How to produce a cost per telegram subscriber you can trust#

A trustworthy cost per telegram subscriber needs three things to be true, and most operators have none of them:

1. Joins matched to ad sets. You need to know which ad set produced each join, which means routing clicks through a page you control and issuing a join link unique to each visitor. The mechanism is in Telegram conversion tracking.
2. Churn counted from a fixed point. A bot can only report on members from the moment it was added to the channel. Departures from before that are invisible, and counting them as churn from your current ads mixes two unrelated groups.
3. A seven-day window you apply consistently. Any window works as long as it never changes. Switching from three days to fourteen halfway through a quarter makes your own trend meaningless.

Once those hold, the calculation is arithmetic, and you can finally answer the question that actually matters: not what a subscriber costs on average, but which ad set is buying people who stay.

This is what the Plan produces: joins per ad set, how fast joiners leave, and who is still there a week later, alongside the landing page those clicks hit. $99 a month from your balance covers the page and one tracked channel, with additional channels at $19 a month.

If your joins-only numbers already look wrong, the upstream version of the problem · clicks that never became visits at all · is in high CTR, no conversions.

What a better number will not do#

• It will not reduce churn. Knowing precisely which ad set brings people who leave is the diagnosis. Why they leave is your offer, your pinned post and what the ad promised.
• It will not attribute everything. Joins through your public @username stay anonymous. Judge ad sets on attributed joins and treat the rest as an organic layer you do not control.
• It will not make small numbers meaningful. An ad set with 11 joins does not have a reliable retention rate. Wait for volume before killing anything on it.
• It will not survive a dead ad account. Measurement does not protect the account being measured, which is a separate problem.

The bottom line#

Cost per telegram subscriber is the number everyone quotes and almost nobody calculates in a way that could change a decision. Spend divided by member growth blends paid with organic, hides churn inside a net figure, and averages your best ad set together with your worst.

The version worth having is narrower and far more useful: what one ad set paid for each member who was still there a week later. It usually reorders your ad sets, and it frequently promotes the one you were about to pause for being expensive.

So before your next budget review, do this. Take last month's spend and your net member growth and write down the naive number. Then write down, honestly, how many of those members arrived from sources you did not pay for, and how many of the ones you did pay for are still in the channel. If you cannot answer either question, that is the real finding · and it means the number you have been reporting has been a guess wearing a decimal point.

PS. The quickest way to feel this problem is to check a single ad set's joins from three weeks ago and count how many are still members today. Most people have never looked. The first time you do, it changes which ad set you think is your best one.