Telegram Ads bought channel by channel, officially

The official Telegram ad platform gives you 160 characters, one link and no interest graph. There are no lookalikes, no behavioural segments and no pixel-fed optimisation. You bid a CPM against named public channels, topics, countries and languages, your message appears under posts in channels above a thousand subscribers, and anyone paying for Telegram's subscription tier never sees it at all. Everything that decides whether the buy works therefore happens before the campaign is created. What you are actually buying: A list of channels. I build that list by hand, reading the channels rather than sorting them by subscriber count, because subscriber counts on Telegram are the least reliable number in the medium. Chosen well, a list of forty specific channels in the automotive or beauty space beats topic targeting at the same bid, and I can tell you which forty and why. Exclusions matter as much: engagement-farming channels and reposting aggregators sit inside every topic and quietly absorb a third of an unmanaged budget. The 160 characters: There is no image doing the persuading and no headline separate from the body. One sentence has to carry the proposition and the reason to tap. I write six to ten variants per channel cluster, run them against each other at the same bid, and keep the two that survive. Copy in Russian, Arabic, Farsi, Italian or Spanish comes from native writers I brief and pay, and the invoice names them. Ad review is manual and rejects for tone as well as content, so I write to the reviewer's standard and keep a record of what has cleared. The destination decides the price: A link into a channel or a bot compounds. A link to a website is a single visit that Telegram's in-app browser will not attribute cleanly, and no reporting I produce will pretend otherwise. So I push toward a channel or a bot with a real reason to stay subscribed, and I measure joins and their retention at thirty days rather than clicks. Categories I decline: Token presales, trading signals, casino and betting, and anything routing to a bot that takes deposits. These are the largest spending categories on the platform and I have no interest in them. Also declined: buying through grey resellers who quote a guaranteed CPM, and brokered sponsored posts negotiated with individual channel owners. That second market is legitimate and some brands do well in it, but it is a different trade with different risk, and if that is what you want you should hire someone who does it openly rather than have me do it quietly. Not included: The media deposit with Telegram or with the official reseller, which you fund and which stays in your name. Bot development, though I specify the flow and test it. Channel content production after launch. Any commitment about a website conversion rate. Who this is not for: Advertisers who need last-click revenue attribution to justify a line item. Brands with no Telegram presence and no intention of building one, since the ad then rents attention it cannot keep. Anyone hoping to spend a four-figure sum and learn something durable; the deposit floor alone makes that a false economy. What arrives weekly: Spend by channel and by cluster, cost per join, thirty-day retention of the joins from each cluster, the copy variants that survived, and the exclusions I added. Once a quarter I write the harder note: whether Telegram deserves the next quarter's money at all, given what your other channels returned.