Ecommerce platform fees explained, line by line

A $40 order does not put $40 in your bank account. Seven separate lines can take a cut before the money lands: the platform subscription, card processing, the platform's own transaction fee, app subscriptions, a theme, a domain, and business email. Some are percentages, some are flat amounts per order, and some are fixed monthly costs that hurt most when you are smallest.

This post is arithmetic rather than opinion: swap in your own figures and you get an answer that applies to you.

The seven lines, and when each one hurts

The ordering matters more than the amounts. Two of these lines scale with sales and five do not, which is why the same store can look cheap at 400 orders a month and brutal at 25.

Line 1 — the platform subscription

The number on the pricing page. It is the most visible line and almost never the largest one. It is also the line people over-optimize: saving $20 a month on a subscription while paying an extra 2% on every order is a bad trade above about $1,000 in monthly sales.

Line 2 — payment processing

Whoever moves the money charges for it, and the charge almost always has two parts: a percentage and a flat fee per transaction. This is a cost of accepting cards, not a cost of the platform, and no platform makes it disappear.

The flat part is the one people miss. At an illustrative 30 cents per transaction, a $40 order gives up 0.75% to the flat fee and a $9 order gives up 3.3%. If your average order is small, the flat fee is your problem, not the percentage.

Line 3 — the platform's own transaction fee

This is a separate charge some platforms add on top of processing when you use a payment gateway that is not theirs. It is pure platform margin: it pays for nothing in the transaction itself. It usually falls to zero if you use the platform's in-house processor, which is exactly the point of it.

Whether it applies to you depends on something you should check before anything else: whether the in-house processor is even available in your country and for your category. If it is not, the fee is not avoidable, it is your default. Low transaction fees are mostly about this one mechanism, and it deserves its own post.

Line 4 — apps and add-ons

The line that grows while you are not looking. Reviews, email, upsells, bundles, a better search box, a shipping-rate calculator, a currency switcher, a popup. Each one is $10 to $40 a month and each one was justified on the day you installed it.

It is additive, so it never goes down on its own, and it is invisible when you plan, because nobody budgets for the apps they have not needed yet.

Lines 5, 6 and 7 — theme, domain, email

A paid theme is an illustrative $150 to $400 once, sometimes with a renewal for updates. Spread over two years it is small money, but it lands in the month you have no revenue.

A domain is an illustrative $10 to $20 a year and is the only line here you should not try to save on. Business email on your own domain is an illustrative $5 to $7 per mailbox per month. On mAutomate, business email and custom domains sit in Business Operations, and the custom domain starts on the $79 Grow plan rather than the $29 Launch plan.

The worked example

Assumptions, all illustrative: average order value $40, processing at 3% plus 30 cents, a platform transaction fee of 2% because this store uses an outside gateway, a $39 subscription, $60 a month of apps, a $250 theme spread over 24 months, a $15 domain, and one $6 mailbox. That makes the fixed block $116.25 a month.

Note what is actually happening. The percentage lines barely move — they are 5% of revenue at every volume. The whole swing from 17.4% to 6.5% is the fixed block being spread over more orders. Your fee percentage is mostly a statement about your volume.

That cuts both ways. A platform that looks expensive is cheap at 400 orders and a platform that looks cheap still takes a sixth of your revenue at 25 orders, because at 25 orders almost any fixed cost is large.

What dropping one line does

Same store, same everything, except the 2% platform transaction fee goes away because this merchant uses the platform's own processor.

At 400 orders that one line was worth $320 a month, which is more than twice the entire fixed block. This is the line to interrogate before you sign anything, and it is the line nobody asks about because it does not appear on the pricing page next to the subscription.

The same arithmetic with mAutomate's plan prices

Here the subscription is from the plan table and everything else is still illustrative. On Grow at $79 a month, the storefront is built by the AI Storefront Studio rather than bought as a theme, business email is in Business Operations, the marketing suite is in the plan rather than in apps, and connecting a custom domain is included — though you still register the domain with a registrar, an illustrative $15 a year. That makes the fixed block $80.25.

This row also assumes no platform fee on top of processing. That is an assumption, not a promise, and you should get it in writing from any platform you are evaluating, this one included.

The honest reading of that table is narrow: it says that collapsing the software lines into one subscription moves the fixed block, and at low volume the fixed block is what the percentage is made of. It does not say the cheapest platform wins. It says the fixed block is the thing you are actually choosing.

Where this arithmetic misleads people

It ignores refunds and chargebacks. A refunded order usually returns the product cost and not all of the processing fee, and a chargeback adds a fee of its own. At a 5% refund rate your effective take rate is worse than any table here.

It ignores cost of goods. Fees are the second bite, not the first. If your product costs $18 and shipping costs $7, the $33.05 left of your $40 order at 25 orders a month is $8.05 of contribution, and the fees have eaten half of what you thought was your margin. Do the cost-of-goods arithmetic first, the way a print-on-demand setup forces you to.

It ignores your hours. A stack that costs $60 a month and takes six hours a week to hold together is not cheaper than a $149 plan unless your time is free. That trade-off is the whole subject of all-in-one versus best-of-breed.

It assumes your average order is stable. It usually is not in year one. Run the table at your worst plausible average order value, not your best.

Do this before you commit

Open a spreadsheet with four cells: your average order value, your realistic monthly orders in month six, your processor's actual quoted rate, and the fixed block of whatever platform you are considering. Then ask the vendor two questions in writing — does any fee apply on top of my processor's rate, and which processors can I use in my country.

Those two answers change the arithmetic more than the subscription price does. mAutomate's own fixed block is on pricing, and the 14-day trial needs no card, so you can test your numbers before the first bill.