How to check a Telegram channel for fake subscribers before buying ads

Updated July 9, 2026 · 6 min read

Buying a post in a botted channel is renting a billboard in an empty field: the invoice is real, the audience isn't. Subscriber counts are the single easiest number to fake on Telegram — a few dollars buys thousands of "members" — so the count on the channel page tells you almost nothing by itself.

The good news: fake audiences are surprisingly bad at imitating real ones, and ten minutes with free tools catches most of them. Here are the four red flags, the checklist, and — just as important — what vetting can't catch.

Why this is the most expensive mistake in channel ads

As we showed in our pricing guide, the honest way to price a post is CPM — cost per thousand real views. Every fake view silently raises your true CPM: pay $100 for a post "seen" by 5,000 bots and 500 humans, and your real CPM isn't $20, it's $200. A botted channel doesn't make your ad cheaper or pricier — it makes it worthless at any price.

The four red flags

1. The views-to-subscribers ratio is out of line

The first thing to check, and checkable without any tools: open the channel and compare post view counters against the subscriber count. In a healthy larger channel, a recent post is typically viewed by 10–30% of subscribers (small channels run higher). The classic botted profile is the inverse: a channel with ten times more subscribers than post views bought those subscribers somewhere.

Check several posts, not one, and include older ones: when Telegram periodically purges bot accounts, old posts keep their inflated counters while new ones collapse — a visible cliff between last month's and this week's views is a tell of its own.

2. The growth chart has spikes with no explanation

Real channels grow in a recognizable rhythm: steady organic drip, plus sharp bumps on the days they ran promotions — bumps that coincide with mentions elsewhere. Analytics services (TGStat, Telemetr) plot both the subscriber curve and the citation history, so you can line them up.

Suspicious patterns:

• A spike with no citation trail. Thousands of new subscribers on a day nobody linked to the channel — they were delivered, not attracted.
• Subscribe-then-purge cycles. Bulk arrivals followed by mass drops days later (Telegram's cleanups deleting the bots), repeating in waves.
• Uniform, metronome growth. Real subscriptions cluster after posts and promos and slow at night; identical increments hour after hour are a machine at work.

3. The view curve of a post is the wrong shape

Genuine post views follow a lifecycle: a burst in the first hours as the audience's phones light up, then smooth decay, with visible dips overnight and a small morning recovery. Purchased views look different: flat hourly increments with no day/night rhythm, or an abrupt two-hour surge that then stops dead — the boost service completing its order.

TGStat and Telemetr both chart views per post over time; one glance at three or four recent posts is usually conclusive.

4. Engagement is hollow

Bots subscribe and view; convincingly reacting is more expensive, so engagement is where fakes run thin:

• Reactions and comments should be diverse and spread across the day, tracking the view curve. All engagement landing in the minutes right after publication is a scheduler, not an audience.
• Read the comments. Real audiences ask questions, argue and go off-topic; bot comments are generic, repetitive or absent despite big view numbers.
• Check the citation index: a channel that has grown large yet has never been mentioned by any other channel grew in a vacuum — which doesn't happen organically.

The 10-minute pre-purchase checklist

1. Open the channel's card in TGStat or Telemetr (both have English interfaces — tgstat.com, telemetr.io). If the service shows a fraud warning label, stop here.
2. Growth chart, 6-month view: look for unexplained spikes, purge cycles, metronome growth.
3. Compare growth spikes against the citation history: every big bump should have a matching mention somewhere.
4. Views-to-subscribers ratio across the last 10 posts — and against posts from 2–3 months ago (look for the purge cliff).
5. View curves on a few recent posts: organic decay with night dips, or flat/step patterns?
6. Read 20 comments and skim reaction variety on recent posts.
7. Ask the owner for a screenshot of native Telegram channel stats (growth sources, languages, views). Cooperative owners share readily; treat it as supporting evidence — a screenshot can be doctored, the analytics history can't.
8. Cross-check in a second tool if the placement is expensive — five extra minutes.

Pass all eight and you're dealing with a real audience at whatever price the math from the pricing guide says it's worth.

What vetting can't catch

Two honest limits. First, freshly botted channels with slow-drip services can pass a casual check — which is why expensive placements deserve the two-tool, old-posts, comments-reading version, not the glance. Second — and this is the bigger one — a real audience doesn't make the deal safe. A perfectly clean channel can still take your prepayment and never post, post and delete early, or swap your link. Audience quality and deal delivery are separate risks: vetting handles the first, and only the deal's structure handles the second — we covered those scam patterns in the safe-buying guide.

Where Adpact fits

Vetting the audience stays your job — bring the TGStat data. What Adpact adds sits on the other side of the line:

• A verifiable track record. Channels on the marketplace accumulate ratings from advertisers on completed, verified deals — feedback tied to placements that provably happened, not testimonials in a DM.
• A deal that can't evaporate. Your TON locks in a per-deal escrow contract; the bot publishes the exact approved creative itself and verifies it stayed live as agreed — or your money comes back automatically. The channel you vetted is the channel that delivers, or you don't pay.

Vetted a channel and want the deal itself to be as clean as the audience? Open Adpact in Telegram — browsing is free.

Quick answers

What's a normal view-to-subscriber ratio? Roughly 10–30% on recent posts for larger channels; small channels legitimately run much higher. Dramatically lower — especially 10× — means bought subscribers.

Can a channel fake everything — views, reactions, comments? At small scale, yes, briefly. Faking the history — months of coherent growth, citations, view curves and live comments — is what's economically impractical, which is why the checklist leans on history.

Are TGStat and Telemetr free? Core channel cards, growth charts and citation data — yes; both sell deeper analytics. For most ad-buying checks the free tier is enough.

The owner sent me a stats screenshot — is that proof? Supporting evidence only. Screenshots are trivially edited; third-party analytics history isn't.

Does escrow protect me from a botted audience? No — and nothing else does either, except vetting before you buy. Escrow protects the delivery: right channel, agreed creative, agreed retention, or automatic refund.

Practices current as of July 2026.