The Telegram Ads Platform handbook — every setting, every gotcha
A complete operator's guide to Telegram's self-serve Ads Platform in 2026 — minimum budgets, targeting precision, creative limits, attribution, and what 80% of advertisers get wrong.
The Telegram Ads Platform is now self-serve, fully programmatic, and — if you read its policy doc carefully — one of the cheapest paid channels per qualified user in 2026.
This is the version of the docs I wish existed when I burned my first $5,000.
What the platform actually does
You write a 160-character text ad with a destination link (a channel, a bot, a Mini App, or a specific message). Telegram displays it as a “Sponsored” message at the bottom of channels with 1k+ subscribers. Users tap → land where you sent them.
That’s it. No images. No videos. No carousel. No remarketing. No lookalikes. Text + link.
This minimalism is the whole point. Cost per result is low because creative friction is low. Targeting is rough because of the text-only constraint.
The numbers up front
- Minimum budget: €2 per ad per day (was €2,000 lifetime in 2023, lifted in 2024).
- CPM range: €0.5 – €4 in most categories. Crypto, finance, gambling: €5 – €20.
- Click-through rate: typically 0.4% – 1.5%. Top quartile ads hit 3%+.
- Cost per subscriber gained (channel growth): €0.08 – €0.40 in EU/EN markets, €0.02 – €0.10 in LATAM/MENA.
- Cost per Mini App install: €0.10 – €0.50 with good targeting.
For comparison, Meta Ads typically delivers Telegram channel subs at €0.40 – €1.20. Telegram Ads is 3-5× cheaper in most categories.
A 10-minute screen-share walkthrough
If you’d rather watch the dashboard than read about it, this third-party tutorial walks through the full ad creation flow:
Account setup
Create an Ads account
Go to core.telegram.org ads.telegram.org ↗ . Sign in with your Telegram account. Add a billing method — credit card or TON wallet. The account is global; one account can run ads in any country/language.
Top up with €100 minimum to unlock most features (some category targeting requires €500+ on the balance).
Pick the destination
Each ad routes to one of:
- Channel —
@yourchannel(most common for growth) - Public group — same syntax (rarely used; groups don’t grow well from ads)
- Bot —
@yourbotwith optionalstartparameter for attribution:@yourbot?start=ad_q2_a - Bot with command —
@yourbot?start=then a value you parse server-side - Specific message —
t.me/yourchannel/123(for promoting a single post) - Mini App —
t.me/yourbot/app?startapp=ad_q2_a(withstart_paramfor attribution)
The start_param mechanism is the only attribution Telegram gives you. Use a unique value per ad creative. You’ll want this in step 7.
Pick targeting
You can layer:
- Languages — multi-select. Match your ad copy language.
- Topics — Telegram’s categorization of channels (Crypto, Tech, News, Lifestyle, etc.). Your ad shows in channels in those topics.
- Specific channels — comma-separated list of @handles. Most precise.
- Excluded channels — channels where you don’t want to appear (your competitors, generally).
- Country — by user’s account country.
The killer move: combine Topics (broad reach) with a specific channels list of 50-200 channels you’ve manually vetted as “audience match.” This delivers 5-10× the CTR of pure topic targeting.
Write the ad
160 characters. UTF-8, so emojis count as 2-3 chars. Markdown is not supported. URLs and @mentions inside the text body are not allowed (they go in the destination only).
What works in Q1 2026:
- Outcome + proof + soft CTA: “5,000 founders read our weekly Telegram playbook. Join free.”
- Curiosity + brevity: “We tested 47 cold-email tools. Only 3 are still live in 2026.”
- Calendar trigger: “April changes to Telegram Ads — what’s new for advertisers.”
What doesn’t:
- ALL CAPS, screaming punctuation
- Vague benefits (“best community ever”)
- Salesy language (“Buy now! 50% off!”)
Telegram’s review team rejects pretty aggressively. Plan for 1-2 rounds of “policy violation” rejections on your first ads.
Set bid and budget
Two modes:
- CPM mode — set max CPM you’ll pay. Telegram serves up to your daily budget at or below that bid. Good for steady growth.
- CPC mode (added late 2024) — set max CPC. Telegram only charges on click, but throttles delivery. Good when CTR is uncertain.
Start CPM mode at €1 CPM for general categories, €4 CPM for crypto/finance. Budget €5/day per ad to gather signal — €100/day per campaign once an ad is proven.
Wait for review
Ads go through manual review. SLA is “up to 7 days” — in practice 6-72 hours. Common rejection reasons:
- Claims you can’t substantiate (“Best ROI in the industry”)
- Implied urgency that isn’t real (“Last chance!”)
- Categories Telegram doesn’t allow (gambling without license, adult, ICOs without disclosure)
- Targeting that overlaps your destination’s category in ways policy disallows (rare)
If rejected, you get a category code; rewrite and resubmit. Budgets aren’t lost.
Measure what actually happened
The Ads Platform dashboard shows: impressions, clicks, CTR, CPC, spend.
It does not show: subscribers gained, bot starts, Mini App installs, conversions. You instrument those yourself:
- Channel subscribers — diff
getChat(chat_id).members_countdaily and attribute deltas to running ads (rough). - Bot starts — read
start_paramon every/startpayload, save to your DB. - Mini App installs — same with
initData.start_param. - Telegram Premium subscribers gained — not exposed; estimate from total subs growth × Premium fraction (~5-12%).
The lack of native attribution is the single biggest weakness of the platform. Build your own dashboard from day one — operators who don’t end up flying blind by month two.
What 80% of advertisers do wrong
- Pure topic targeting. Reaches everyone in “Tech” — too broad, mediocre CTR. Always layer specific channels.
- One ad creative. You need 5-10 variants per campaign to find the 2× outlier. Cycle weekly.
- No attribution. Without
start_param, you have no idea which ad worked. You’ll over-spend on the loudest one. - Sending to a channel from a cold ad. Cold users don’t subscribe to channels they don’t know. Send to a Mini App or a bot conversation that warms them up first.
- Ignoring the 80/20 of geography. A €1 CPM in Indonesia delivers 5× the volume of the same CPM in Germany. If your product is global, lean into low-CPM markets first.
A starter campaign structure
Three campaigns running in parallel:
- Discovery — broad topic targeting, low budget (€10/day), high creative variance (10 ads). Goal: find what resonates.
- Performance — top 2 ads from Discovery + specific channels list. Higher budget (€50/day). Goal: cheap reach.
- Premium — top 1 ad from Performance + curated channels of high-paying audience. €100/day at €5 CPM. Goal: quality.
Move learnings up the funnel weekly.
Withdrawing unspent budget
You can withdraw remaining ad credit back to TON or fiat (depending on top-up method). 30-day cooling period. 1% withdrawal fee. Plan budgets in chunks of €100 to avoid micro-withdrawal friction.
Comparing to alternatives
For Telegram-channel growth specifically:
| Channel | Cost per subscriber | Quality | Speed |
|---|---|---|---|
| Telegram Ads | €0.08 – €0.40 | Mid | Fast |
| Cross-promotion (organic) | ~€0 | High | Slow |
| Boost swaps | ~€0 | Mid | Slow |
| Meta Ads → Telegram link | €0.40 – €1.20 | Low | Fast |
| Influencer mention | €0.20 – €2.00 | High | Variable |
Telegram Ads is the right answer for fast, cheap, mid-quality growth. Cross-promotion and boost swaps remain the best for high-quality growth at scale. Mix all three — they don’t cannibalize.
Read next
The Channel Boost playbook
What you actually unlock with Telegram Channel Boosts in 2026, how to ladder boosts intentionally, and the cross-channel campaigns that actually work.
The Telegram glossary — every term that matters in 2026
A reference glossary of every Telegram concept, API surface, and product feature you need to know — with one-sentence definitions and direct links to the official docs.
Telegram channels worth following in 2026
A curated, opinionated starting point — by category and by language — for the Telegram channels worth opening regularly in Q2 2026.