Every week advertisers pay for placements in channels where most of the "audience" is bots, and buyers purchase channels whose subscriber count is a decoration. Faking a Telegram channel costs almost nothing; checking one properly costs ten minutes. Here are the seven signals that expose inflation — ordered so you can stop at the first hard fail.
1. Reach rate: views ÷ subscribers
Open the last 10 posts and divide average views by the subscriber count. Living channels sit between 10% and 40%. Below 10% the audience is dead or fake; a "reach" above 70–80% on a big channel is also suspicious — views are being bought to mask dead subscribers.
2. View velocity: the staircase test
Watch how views grow on a fresh post (or compare posts published hours apart). Organic views rise fast in the first 2–3 hours and flatten smoothly. Purchased views arrive in vertical steps — +5,000 in minutes, then flat, then another jump. One staircase is enough to walk away.
3. Reactions and forwards vs views
Real audiences react. Thousands of views with 2–3 reactions post after post means the views are hollow. Forwards are the strongest honesty signal — they are expensive to fake convincingly and rarely bought.
4. Growth history
Ask for (or pull via analytics tools) the subscriber graph. Healthy: steady growth with visible bumps after placements. Red flags: vertical spikes of tens of thousands with no public campaign behind them, and long slow decline between spikes — the signature of bot top-ups compensating churn.
5. Comments, if enabled
Real comments are messy: typos, arguments, off-topic questions, replies to each other. Bot comments are short, generic and oddly polite ("Great post!"), often within minutes of publishing. No comments at all on a huge "engaged" channel is information too.
6. Audience composition
Deeper checks need tooling: share of accounts with usernames and photos, account age distribution, Premium share, join sources. A channel where half the members are photo-less accounts registered in the same month is not an audience — it is a warehouse.
7. The advertiser test
Ask for results of past placements: clicks and joins other advertisers got. A serious admin has them and shares willingly. Refusal, or "we never tracked", from a channel that sells ads weekly — treat as a fail.
Automating the check
Manual checks work for one deal and collapse at scale. Telegrad automates the pipeline: Channel Spy pulls reach, velocity and growth analytics for channels you evaluate; every subscriber joining your own channel gets a quality score (account age, username, photo, Premium, activity); and the marketplace only lists channels that pass these checks continuously, with public ratings from past deals. Check once by hand to learn the signals — then let the platform do it on every deal.