Traders compare platforms on the published fee because it is the only number anybody publishes. It is one of six costs a complete round trip incurs, and depending on what you trade and when, it can be the smallest of them. The other five are all measurable and none of them appear on a statement.
The trading fee is a percentage of notional, applied on entry and again on exit, with a lower rate for orders that rest in the book and a higher one for orders that cross it. It is the only component of your cost that somebody has written down in advance, which is precisely why it dominates every comparison and every discussion, and why the other five go unexamined.
| Market | Maker | Taker | Comes back |
|---|---|---|---|
| Spot | 0.100 % | 0.100 % | 0.0100 % |
| Perpetual futures | 0.020 % | 0.060 % | 0.0060 % |
Two things about this schedule matter more than the numbers themselves. The rate applies to notional rather than to the capital you committed, so a leveraged position is charged on the full size rather than on your margin. And the gap between the two sides is what makes the choice of order type an economic decision rather than a stylistic one, because that gap accrues on every trade for as long as you trade.
A perpetual futures contract has no expiry, so nothing forces its price toward the spot price the way a delivery date does. The mechanism that replaces that pull is a periodic payment between the two sides of the market: when the contract trades above spot, longs pay shorts, and when it trades below, the reverse. It is not a fee charged by the platform, it is a transfer between participants, and the platform's rate schedule has nothing to do with it.
For a position held for minutes this is irrelevant. For a position held for weeks it is frequently the largest single cost, larger than the trading fee by a wide margin, and it compounds silently because each payment is small enough to ignore individually. A trader who holds a directional position through a period of one-sided positioning can pay more in funding than the move itself was worth, and nothing in the interface presents that total until it has already been paid.
The measurement is straightforward and almost nobody does it. Record the funding paid or received per position, add it to the trading fees, and compare the sum with the gross result. Positions held longer than a day should be judged against that total rather than against the price move, and doing so changes which strategies look viable.
Any order that crosses the book pays half the distance between the best bid and the best offer, measured against the mid price, and pays it again on the way out. This is not optional and it is not a penalty: it is the standing price of immediate execution, quoted continuously by whoever is willing to be on the other side.
It goes uncounted because it is embedded in the fill price rather than deducted afterwards. The trade appears to execute at the market price, and the market price you compared it to was the mid, and the difference silently became a cost. On a liquid pair during active hours the amount is small. On anything thinner, or during a quiet hour, it can exceed the trading fee several times over while still looking like a normal fill.
If your size is larger than what rests at the best price, the remainder executes at the next level and the one after. The book returns a worse average price because you consumed the good one, and the effect grows faster than linearly with size: doubling an order more than doubles this component.
It is invisible in the same way the spread is, and it is worse in one respect. The spread is quoted, so you can see it before deciding. Impact depends on the depth that happened to be present in the second your order arrived, which nobody records, so the only way to know what it cost is to compare the price you intended with the price you received. That comparison is the single most informative thing a trader can measure and it takes two numbers per trade.
The fee is deducted. The spread and the impact are absorbed into the price, which is why they feel like the market moving rather than like a cost you paid.
Moving assets onto a platform and off it again costs a network fee, and that fee is a fixed amount rather than a percentage. On a large transfer it rounds to nothing. On a small one it can be a significant fraction of the amount moved, which means the same operational cost falls with completely different weight depending on the size of the account performing it.
The consequence is a threshold that nobody states: below a certain balance, moving capital frequently between venues costs more than any advantage the movement could produce. Traders who chase small differences between platforms discover this after the fact, having paid the fixed cost several times to capture a percentage difference on an amount too small to cover it.
The related detail is that platforms frequently charge a withdrawal fee that exceeds the network cost, and that the difference between the two is discretionary and rarely explained. Comparing withdrawal fees between platforms, in absolute terms rather than as a percentage, is worth doing once, because it is a recurring cost that scales with how often you move rather than with how much you trade.
If your capital is in one currency and the platform quotes in another, every deposit and every withdrawal crosses a conversion, and the rate applied is rarely the mid-market rate. The margin taken on that conversion is a cost, it is charged twice over the life of a position, and it appears nowhere except as a slightly worse number than expected.
It is the least discussed of the six and it is not the smallest. On a full cycle of deposit, trade, and withdrawal, the two conversions can be comparable to everything else combined, particularly for traders working outside the currency their platform quotes in. Checking the applied rate against the mid-market rate on the day, once, tells you what it costs, and the answer is stable enough to plan around.
The exercise takes half an hour and it changes how most traders allocate their attention. Take one recent, ordinary position. Write down the fee paid on entry and on exit. Add the funding paid over the holding period. Add the difference between the mid price at the moment you decided and the price you actually got, on both sides, which captures the spread and your impact together. Add the network fees if capital moved. Add the conversion margin if it crossed a currency.
The sum is what that position had to overcome before it produced anything, and it is almost always larger than the trader expected. More usefully, the breakdown tells you which component is worth attacking. For a high-frequency trader on liquid pairs, the fee and the spread dominate and the answer is passive execution. For somebody holding directional positions for weeks, funding dominates and the answer is different. For a small account moving between venues, network fees dominate and the answer is to stop moving.
Without the breakdown, everybody optimises the same thing, which is the published fee, because it is the only one they can see. That is the component least likely to be the largest, and the effort spent on it is effort not spent on the one that is.
Optimising the published fee is optimising the only line you can read. It is rarely the biggest line, and the biggest one is measurable in an afternoon.
Ranked by how much they typically save, and none of them require a view on the market. Trade when the book is deep, because the spread and impact components fall by more during active hours than any fee tier will ever save you. Rest orders rather than crossing them wherever missing the trade is acceptable, because that flips the fee to the lower side and removes the spread at once. Hold perpetual positions with an eye on funding rather than only on price, because over days it becomes the dominant term.
Then the operational ones, which are smaller and permanent. Move capital rarely and in larger amounts, so the fixed network cost is spread across more value. Keep the balance in the currency your platform quotes in, if the choice exists, so no conversion happens on either side. And size positions against available depth, because impact grows faster than size and the largest single execution cost most traders ever pay is the one they created themselves by sending too much at once.
Rarely. It is the only one published, which is why it dominates comparisons. Depending on what you trade and how long you hold, the spread, your own market impact, or funding on a perpetual position can each be larger. The only way to know which dominates for you is to measure a real round trip.
No. Funding is a periodic payment between the two sides of a perpetual contract, which keeps its price tethered to spot in the absence of an expiry date. Longs pay shorts when the contract trades above spot and the reverse when it trades below. It is a transfer between participants, not a platform fee, and over a multi-day hold it is frequently the largest cost.
Because two costs are embedded in the price rather than deducted from it. Crossing the book pays half the spread, and any size beyond what rests at the best price walks to the next level, which is your own market impact. Neither appears as a fee, and together they routinely exceed one.
Record the mid price at the moment you decided and compare it with your average fill, on entry and on exit. That difference captures the spread and your impact together. Add the trading fee and any funding paid, and you have the number the position had to overcome before producing anything.
Not necessarily. The network fee is what the blockchain charges to move the assets; the withdrawal fee is what the platform charges you, and the difference between them is discretionary. Comparing withdrawal fees in absolute terms across platforms is worth doing once, since it is a fixed cost that scales with how often you move rather than how much you trade.
It depends entirely on which component dominates your costs. For an active trader on liquid pairs paying the crossing side repeatedly, it matters. For somebody holding perpetual positions for weeks, funding will dwarf it. For a small account paying fixed network costs on frequent transfers, it is close to irrelevant.
Because two of the six components depend on how much depth is present. The spread widens and your order walks further up the book when fewer participants are quoting. The published fee does not change, which is exactly why comparing platforms on the fee alone misses the larger effect.
Choosing when to trade, if your strategy allows it. The spread and impact components fall substantially during active hours, by more than any fee tier is likely to save, and the change costs nothing to implement. The second cheapest is resting orders instead of crossing them wherever missing a fill is acceptable.