How exchange fee tiers work, and why your rate can change overnight
A fee tier is a discount band that a venue assigns you based on measured behaviour, usually how much volume you traded over a recent rolling window and sometimes how much of the venue's own token you hold. Because the window rolls forward and the holdings check repeats, the measurement is re-run on a schedule rather than fixed at signup, so your rate can rise or fall without you changing anything you do.
A fee tier is a band of discounted trading rates that a venue assigns to you based on how you have behaved recently, most often the volume you traded over a rolling window and sometimes the amount of a native token you hold. Your rate can change overnight because the venue re-measures you on a schedule, and the window it measures moves forward with the calendar, so activity that qualified you last month can fall out of the count without you doing anything differently.
That single fact, that a tier is the output of a repeated measurement rather than a status you were granted, explains almost everything else about how tiers behave.
Why venues price customers in bands instead of charging everyone the same
A venue’s costs are not proportional to the fees it collects. Matching an order, settling it, and holding the capital to clear it costs roughly the same whether the order is small or large. A venue also needs resting orders on its book, because a book with no depth produces bad fills and loses customers to venues with better depth.
Both pressures point the same way. High-volume accounts cost less to serve per unit of notional and contribute the liquidity that makes the venue usable, so they are worth keeping even at thin margins. Low-volume accounts trade rarely, cost relatively more to serve, and are less sensitive to a few basis points. Charging everyone the same rate would either overcharge the accounts the venue most wants to keep or undercharge the accounts that cost it the most.
Tiers resolve this without the venue having to negotiate individually. It publishes a schedule, lets the measurement sort accounts into bands, and the discount flows automatically to whoever meets the condition. There is also a second motive when a native token is part of the qualifying test: holding the token for a fee discount takes supply off the market and ties the holder to the venue, which is useful to the venue in ways that have nothing to do with your trading.
The thing to notice is that none of this is a reward for loyalty. It is a price set from a current measurement. When the measurement changes, the price changes with it.
What the rolling window does to your rate
Most volume-based tiers are measured over a trailing window rather than since account opening. The venue looks back over a recent stretch of time, sums your notional traded, and places you in whichever band that sum reaches. Then it does the same thing again at the next recalculation.
This has three consequences that catch people out.
- Volume expires. Activity that pushed you into a band drops out of the count once it ages past the back edge of the window, and nothing replaces it unless you trade again.
- The rate is forward-looking only. When a recalculation moves you down a band, the trades you already paid for are not repriced. The new rate applies to what you trade next.
- Timing matters more than totals. Two accounts with identical lifetime volume can sit in different bands if one concentrated its trading inside the current window and the other spread it out.
Holdings requirements behave similarly but fail faster. If a tier requires you to hold a minimum amount of a token and the qualifying test is denominated in value rather than units, a fall in that token’s price can drop you below the threshold while your holding is untouched. You held exactly what you held yesterday; the measurement just returned a smaller number.
Working the arithmetic on a single tier step
All figures below are invented for the example. No venue is being described.
Suppose a venue publishes a base taker fee of 0.10% and a next-tier taker fee of 0.08%, with the better tier requiring $500,000 of traded notional in the rolling window. Say you have accumulated $420,000 so far and you are trading $200,000 of notional each month from here.
At the base rate, $200,000 of monthly volume costs you 0.10% of $200,000, which is $200. At the better rate it costs 0.08% of $200,000, which is $160. The tier is worth $40 a month at your activity level.
Now suppose you consider closing the $80,000 gap with filler trades. Those trades pay the base rate, so the commission alone is 0.10% of $80,000, which is $80. Assume a spread cost of 0.02% on each round trip, adding another $16. The gap costs roughly $96 to close, against a $40 monthly saving, so you break even in about two and a half months, and only if you keep trading $200,000 a month and the window does not age your volume out in the meantime.
Change one input and the answer flips. If you were already trading $2,000,000 a month, the same tier step saves 0.02% of $2,000,000, which is $400 a month, and the $96 of filler trades pays for itself almost immediately. The method matters more than the result: measure the saving against your actual forward volume, then subtract the full cost of manufacturing the qualifying volume, including spread. Our crypto fee calculator will run the same comparison on your own numbers.
The mistake: treating a tier as a property of the account
The specific error is reading a tier as something you reached and now have, like a frequent-flyer status, and then budgeting at that rate indefinitely. Three things follow from it.
The first is a stale cost model. Someone works out their break-even on a strategy using the discounted rate they qualified for during a busy quarter, then keeps running the same strategy at lower activity. The fee line in the model no longer matches the fee line on the statement, and a thin-edge strategy can cross from profitable to unprofitable on that difference alone while appearing unchanged.
The second is the inverted comparison. People compare venues on the best rate in the published schedule rather than the rate their own volume reaches. The top band is usually the headline, and it is usually far above what a retail account will ever sustain in a rolling window. The only number that matters when comparing is the one at your volume, which is why our comparison pages sort by band rather than by best case.
The third is volume chasing as a reflex. Trading extra size purely to hold a band only makes sense when the arithmetic above comes out positive, and it rarely does near a distant threshold, because the filler trades pay commission and spread now while the discount returns value slowly and only on future volume. A tier is worth having when your normal activity reaches it. Manufacturing it is a separate decision with its own cost.
The practical habit is to treat your tier as a reading rather than a setting. Check which band you are currently in rather than which band you once reached, note what the rate is for both maker and taker at that band rather than averaging them, and if a token holding is part of your qualification, recognise that it is a price-dependent condition and not a permanent one. See also maker and taker fees for why the two sides of a tier move at different speeds, and the glossary for the terms used here.
Education only, not financial advice, and trading carries a risk of loss.
Key facts
- A fee tier is assigned by measurement, not by application, and the venue re-runs the measurement on a fixed cycle.
- Volume is normally counted over a rolling window, so old activity falls out of the count as new activity enters it.
- Volume is usually measured as notional traded, not as deposits, profit, or account balance.
- Some venues add a second condition, such as holding a minimum amount of a native token, and apply the lower of the two qualifying tiers.
- Maker and taker rates usually fall at different speeds across tiers, so the gap between them widens at higher tiers.
- A tier can drop as well as rise, and the drop applies to future trades at the moment the next recalculation runs.
Frequently asked
How is the volume for a fee tier actually counted?
Almost always as notional value traded, meaning the size of each fill multiplied by its price, summed across the window. Both sides count: buying and selling each add to the total, so one round trip contributes twice. Deposits, withdrawals, unrealised profit, and account equity do not count, which is why a large balance with little activity sits at the base tier.
Why did my fee go up when I did not change anything?
Because the measurement window moved. A busy stretch of trading eventually ages out of the rolling window, and once it does, your counted volume falls below the threshold that was holding your tier in place. The same happens if a tier depended on a token holding whose market value fell below the required amount.
Do maker and taker fees move together between tiers?
Usually not. Maker rates tend to fall faster than taker rates as tiers rise, so the gap between the two sides widens the further up a schedule you go. This means the cost advantage of resting an order instead of crossing the spread grows as you move up.
Is it worth trading extra volume to reach the next tier?
Compare the fee saved against the fee paid to get there. The saving applies only to future volume at the better rate, while the cost of the filler trades is paid immediately and includes spread and slippage as well as commission. If the gap is large and the extra volume is small, the sum can work; if you are chasing a distant threshold, the filler trades usually cost more than the discount returns.
Risk warning: trading is high risk and most retail accounts lose money. Education only, not financial advice.