Fees are the cost merchants complain about most and understand least, and both halves have the same cause: the fees are many, small, differently shaped, and taken before the money ever reaches the account, so nobody sees them leave. The complaint is sometimes justified, sometimes a distraction from a bigger leak two lines up the statement, and the only way to know which is to itemise. This post lays out the full fee stack a Shopify store actually pays, the three mechanics that make fees larger than the headline rate suggests, the one report that shows your true blended rate, and the four levers that genuinely move it, in order of payoff.
- The stack has six members: the subscription, payment processing, the extra fee on external gateways, currency conversion, chargebacks and apps, and only some of them scale with sales.
- Three mechanics inflate the real rate above the headline one: fees are charged on the gross including VAT and shipping, the fixed part is regressive on small baskets, and split payments pay it twice.
- Your true blended rate lives in one place: the payout report's transaction list, total fees over total gross charged; on the example store's €85 order the fee is €1,62.
- The four levers, in payoff order: basket size, gateway mix, plan and rate review, chargeback prevention; and proportion matters, fees are one line of a five-line cost stack.
The full fee stack, itemised
| Fee | How it is charged | Where it hides |
|---|---|---|
| Subscription | fixed, monthly | in plain sight; belongs with fixed costs, not per order |
| Payment processing | a percentage plus a fixed amount per transaction | deducted inside the payout, never invoiced |
| External gateway surcharge | an extra percentage when checkout runs on a third-party provider | billed separately from the provider's own fee |
| Currency conversion | a percentage on cross-currency orders | inside the exchange rate applied to the payout |
| Chargebacks | a fixed fee per dispute, plus the disputed amount | weeks after the order, on the payout of a different day |
| Apps | fixed, monthly, per app | a dozen small invoices that read as nothing and sum to a salary |
Two of the six, subscription and apps, do not belong in per-order arithmetic at all: they are fixed costs, spread daily, and judged like rent. The other four attach to orders, and of those the payment processing fee is the one that matters everywhere, because it touches every single transaction the store will ever take. The chargeback is its ugly cousin: rare, but each one takes the disputed amount and a fee on top, and what a chargeback costs is always more than the order was worth.
The three mechanics that inflate the real rate
First, the base. The percentage is charged on the gross amount, the price including VAT plus the shipping you charged, so the fee is computed on money that was never yours. A rate that reads as around 2% of the price becomes noticeably more when measured against net revenue, which is the denominator every other cost on your statement uses, and the mismatch flatters the fee line in every casual comparison.
Second, the fixed part is a percentage in disguise, and a regressive one. A fixed fee of twenty-five cents is 1,25% of a €20 basket and a rounding error on a €250 one, which means small-basket stores pay structurally higher blended rates at the same headline pricing. On the example store's €85 order the whole payment fee comes to €1,62, about 1,9%; the same rate card applied to a €19 order would take visibly more than 3% once the fixed part does its work. Third, split payments: an order paid with a gift card plus a card, or any two methods, is two transactions, and the fixed part is charged on each. None of these is hidden exactly, they are all in the price sheet, but they compound quietly, and none of them appears anywhere in the dashboards, a gap that belongs to the larger family of what Shopify Analytics does not tell you.
Find your true blended rate in one report
Skip the price sheet and measure. The payout report's transaction list itemises the fee taken from every single payment; divide a full month's fees by the same month's gross charged volume and you have your blended rate, the only fee number worth acting on. Merchants who do this for the first time usually find a figure above what they would have guessed, for the three mechanical reasons above, and occasionally find genuine surprises: a currency-conversion habit nobody chose, or one gateway quietly costlier than the rest. A Shopify payment fees calculator turns any rate card into the same per-order figure in advance, which is the fast way to compare a proposed change before committing to it.
The four levers, in order of payoff
Basket size is the strongest lever and the least fee-shaped: every euro added to the average basket shrinks the fixed part's percentage take on its way to helping every other per-order cost too. Second, gateway mix: rates differ by method and by whether checkout runs on the platform's own processing, and the surcharge on external gateways changes the arithmetic per method, so measure which mix your customers actually use before assuming. Third, the plan and rate review: higher subscription tiers trade a bigger fixed cost for lower processing rates, and the crossover point is pure arithmetic on your own volume, worth redoing once a year and at every growth spurt. Fourth, chargeback prevention: clear descriptors, reachable support and honest delivery estimates are unglamorous and cheaper than every dispute they prevent.
And one anti-lever, because proportion is the real finding of most fee audits: fees are one line of a five-line variable cost stack, and rarely the largest. A store agonising over a tenth of a percentage point on processing while its goods share drifts or its parcels are mispriced is optimising the small line because it is the line that came with a villain attached. Put the fee line into a full P&L statement next to goods, parcels and ad spend, and let the sizes set the agenda; the per-product view, which of your products actually make money, shows the same proportions one SKU at a time, where the fixed fee's bite on cheap solo products becomes visible.
The one-hour fee audit
- Pull one full month's payout reports and sum the fees from the transaction lists; divide by the month's gross charged volume for your blended rate.
- Split the total by payment method to see which part of the mix is expensive, and check for currency-conversion charges you did not choose.
- Count the month's chargebacks; multiply by the dispute fee plus the average disputed amount to price the problem before deciding how much prevention it deserves.
- Recompute the plan crossover: current volume times the rate saving on the next tier, against the tier's extra subscription cost.
- Write the blended rate into your statement as its own line against net revenue, next to goods, parcels and ad spend, and judge its size in that company before acting on it.
Three questions that always come next
Are these fees high compared with alternatives? Processing costs in this range are what card infrastructure costs almost everywhere; alternatives shift the shape of the fee more often than the size, a lower percentage with a higher fixed part, or the reverse. The productive question is not whether the fee exists but whether your mix and your basket profile fit the rate card you are on, and that is answerable from your own payout data in an hour.
Should I switch gateways to save a few tenths of a percent? Compute it in euros first: the saving is your annual charged volume times the difference, and for many stores that number turns out to be a weekend of revenue, while the switch costs integration work, checkout friction risk and sometimes customer trust. Switch when the euros are real or the current provider misbehaves, not because a comparison table produced a smaller number in one cell.
Do fees come back when an order is refunded? Typically no: the processing fee stays charged even when the sale reverses, which makes a refund slightly worse than a non-sale and adds one more quiet reason returns cost more than the refunded amount. A conservative statement books it exactly that way, the fee stays on the original order's day, and treats any gateway that does credit fees as a pleasant exception rather than the rule.
Fees deserve a line, not a mythology
The fee stack is real money, taken with real asymmetry from small baskets, and it rewards an hour of measurement with a rate you can actually act on. It is also, in most stores, the third or fourth largest cost line pretending to be the first, because it is the one that arrives pre-deducted and unexplained. Itemise it, price it against its true gross base, put it on the statement in company that shows its size, and fees stop being a grievance and become what every other cost already is: a number with a name, a lever and a proportion.