Every exchange publishes two rates and almost none explains which one you will be charged. The honest answer is that most traders are on the expensive side far more often than they believe, and not because of anything they decided.
An order book is a list of offers waiting to be met. When you place an order that sits in that list and waits, you have added something to the book: you are a maker, and you are charged very little or paid outright, because a book with more waiting orders is a book where everyone else gets filled at a better price. When you place an order that takes an offer already sitting there, you have removed something from the book: you are a taker, and you pay more, because you consumed what somebody else was providing.
That is the whole definition. It says nothing about buying or selling, nothing about being right, nothing about how long you hold. It asks only whether your order waited, and it decides at the exact instant the order reaches the book.
The word describes what your order did to the book, not what you intended. This is why the rate you actually pay is often not the one you had in mind when you placed it.
A market order is taker, always, with no exception. That one is obvious and most traders account for it. The three that follow are the ones that quietly move the average, and none of them feels like a decision at the moment it happens.
There is exactly one way to be certain you will be charged the maker rate, and it is to mark the order post only. That instruction tells the book to reject the order rather than execute it if it would cross. You lose the fill you wanted; you keep the rate. Most interfaces bury the option behind a checkbox in a submenu, and most traders have never used it once.
Here is ours, and it is the whole of it. There is no volume ladder to climb, no tier that unlocks at a threshold nobody tells you about, and no rate that improves because you asked nicely or because somebody noticed you. The fourth column is the share of what you paid that comes back to you every month.
| Market | Maker | Taker | Comes back |
|---|---|---|---|
| Spot | 0.100 % | 0.100 % | 0.0100 % |
| Perpetual futures | 0.020 % | 0.060 % | 0.0060 % |
Two things in that table deserve attention. The first is that spot and futures are not charged alike, and that the gap between the maker rate and the futures taker rate is much wider than the gap on spot. The second is that the rate applies to notional value and not to your capital: with leverage, a position five times the size of your account is charged as a position five times the size of your account.
Take a trader with fifty thousand of capital, turning over three times that in monthly volume, all of it on perpetual futures. Those are illustrative round numbers chosen because they are easy to scale, not a claim about anybody in particular. That is one hundred and fifty thousand of notional a month, one point eight million across a year.
Charged entirely at the taker rate, then charged entirely at the maker rate, the two annual bills are not close to each other. The difference is not a rounding error at the bottom of a statement, it is a recurring cost that weighs on every strategy living off a thin edge. A trader who nets a small percentage per position and pays the taker rate on both legs hands back a serious slice of that edge before the market has done anything at all.
The calculator on this site runs that arithmetic with your numbers rather than these. It uses the taker rate from beginning to end, because a market order always is one and a limit order can become one without warning.
In order of how much they move the number, and the first one is the least popular thing anybody in this industry will tell you.
Fees are charged per round trip, so the bill follows the number of round trips and not the size of your account. Halving the number of positions you open in a month halves the fee bill exactly, whatever else changes. No fee schedule anywhere will do as much for you as that, and no exchange has any reason to say so out loud.
Not every strategy can wait. A momentum entry cannot, and a stop cannot by construction. But a mean reversion entry, a scale in, a partial take profit and most exits that are not emergencies can all sit in the book and wait their turn. Marking those post only costs patience and nothing else at all.
Leverage does not change your rate, it changes the base the rate is applied to. Doubling leverage on the same account doubles the fees for the same number of trades, with no line anywhere announcing it. This is the detail people forget when they compare two venues on the published rate alone.
A rebate on fees paid is the only lever that works without changing anything you do. It does not make a bad entry good, it does not turn a losing month around, and anyone telling you otherwise is selling something. What it does is reduce the fixed cost of operating, every month, for as long as you keep trading.
The maker and taker split did not start in crypto. It came from equity markets in the late nineties, where venues competing for order flow discovered that paying people to leave orders in the book was the cheapest way to attract the volume that made the book worth using. Crypto inherited the idea wholesale, and then pushed it further than any regulated market ever has.
Three things about crypto make the gap between the two rates wider than it is elsewhere. The first is that the market never closes. There is no opening auction to concentrate liquidity and no closing print to settle against, so the book has to be populated continuously by people who have a reason to be there at four in the morning. Paying makers is that reason.
The second is fragmentation. The same asset trades on dozens of venues that share no order flow, so each of them has to buy its own liquidity rather than inherit it. The third is that a very large share of crypto volume comes from automated strategies that are indifferent to which venue they run on and will move for a few hundredths of a percent. A venue that stops paying makers competitively does not lose them gradually, it loses them in an afternoon.
This is worth knowing because it explains a pattern that confuses people arriving from other markets: why the taker rate on crypto perpetuals is often several times the maker rate, while on spot the two sit much closer together. Perpetuals are where the automated liquidity lives, so that is where the incentive to provide it has to be strongest, and somebody has to fund the difference. That somebody is whoever sends a market order.
The gap between the two rates is not an accident of pricing. It is the price the venue pays to keep a book alive around the clock, passed on to the traders who consume it.
Most exchanges run a fee rebate as a campaign: it lasts thirty days, it applies to new accounts, and it ends without an announcement. A standing share of what you were charged is a different object entirely. It is part of the account rather than bolted onto it, it is paid on the fees you actually paid rather than on a promotional estimate, and maker rebates received are subtracted from the base rather than added to it, which is the detail that separates a true number from a flattering one.
It is worth being just as precise about what it is not. It is not a discount on the published rate, which stays exactly what it says. It is not a bonus conditioned on hitting a volume target. And it is not a reason to trade more, which would be the fastest possible way to hand back more than you receive.
No. A limit order pays the maker rate only if it waits in the book. Placed at a price that crosses the spread, it executes immediately against existing orders and is charged as a taker fill. Marking it post only is the only way to be certain, and the price of that certainty is that the order may never be filled.
Yes. A stop sits on the exchange as an instruction and, when its trigger is reached, enters the book as a market order. Every stop you are hit on is a taker fill. It is one of the reasons a strategy with tight stops costs more to run than its backtest suggests.
No, it changes what the rate is applied to. Fees are charged on the notional value of the position. At five times leverage, a position opened with ten thousand of margin is charged as fifty thousand. The rate on the schedule has not moved; the base it is applied to has.
Because a resting order is useful to everybody else and a market order is not. The book needs orders waiting in it for anyone to be filled at a reasonable price, so venues make it cheap to provide them and expensive to consume them. On some venues makers are paid rather than charged, for exactly the same reason.
On some venues and on some pairs, yes: the maker rate is negative, meaning a resting order that gets filled earns rather than costs. It is not charity. A venue does it on pairs where it needs depth badly enough to buy it, and it funds that payment from the taker side. When you see one, read what the taker rate is doing at the same time.
Almost always, and usually by a lot. Spot rates tend to be higher and closer together between maker and taker; perpetual rates tend to be lower overall with a much wider gap between the two. The two products have different competitive dynamics, and no venue prices them from the same logic.
It depends entirely on how often you trade. A trader opening four positions a month will barely notice it. A trader opening four a day is paying a recurring cost heavy enough to decide whether a thin edge survives. The only honest way to know is to run your own numbers rather than trust an example.