When a channel tells you where to open your account, it usually earns something when you do, and it keeps earning while you trade. That income does not depend on whether the signals work.
Why does the channel care which exchange you use?
Because in most cases it gets paid when you sign up, and keeps getting paid out of the fees you generate afterwards. The payment arrives whether your trades win or lose.
That is the whole mechanism, and it is worth stating first because everything else follows from it. A referral arrangement pays the referrer a share of what you spend on trading fees. Spending on fees requires you to trade. It does not require you to profit. So a channel earning this way has a steady income from your activity that is disconnected from the thing you are paying attention to, which is whether the calls are any good.
None of that makes an exchange bad, and nothing below is a claim about any venue's solvency, licensing or conduct. The subject here is the arrangement between the channel and the exchange, and what it means for the person reading the channel.
What we counted
We hold archives of 39 Telegram signal channels in this repository: 23,707 messages, 16,694 of them distinct once exact duplicates are removed, spanning January 2021 to August 2026. Everything in this section comes from running one parsing script over that archive, as of August 2026. The script, its sensitivity runs and its raw output are listed at the end of this article.
Pulling every link out of the text gives 2,045 URLs. Of those, 463 point at a crypto exchange. And of those 463, 222 carry a code that identifies who sent you: an invite path, a partner path, a ref parameter, or a subdomain that exists for attribution and nothing else.
Those 222 links are spread very unevenly:
| Exchange | Attributed links | Channels using them |
|---|---|---|
| BingX | 213 | 3 |
| Bybit | 6 | 3 |
| Binance | 2 | 2 |
| MEXC | 1 | 1 |
One channel accounts for 197 of the 213. It is the same invite link posted 197 separate times across its archive, appended to signal after signal.
Widening the definition of an attribution code to include weaker evidence, such as a bare registration page or a short r= parameter, raises the total to 233 links and 7 channels, and adds WEEX and CoinW to the list. We read those weaker matches by eye and found one that is not a venue referral at all: an app-share link inside a message asking readers to vote for someone in a contest. That is why the table above uses the conservative count.
Why every number here is a floor
Four channels out of 36 sounds like a fringe practice. Before you read it as a rate, it is worth knowing what the count cannot see.
Our archive keeps four fields per message: the channel, the text, a message id and a timestamp. Telegram carries links in two places that a text dump does not preserve, inline buttons and anchor text hidden behind a word. A channel can route every reader it has through an attributed link and leave nothing in the text for a script to find. Everything above counts links that happened to be typed out in full.
We did try to put a second number on the practice by counting the instruction instead of the link, and we are not publishing that count. It does not survive its own sensitivity check. Depending on where the line falls between "open an account on this exchange through my link" and a channel selling its own membership, the same archive returns anywhere from 34 messages to 343. Reading the matches by hand, every threshold mixed genuine sign-up asks with exchange news and subscription sales. A figure that swings by a factor of ten on one judgement call is not evidence of anything, and the script that produced it is in the repository so you can see the swing yourself.
So treat 222 links and four channels as the part we can prove, and assume the practice is wider than that. How much wider, we have no honest way to say.
Does the code match the venue the channel talks about?
Usually, and the exceptions are informative.
For 5 of the 7 channels carrying a code, the exchange in the code is also the exchange the channel names most often when it posts about trading. For 2 of them it is different: the posts discuss one venue and the sign-up link points at another.
More telling than either case is what happens inside a single channel. Two of the seven carry codes for several exchanges at once. One carries codes for five: Bybit, BingX, CoinW, MEXC and Binance. Five messages in the archive carry codes for two or more exchanges simultaneously, and one pricing announcement carries three in the same post, describing them in the channel's own words as a referral offer.
We also found the same marketing message posted twice by one channel, identical text, once with a partner link to one exchange and once with a partner link to another. Read that as the channel telling you plainly what the arrangement is: the venue is yours to pick, and the attribution happens either way.
That is the clearest evidence in the archive about what the link is for. A channel that genuinely needed you on a particular exchange, because that is where it trades and where its entries are priced, would need you on that one exchange. A channel offering you five is not solving a technical problem.
What the arrangement does and does not tell you
It does not tell you the channel is dishonest. Affiliate income is ordinary, legal and disclosed by plenty of legitimate publishers, including ones whose analysis is good. Some channels state the arrangement outright, and one in our archive announces a named exchange partnership as news.
What it tells you is the shape of the channel's incentives, and there are three consequences worth holding onto.
Volume pays the referrer. Fee-share arrangements pay out of what you spend on trading fees, so more signals and more round trips mean more income, independent of results. That is a real pull toward a busier feed than the analysis alone would justify. What frequent entries do to a small target is worked through in what fees and slippage do to a 2% target.
The venue recommendation is not analysis. When a channel explains why you should use a particular exchange, that explanation is being written by someone who is paid if you agree. The reasons given may be sound. They are just not disinterested, and you should read them the way you read any other sponsored recommendation.
The channel's obligations end at the referral. This is the part that hurts people. If your account is frozen, if a promotional payout is withheld, if a migration moves your assets somewhere you did not choose, your dispute is with the exchange. The channel that sent you has already been paid and has no standing in it, and often no interest either. What actually happens to customer funds when a venue fails or freezes is a separate subject, covered in exchange risk explained.
What to do about it
Five minutes of checking, done before you open an account rather than after.
Look at the link before you tap it. An attributed link usually says so in the URL itself: /invite/, /partner/, ref=, referral=, or a host beginning partner. or promote.. Long press to see where it goes. If the channel uses a shortener or a button with no visible destination, you have learned something too, though not necessarily something damning.
Ask whether you can follow the signals somewhere else. If the answer is yes, the venue choice was never about execution, and you are free to use whichever exchange you would have chosen anyway. If the answer is no, ask what specifically breaks, and see whether the explanation is about the trade or about the sign-up.
Watch for the deposit gate. The pattern worth refusing is register through this link, deposit a minimum, then send proof to get access. That sequence makes your deposit a condition of entry and puts your money in place before you have seen anything work. It appears throughout our archives, and it is discussed further in how crypto signal scams make money.
Separate the two questions. Is this channel's analysis worth following, and is this exchange somewhere I want to keep money. They have different answers and different evidence, and a channel that bundles them is doing you no favours. On the second question, the things worth checking before funding an account are set out in 10 red flags of a signal channel and, for account security specifically, in API keys and account security.
Assume a disclosure gap. Very few of the messages we read disclose the arrangement in the post carrying the link. A handful do. Most simply say to register here.
What this does not prove
These 39 channels are not a sample of the market. They were collected because they were reachable and active, and they skew toward the promotional end of Telegram. Nothing here supports a claim about signal channels in general.
The link counts are floors and we have said why. Our archive stores message text, so a link living in a button or behind anchor text is invisible to us. Our attempt to size that blind spot by counting registration instructions instead of links failed its sensitivity check and is reported above as a failure, not as a number. We cannot resolve shortened links either, since this run had no network access to follow them.
We did not verify a single referral arrangement with any exchange, and we make no claim that a particular payout exists or what it is worth. What we observed is codes in links, which establishes attribution and nothing more. Standard published affiliate terms for this industry pay a share of trading fees, and that is the mechanism described above, though we could not open a primary source page in this run to quote specific rates.
We also make no claim about the exchanges named. They appear here because their attribution links appear in the archives, which is a fact about the channels.
The practical read
A channel that tells you where to open your account very often earns from that instruction, and keeps earning from your trading afterwards. In our archives 222 links carry an attribution code, 213 of them pointing at a single exchange, and the real figure is higher than that by an amount a text archive cannot measure.
That does not make the signals bad. It means one part of the channel's income is safe from the signals being bad, and the reader is the only person in the arrangement whose outcome depends on them working. Knowing which parts of the advice are paid for is most of what you need, and it costs about a minute per link.
Channels we can score, and what their published record actually supports, are listed at Signal Providers. Nothing here recommends any exchange, any channel or any asset.
Sources
- Referral link counts, exchange breakdown, venue matching and all sensitivity runs:
work/why-a-signal-channel-sends-you-to-one-exchange/parse_referrals.py, run againstdata/channel_dumps/(39 channels, 23,707 messages, 16,694 after exact deduplication), as of August 2026. - The withdrawn registration-instruction count, and the sensitivity run showing it moving between 34 and 343 depending on definition:
work/why-a-signal-channel-sends-you-to-one-exchange/parse_signup_ask.py, as of August 2026. - Fourteen matched links read by eye, including the false positive discarded from the headline figure:
work/_batch8-shared-research.md.