⎯ TL;DR
  • Estimated ~40 accounts/min banned on Telegram from mass outreach patterns. 57,600/day. 21M/year.
  • Based on our observations, 92% of bans fall into 5 categories, each with a specific triggering pattern.
  • 34% — DC proxies. 28% — burst sending. 18% — text hash collision. 9% — fingerprint fail. 3% — instant start.
  • The remaining 8% — edge cases: recipient complaints, privacy settings, keyword filters, geo mismatch.
  • All 5 categories are preventable. You need infrastructure, not "luck".

Telegram doesn't publish official ban statistics. But there are indirect estimates based on:

The figure ~40 accounts/minute is an estimate, but a plausible one for the active outreach ecosystem. Telegram has 950M+ MAU, and according to various estimates 100-500K accounts are engaged in mass outreach at any given moment. At a detection rate of 5-10%/month, that's 10-50K bans/day, or 7-35 per minute.

01 · Distribution

5 categories by frequency

#Category% of bansMitigation
1DC proxies / no proxy~34%Residential pool
2Burst sending (>100 msg/hour)~28%Log-normal rate-limit
3Identical text hash~18%Spintax + LLM
4Fingerprint collision~9%Per-account unique fingerprint
5Instant start + burst~3%3-5 days of behavioral warmup
92% totalPreventable
6User complaints (report spam)~5%Offer quality
7Privacy restrictions (recipient blocks)~2%Filter targets
8Keyword filters (banned content)~1%Avoid restricted topics
02 · Category 1

34% — DC proxies or no proxy

The most common cause — a media buyer tries to save money by buying datacenter proxies ($2-3/month/account) instead of residential ($10-15). Telegram sees the ASN in DC range → priority scan → catches other signals faster.

Another 5-7% in this category — working without any proxy at all, from a single public IP. This is an instant group flag when 5+ accounts run from one IP.

Mitigation: switch to residential proxies via Bright Data / Oxylabs / Soax. Cost $8-15/month/account. Pays off after the first account that doesn't get banned.

03 · Category 2

28% — Burst sending

When a script sends 100+ messages per hour, Telegram starts with FLOOD_WAIT. After 3-5 of those in a short window — PEER_FLOOD. After a few consecutive PEER_FLOODs — a ban through @SpamBot (if there are complaints) or simply restricted state.

Typical mistakes:

Mitigation: log-normal rate distribution (median 90s, sigma 0.7). Spread across 3-4 time windows per day. Long "lunch" pauses every 8-12 messages.

04 · Category 3

18% — Identical text hash

Telegram hashes the body of every message. At 50+ identical hashes in a short period — content cluster flag.

Mitigation: spintax with 4-5 variable blocks + LLM rewriting. Covered in detail in the article on conversion.

05 · Category 4

9% — Fingerprint collision

50 accounts with identical device_model="Desktop", app_version="4.8.11", system_version="Windows 10" → SpamBot sees a farm.

Mitigation: fingerprint randomization from the real Telegram user population, unique per-account, stable across sessions.

06 · Category 5

3% — Instant start after SIM activation

Fresh SIM → authorization → outreach begins within 24 hours. A perfect bot profile for Telegram.

Mitigation: 3-5 days of behavioral warmup. Covered in detail in the article on warmup.

07 · Conclusion

8% uncontrollable

The remaining 8% — categories that can't be solved with infrastructure:

These 8% are the natural baseline. You can't eliminate them completely, but you can reduce them: a better offer → fewer complaints; correct targeting → fewer privacy blocks.

«92% of bans come from infrastructure. The other 8% from offer and targeting quality. That distribution hasn't changed in 3 years.»
⎯ download

TG:ON for macOS · Windows · Linux

Desktop app, 160 MB. Runs locally, your keys stay yours. 3-day trial, no credit card.

Download for free
⎯ close 92%

5 categories → 5 TG:ON features.
3-day trial.

Residential proxies, smart rate-limiter, spintax+LLM, unique fingerprints, behavioral warmup. Closes 92% of ban causes with one checkbox.

Start trial