Almost every withdrawal guide tells you to send a small test first, and plenty of them hand you a number to use — try 10 USDT. The advice isn't wrong, but it gets told so briefly that some people end up less safe after running a test, because they now believe the hard part is behind them.
So let's take it apart: what a test actually verifies, which risks it leaves completely untouched, how to size it (there's a trap here — send too little and the test itself can be lost), and when it isn't worth running at all. The short version: a test proves the route worked at that moment. It does not prove you won't make a mistake later.
What's in this guide
What one test actually proves
Stated precisely: a successful test proves that this one specific route worked at this one moment — the receiving address accepted it, the network was the right one, the memo or tag (if the destination needs one) was filled in, and the far end actually credited the money to your account.
That last clause is the part most people skip. A "Success" on the blockchain is not the same as the destination crediting you — those are two different things, with confirmations and the platform's own crediting step in between. Calling the test passed the moment the explorer turns green means you verified that the chain delivered to that address, not that the platform behind the address is willing to recognise it.
The test has passed when you can see the balance change in the destination account — not when the explorer says success. For the gap between those two, see confirmed but still not credited and how long a deposit takes.
Four things it does not prove
This is the section worth remembering. A successful test covers none of the following:
- It does not protect the second send. The test and the real transfer are two independent transactions, and in between you have to handle the address again. Clipboard-hijacking malware swaps the address at the exact moment you paste — the fact that the first paste was correct has no bearing on the second. See clipboard hijacking.
- It does not mean a large amount takes the same path. Change the amount and the platform's risk controls and manual review may behave differently. A small send sailing through is no guarantee that a large one won't be held for questions. See why a withdrawal sits in review.
- It does not lock in the fee or the timing. Network congestion moves. Three minutes during the test can become considerably longer — and more expensive — half an hour later on the same route.
- It does not prove the destination supports that coin. If the destination doesn't support it at all, the test hasn't verified anything; it has just added a small amount to the pile you now need recovered. See deposited an unsupported token.
"I already tested it." That sentence is only true about the send you just made. Treating it as a permit for everything that follows is the most common misuse of testing — and the one way a test can actually make you worse off, because you stop checking the second paste as carefully.
How much to send: the one number you can't guess
The usual approach is to hand you a figure. That has a structural problem: minimum deposit amounts are set per coin and per network, and they get adjusted as network conditions and platform policy change. Any number written into an article can be below the line that applies when you're actually sending.
And falling below that line costs more than people expect: such a deposit usually isn't credited automatically, and Binance's help centre currently states that a deposit which didn't meet the minimum cannot be recovered — unlike the unsupported-coin case, there's no application channel left open. Which means shaving the test amount down to save money can turn the test itself into a loss you have to worry about. The full picture is in minimum deposit: how low is too low.
There's only one correct method: look at the minimum shown for that network before you send, and put the test somewhere above it but low enough that losing it wouldn't hurt. For a sense of how the thresholds compare across networks, use the network comparison table — but the number that counts is always the one on the page in front of you.
After the test, change nothing
The conclusion from a test is bound to that exact set of parameters. Change any one of them for the real transfer and the conclusion no longer holds — and this step is precisely where people like to "optimise things slightly".
| What you changed for the real send | Does the test still tell you anything? |
|---|---|
| A different receiving address (even another one at the same platform) | No. The address is the thing being verified; change it and you never tested |
| Switched network at the last minute to save on fees | No. This is one of the classic ways wrong-chain transfers happen |
| Skipped the memo / tag this time | No. For a destination that needs one, omitting it is a separate kind of accident |
| Only increased the amount | The route conclusion holds, but risk control and review were never inside the test's scope |
| Waited a few days before sending the real one | The route probably hasn't changed, but re-check the address character by character — don't reuse an old clipboard |
Binance Academy's own deposit and withdrawal guidance gives exactly the two actions this table is trying to protect: "Verify addresses before sending: even small typos can result in lost funds," and "Always match the network: when depositing crypto, confirm that the network selected in Binance matches the one you're sending from." A test doesn't do either of those for you — it just lets you do them once on a small amount first.
When a test isn't worth it
The cost of testing is explicit: one extra network fee and one extra wait. What you buy with it is turning one large irreversible risk into one small one. So the test for whether to test is simple — when the cost of getting it wrong is about the same as one network fee, testing has stopped being worth it.
- Worth testing: first time sending to this address; you changed network; you changed coin; the destination is a platform account that needs a memo or tag; the amount is large by your standards.
- Not worth testing: the amount is small enough that the fee is already a big share of it; or you've used this address before and saved it — reusing a saved address beats pasting it fresh and re-testing every time.
One test protects one transfer; one saved, verified receiving address protects every transfer after it. So the valuable move after a test passes isn't relief — it's saving that address to your platform's address book (and to the withdrawal whitelist, where that's offered), so you never have to paste it again. Where those settings live, and whether they're available to you, is whatever your account's settings pages show at the time.
One note on framing. Both Binance Academy articles we checked for this piece stress the same two things — verify the address, match the network — and neither of them lists "send a small test first" as a step. That doesn't mean testing is a bad idea. It means a test is insurance you take out on yourself, not a protection the platform hands you. And since you're the one paying the premium, working out what it covers, what it doesn't, and whether it's worth it is on you — which is the whole point of this guide.
FAQ
How much should a test transfer be?
There is no universal figure. Minimum deposit amounts are set per coin and per network and they get adjusted, so any number written into an article can be below the line that applies when you send — and a deposit below the minimum usually isn't credited automatically, with Binance's current wording being that it cannot be recovered. Check the minimum shown for that network first, then put your test above it but low enough that losing it wouldn't hurt.
The test passed — do I still need to check the address on the real transfer?
Yes, and it's the single most important step. The test and the real send are two independent transactions and you handle the address again in between; clipboard-hijacking malware swaps it at the moment you paste. A correct first paste has no bearing on the second.
The explorer showed success during the test. Does that count as passed?
No. Success on-chain only means the transaction was included and executed; it doesn't mean the destination has credited you, with confirmations and the platform's crediting step still in between. The test has passed when you can see the balance change in the destination account.
Isn't testing every single transfer the safest approach?
No, and it costs you more than it returns. Once the amount is small enough that the network fee is a large share of it, the risk reduction is worth less than the fee. And for an address you've already used and saved, reusing the saved entry is more reliable than pasting it fresh and testing again.
A test proves the route worked at that moment, not that it always will. It doesn't protect the second send, doesn't mean a large amount takes the same path, and doesn't lock in fees or timing. Don't copy a test amount from anywhere — keep it above the minimum shown for that network — and change nothing about the address, network or memo afterwards. The lasting payoff isn't the test itself; it's saving the address you verified.
Read next
Minimum deposit: how low is too low Clipboard-hijacking malware swaps your address Sent to the wrong address? Is it recoverable Forgot the memo / tag on a deposit Run the address through the format checker firstSources: Binance Academy, Crypto Payments Explained (blockchain transactions are typically irreversible; a wrong address or wrong network can mean permanent loss), Binance Academy, Your Guide to Binance Deposit/Withdrawal (verify addresses before sending; always match the network), and the Binance help centre. Minimum deposits, confirmation counts, address books and whitelist features are live / policy-dependent — whatever Binance's own pages show at the time is what counts. Note that neither Academy article above lists a small test transfer as a step; both stress verifying the address and matching the network.