Every option, read off their own pages on 11 September 2026

The fee is the small question. Who owns the customer is the large one.

Two stores can cost you the same money this month and leave you in completely different positions in two years. One has your customer list on it. The other has somebody else's, and you have been renting the audience the whole time without anybody calling it rent.

I sell none of these and I take no cut of anything you sell. That is the only reason this page is worth reading rather than the comparisons you will find, which were mostly written by one of the companies on it.

Read this before the table

This prices one thing: what the platform takes.

Its monthly fee, plus its own percentage of what you sell. Not what your card processor charges, because some of these bundle that in and some leave you to pick, and a total mixing the two compares a bundle against a piece of one. That is why no two store comparisons on the internet agree with each other.

The card processing half is a whole sheet of its own. Every rate on it was read the same day as these, and the arithmetic there runs on your own numbers in exactly the same way.

And the largest cost of running a store is not on either sheet, because nothing can compute it. It is your hours. Every card below says what that platform does to them, and the tables print your own figure beside the money rather than folding it in.

What it costs you to be paid

The arithmetic

Four numbers of yours, and the ranking changes.

What goes through the store in a month, what an average order is, how many things you list, and roughly how many hours a month it takes you. Two worked examples follow so you can see the shape — put your own four numbers in their place and the sum is identical.

A maker with a short catalog

Assumed, not measured: $1,200 a month through the store, an average order of $45, forty things listed, and about ten hours a month spent on it. These are my numbers for the example, not a claim about what your business does.

Where you sell	A month	A year	What that really is, as a share of sales
Square Online Square Free	$0	$0	0%
No online store at all Nothing to subscribe to	$0	$0	0%
Big Cartel Platinum	$15	$180	1.25%
Shopify Basic, billed annually	$29	$348	2.42%
Ecwid Venture, up to 100 products	$29	$348	2.42%
BigCommerce Core, billed annually	$39	$468	3.25%
Etsy No subscription required	$78	$936	6.5%
WooCommerce The software itself is free	$25 to $350	$300 to $4,200	The software is free and the cost moved somewhere else. WooCommerce’s own pricing page puts hosting at $25 to $350 a month for most stores, and that is their range rather than an estimate of mine. Where in it you land depends on a host you have not picked yet, so this row gives both ends instead of averaging them into a number nobody published.
TikTok Shop	TikTok Shop charges a referral fee on every order and its own fee page sits behind a seller login that refused to open. The figure people quote comes from summaries of pages nobody outside a seller account can read, so it is not printed here. Ask somebody who sells on it, or open a seller account and read it yourself — and treat any number you find elsewhere as unchecked.
Facebook and Instagram Shops	There is no store fee to work out here any more, and the reason is the most useful fact on this sheet. Since September 2025, Shops on Facebook and Instagram send the buyer to the seller’s own website to pay. Meta no longer takes the payment, so there is nothing for Meta to charge — and a shop that has not been updated needs a checkout URL, which means it needs a website.
A page on your own site that reads your store	I take no percentage of anything sold through it, so this adds nothing to what the platform above already charges you. There is no figure to print here and inventing one would be inventing a fee. The build itself is quoted once, in writing, before it starts.

Cheapest to dearest, with the ones that cannot be worked out kept at the bottom rather than dropped. The gap between the top and the bottom of what could be priced is $936 a year on those numbers — and every figure above is what the platform takes, before your card processor takes anything.

And 10 hours a month of your own time, which is 120 hours a year.

These hours are not turned into money anywhere on this sheet. Doing that needs a figure for what an hour of your own time is worth, and inventing one for you would be the only made-up number on the page.

What is worth doing instead is reading the hours row on each card below. The difference between the easiest and the hardest of these to run is larger than the difference between the cheapest and the dearest, and it is the one nobody puts on a pricing page.

A shop doing real volume

Assumed, not measured: $12,000 a month through the store, an average order of $85, three hundred things listed, and about twenty-five hours a month spent on it. These are my numbers for the example, not a claim about what your business does.

Where you sell	A month	A year	What that really is, as a share of sales
Square Online Square Free	$0	$0	0%
No online store at all Nothing to subscribe to	$0	$0	0%
Big Cartel Platinum	$15	$180	0.13%
Shopify Basic, billed annually	$29	$348	0.24%
BigCommerce Core, billed annually	$39	$468	0.33%
Ecwid Business, up to 2,500 products	$49	$588	0.41%
Etsy No subscription required	$780	$9,360	6.5%
WooCommerce The software itself is free	$25 to $350	$300 to $4,200	The software is free and the cost moved somewhere else. WooCommerce’s own pricing page puts hosting at $25 to $350 a month for most stores, and that is their range rather than an estimate of mine. Where in it you land depends on a host you have not picked yet, so this row gives both ends instead of averaging them into a number nobody published.
TikTok Shop	TikTok Shop charges a referral fee on every order and its own fee page sits behind a seller login that refused to open. The figure people quote comes from summaries of pages nobody outside a seller account can read, so it is not printed here. Ask somebody who sells on it, or open a seller account and read it yourself — and treat any number you find elsewhere as unchecked.
Facebook and Instagram Shops	There is no store fee to work out here any more, and the reason is the most useful fact on this sheet. Since September 2025, Shops on Facebook and Instagram send the buyer to the seller’s own website to pay. Meta no longer takes the payment, so there is nothing for Meta to charge — and a shop that has not been updated needs a checkout URL, which means it needs a website.
A page on your own site that reads your store	I take no percentage of anything sold through it, so this adds nothing to what the platform above already charges you. There is no figure to print here and inventing one would be inventing a fee. The build itself is quoted once, in writing, before it starts.

Cheapest to dearest, with the ones that cannot be worked out kept at the bottom rather than dropped. The gap between the top and the bottom of what could be priced is $9,360 a year on those numbers — and every figure above is what the platform takes, before your card processor takes anything.

And 25 hours a month of your own time, which is 300 hours a year.

What the sum assumes, stated plainly: today's advertised prices, the plan each platform's own rules put you on, and no card processing of any kind. Where a figure leaves something real out, the card for it says what and why. It is arithmetic on published figures rather than a forecast, and the moment it needs a forecast it stops belonging here.

One at a time

The same 6 questions, asked of every one of them.

Grouped by who owns the customer when the sale is done, because that is the thing you cannot change your mind about later. Mine is on the list and gets the same treatment as the rest.

Your own store, your own address

Shopify

The one most people mean when they say "an online store"

$29 a month on Basic, and no cut of what you sell

Priced on the Basic plan billed annually. The dearer plans lower the card rate rather than the monthly fee.

What that figure leaves out: Apps. A Shopify store that does anything unusual ends up paying for three or four of them, and nothing can work out which ones you will want before you have run it for a season.

What the platform takes, before your card processor: Twenty-nine dollars a month, and Shopify takes no percentage of your sales on top of that — as long as you use their own payment processing. Use anybody else and they add 2% to every order, which is the single largest number on this sheet and the one people find out about after they have moved.
Who owns the customer when the sale is done: You do. The customer bought from your store at your address, you have the email, and nothing stops you writing to them next month. That is the difference between this and a marketplace and it is worth more than the rate.
What you take with you if you leave: Products, customers and orders export cleanly, which is better than most of this sheet manages. The storefront itself is written in Shopify’s own template language and runs nowhere else, so the look is what you rebuild.
What it costs you in hours: Less than anything else here. It updates itself, nothing breaks on a Tuesday, and the hours you spend go on photographs and descriptions rather than on keeping software alive. For most people that is what the monthly fee is actually buying.
What it does better than I do: Selling at any scale, and it is not close. Sales tax by jurisdiction and shipping worked out from a buyer’s address are the two things I will not build at any price, and a company whose whole business is keeping those correct maintains them. If you are selling more than a few things a week, you want this rather than me.
Who fixes it when it breaks: Shopify, around the clock, with the largest ecosystem of any name on this sheet. Whatever has gone wrong, somebody has already written about it.

Your own store, your own address

BigCommerce

The full platform that is not Shopify, and Shift4Shop beside it

$39 a month on Core, and the plan moves up on its own

Priced on the Core plan billed annually. Shift4Shop advertises an unlimited plan at $41 and sits beside this one; the figure in the table is BigCommerce’s.

What that figure leaves out: What happens when you have a good year. Plans upgrade themselves as your trailing twelve months of sales pass each threshold, so the price in the table is the price today rather than the price after a strong Christmas.

What the platform takes, before your card processor: Thirty-nine dollars a month on Core, rising to $105 on Growth and $399 on Scale, and no percentage of your sales below the Scale ceiling of $33,333 a month. Above that ceiling they take 0.9% of the excess. The part worth knowing is that you do not choose when to move up — the plan upgrades itself as your trailing twelve months pass each line, and that is not a fee so no comparison lists it.
Who owns the customer when the sale is done: You do, exactly as on Shopify. Your address, your customer list, your relationship.
What you take with you if you leave: Products, customers and orders export. The storefront is theirs. Same shape as Shopify, with a smaller ecosystem to rebuild into.
What it costs you in hours: About the same as Shopify — low, and mostly spent on your own photographs. The one thing to diary is the plan threshold, because the month it moves is the month the bill changes without anybody sending you an email first.
What it does better than I do: Giving you more of the store without an app for every piece of it. A good deal of what costs extra on Shopify is included here, which is the reason to look at it at all, and for a business with an unusual catalog that difference is real money.
Who fixes it when it breaks: BigCommerce, by phone and chat. Fewer people have used it than have used Shopify, so the answer to an odd question is more often theirs to give than a forum’s.

Your own store, your own address

WooCommerce

A store bolted onto WordPress, running on hosting you rent

Free software, and hosting somewhere between $25 and $350 a month

The range is WooCommerce’s own published figure for hosting, not a guess. Nothing else here is priced, because nothing else here has a price until you choose a host.

What that figure leaves out: Extensions, at $29 to $299 a year each, and whoever keeps the thing patched. A Woo store with six extensions is normal and costs more a year than the hosting does.

What the platform takes, before your card processor: Nothing, for the software, and a 0% revenue share — those are both WooCommerce’s own words. Then hosting at $25 to $350 a month, extensions at $29 to $299 a year each, and somebody to keep it all updated. The honest reading of this card is that the platform is free and the cost moved to places that do not appear on a pricing page.
Who owns the customer when the sale is done: You do, more completely than anywhere else on this sheet. It is your database on your hosting, and there is no company in the middle of the relationship at all.
What you take with you if you leave: Everything. The content, the theme, the orders and the database together. This is the one row where that answer is honestly everything, and it is a large part of why WordPress runs so much of the web.
What it costs you in hours: The most of anything here, and it is the reason people leave it. WooCommerce is plugins, plugins update, and an update can take a working store down on a Tuesday for a reason nobody chose. Budget an evening a month that you will not always need and will badly want when you do.
What it does better than I do: Freedom, and price at the bottom end. There is a plugin for nearly anything, a plugin costing seventy-nine dollars once beats me building the same thing for four figures, and nobody can switch your store off. If what you need already exists there, buy it and keep your money.
Who fixes it when it breaks: You, or whoever you pay. It is the option with the most freedom on this sheet and the most homework attached to it.

It came with something else

Square Online

Already sitting in the account of anybody taking cards through Square

Nothing a month, and all of it in the card rate instead

Priced on the free plan. The whole of Square’s take is in what they charge to process a card, which is the column this sheet deliberately leaves to /getting-paid/.

What that figure leaves out: The card rate, which on this one is the entire cost. Square Free charges 3.3% plus 30 cents on an online sale against 2.9% plus 30 cents on their paid plan, and on a store doing real volume that gap is the whole decision.

What the platform takes, before your card processor: Nothing as a monthly fee, and no percentage taken as a platform. Then read the payments sheet, because this is the card where the money is entirely in the processing rate: 3.3% plus 30 cents online on the free plan. Cheapest row in this table and not necessarily the cheapest store.
Who owns the customer when the sale is done: You do, and the customer list is the same one your card reader has been building. For a shop that already runs on Square that single fact is worth more than the rate difference.
What you take with you if you leave: Your item list and your customer records export. The site itself does not. Card-on-file records do not travel either, so anything on a subscription has to be re-entered by the customer.
What it costs you in hours: Very few, because there is not much of it. It is a storefront bolted to a card reader rather than a store you tend, and the thing you will spend hours on is working around what it cannot do.
What it does better than I do: Costing nothing and already existing. If you take cards through Square you have an online store right now whether you knew it or not, and you should go and look at it before you pay anybody for one — including me.
Who fixes it when it breaks: Square, and there is a telephone number a person answers during business hours. The reader, the store and the bank account are one company, so a problem is one conversation rather than three.

Your own store, your own address

Ecwid

A store you drop into a site you already have

$5 to $119 a month, priced by how many things you list

The tier in the table is the one your catalog size lands on. Annual billing saves 16% and is not applied here.

What that figure leaves out: Where you want to sell. Selling on Instagram and Facebook starts at Venture, and selling in person needs Unlimited, so the tier your catalog picks is not always the tier your plans pick.

What the platform takes, before your card processor: Five dollars a month up to ten products, twenty-nine up to a hundred, forty-nine up to twenty-five hundred, and a hundred and nineteen above that. No percentage of your sales at any tier. It is the only thing on this sheet priced by how much you LIST rather than by how much you SELL, which makes it the cheapest row here for a maker with a short catalog and one of the dearest for a shop with a long one.
Who owns the customer when the sale is done: You do. It sits inside a site that is already yours, so the customer never leaves your address to buy.
What you take with you if you leave: Your catalog and your orders export. What does not move is the store itself, and because it was embedded in your site rather than being your site, what you are left with is the site with a hole in it.
What it costs you in hours: Few, and the reason is that there is less of it than the full platforms. The hours arrive later, when you want something it does not do and find the answer is to move.
What it does better than I do: Being small and cheap without being a toy. For somebody selling a dozen things beside a website they already have, five dollars a month against twenty-nine is a real difference and nothing above is doing anything they need.
Who fixes it when it breaks: Ecwid, by chat and email, with phone support on the dearer tiers. It is a smaller company than the two above it and that shows in both directions.

Your own store, your own address

Big Cartel

Built for artists and independent makers, and it shows

$15 a month on Platinum, and a free plan underneath it

Priced on Platinum. A free Gold plan exists and is limited; exactly how it is limited was not read on 11 September 2026 and is not claimed here.

What that figure leaves out: What you give up for the price. Abandoned-cart recovery is a Diamond feature at $30, and the older reporting is thinner than anything above it on this sheet.

What the platform takes, before your card processor: Fifteen dollars a month on Platinum and thirty on Diamond, with a free plan under both, and no percentage of your sales on any of them. It is the cheapest real store on this sheet and it is aimed squarely at somebody selling their own work.
Who owns the customer when the sale is done: You do, and the store is at your own address. What is thinner here than above is what you can DO with the list once you have it.
What you take with you if you leave: Your products and your orders. Like the rest of this group, the storefront stays behind.
What it costs you in hours: Almost none, because there is almost nothing to configure. That is the product rather than a shortcoming — it does one thing and it stops.
What it does better than I do: Knowing exactly who it is for. An artist selling twenty pieces does not need sales tax by jurisdiction or abandoned-cart email, and a platform that refuses to grow features it does not need is rarer and more valuable than it sounds.
Who fixes it when it breaks: Big Cartel, by email, and the company is small enough that the answer usually comes from somebody who knows the product.

Somebody else’s crowd

Etsy

A marketplace that brings the buyers with it

6.5% of every sale, no monthly fee, and the 6.5% lands on your shipping too

Priced on the 6.5% transaction fee alone, against the order value you gave. The real figure is higher, and the line below says by how much.

What that figure leaves out: Two things, and both are real money. Twenty cents to list each item, charged whether or not it sells and again every time the listing renews. And the 6.5% is charged on the shipping and gift wrapping you billed, not only on the price of the item — so the figure in the table is a floor rather than the answer.

What the platform takes, before your card processor: No subscription, then 6.5% of the price you displayed plus what you charged for shipping and gift wrap, plus twenty cents for every listing whether it sells or not and again on every renewal, plus payment processing on top that varies with where your bank is. Etsy Plus is another ten dollars a month and is optional. Add it up and this is the dearest per-sale option here — which is the fair trade for a marketplace that brings its own buyers, and the card should say both halves.
Who owns the customer when the sale is done: Etsy does, and this is the row that matters more than the fee. The buyer thinks she bought from Etsy, the email address is not yours to write to, and the search results that found you can stop finding you without anybody explaining why. You are renting an audience, and the rent is the 6.5%.
What you take with you if you leave: Your listings, as data, and very little else. The reviews stay, the search position stays, and the customers stay — they were never yours. A seller moving off Etsy starts the audience again from nothing, which is why so few do.
What it costs you in hours: Real hours, and they go somewhere unusual: into the marketplace itself. Relisting, tags, the ads system, and watching a search position you do not control. That work does not transfer anywhere if you leave.
What it does better than I do: Bringing buyers, and nothing on this sheet comes close. A store on your own address gets exactly the visitors you drive to it; a listing on Etsy gets people who came to Etsy to buy. For a maker with no audience yet that is the whole game, and the fee is what it costs.
Who fixes it when it breaks: Etsy, by a form. You are one of millions of sellers, which is the same fact that makes the buyers arrive.

Somebody else’s crowd

TikTok Shop

Selling inside the app people are already scrolling

A referral fee on every order, and the number is not publicly readable

Nothing is assumed, because nothing is computed. The fee page returned a permission error on 11 September 2026.

What that figure leaves out: Everything, because nothing was priced. That is the card rather than a gap in it.

What the platform takes, before your card processor: A referral fee on each order and no setup fee, and I am not going to print the percentage because I could not read it. Their fee page requires a seller login and returns a permission error without one. Every published figure for this traces back to somebody summarizing a page they also could not open, and a guessed number on a sheet whose whole claim is that every figure was read would be worse than no sheet.
Who owns the customer when the sale is done: TikTok does, completely. The purchase happens inside the app, the buyer never reaches an address of yours, and the audience belongs to a recommendation system rather than to you.
What you take with you if you leave: Nothing that matters. There is no customer list to take and no store to move. What you built was attention, and attention does not export.
What it costs you in hours: More than any other row here, and the hours are not shop hours. Selling on TikTok is making video, continuously, and the day you stop the sales stop with it. That is a job rather than a channel.
What it does better than I do: Reaching people who were not looking for you. Nothing else on this sheet can put your product in front of somebody who has never heard of you and has no intention of shopping. For the right product that is genuinely transformative, and it is why people put up with the rest of it.
Who fixes it when it breaks: TikTok, through the seller center, and getting a human is difficult. Plan for that before you make it the only place you sell.

Somebody else’s crowd

Facebook and Instagram Shops

What most small businesses already have, and it changed in 2025

No fee, because Meta no longer takes the payment at all

Nothing is assumed. There is genuinely nothing to price, and the reason is dated September 2025.

What that figure leaves out: The cost of the website the buyer is now sent to, which is what this change moved the money to. That is what the other sheet here is about.

What the platform takes, before your card processor: Nothing to Meta. As of September 2025 your product tags direct the customer to your own website to complete the purchase, and managing the post-purchase side inside Commerce Manager has been discontinued. Most comparisons you will find still describe Instagram Checkout as a live feature. It is not, and that is worth knowing before you plan around it.
Who owns the customer when the sale is done: You do now, which is the change. The buyer finishes on your site, so the email address, the order and the relationship are yours — and that is the opposite of how this worked two years ago.
What you take with you if you leave: Your posts and your followers stay with Meta. What is different since the change is that the CUSTOMERS do not, because they bought on your own site. That is the first time that sentence has been true here.
What it costs you in hours: The hours go into posting rather than into a store, and the store half is now somebody else’s problem — yours, on your own site. Which is the point.
What it does better than I do: Being where your customers already are, for free, with an audience you may have spent years building. Nothing here argues against having it. What it can no longer do is take the money, so it needs somewhere to send people.
Who fixes it when it breaks: Meta, by a help center and very little else. The half that matters is now on your own site, where you can actually get somebody to fix it.

What plenty of good businesses do

No online store at all

A page, a photograph, a price and a phone number

Nothing, and for a lot of businesses it is the right answer

Nothing is assumed. There is no subscription and no percentage, so the figure is zero and it is exact.

What that figure leaves out: The sales you do not make because somebody wanted to buy at eleven at night. That is a real cost and nothing on this page can measure it for you.

What the platform takes, before your card processor: Nothing at all. A page showing what you make, what it costs and how to ask for it, with a payment link for a deposit if you want one. No basket, no stock counts, no checkout. For a business selling ten commissioned things a month at four figures each, a basket is machinery nobody needed.
Who owns the customer when the sale is done: You do, entirely, and you probably spoke to them. This is the only row where the customer relationship starts as a conversation.
What you take with you if you leave: Everything, because there is nothing to leave. The record is your own bookkeeping and the relationship is with the customer.
What it costs you in hours: The hours are spent on the customers rather than on the store, which is the trade. What it costs you is answering the same question forty times a week, and a good page is what stops that.
What it does better than I do: Cost, and simplicity, and it is not close. It is free, there is nothing to break, and for a maker taking commissions rather than shipping stock it may genuinely be the whole answer. Ask what a basket would actually do for you before you rent one.
Who fixes it when it breaks: Nobody, because nothing breaks. What it quietly does is less than you think, and the way to find out is to count how many people asked to buy and did not.

Mine

A page on your own site that reads your store

Built once, pointed at whichever platform above you are on

Their fee, with nothing added by me, and the build quoted once

Nothing is assumed about cost, because there is no recurring cost from me. What it can be built on at all depends entirely on which platform you are on — the table further down answers that.

What that figure leaves out: Your platform’s own fee, which does not go away. This sits on top of a store you already have rather than replacing one.

What the platform takes, before your card processor: Nothing, month to month. I take no percentage, no markup and no share of anything you sell, and that is a term in the agreement rather than a promise. The build is quoted once before it starts. A real store — a basket, a checkout and stock counts — is separate work with its own price, and I do not build shipping by address or sales tax by state at any price.
Who owns the customer when the sale is done: You do, and more of them than before, because the customer stays on your address for the part where they decide what they want. The checkout still happens on your platform, where it belongs.
What you take with you if you leave: All of it, because none of it was ever mine. The files, the hosting account and every login are in your name, and if you stop working with me tomorrow nothing switches off.
What it costs you in hours: The point of the thing is that it costs you none. You add a listing in the store you already run and the page follows, because it reads your catalog live rather than holding a copy I have to update. That is the one requirement that decides which version you get.
What it does better than I do: Nothing at all, as a store — I am not a store platform and I do not compete with one. What this does instead is the part no platform does: the page where your customer configures the thing they actually want, in your words, before they land in your checkout with it.
Who fixes it when it breaks: Me for the page, the platform for the store and the money. Where the platform changes something underneath it, keeping up is quoted in the same breath as the build rather than assumed.

The one thing I sell here

Whether a page I build can read your catalog depends entirely on which platform you are on.

The thing worth building here is a page on your own site where a customer configures what they actually want — a size, a finish, a piece of text, whatever your work varies by — and then lands in your store's own checkout with it. You add a listing in your store and the page follows, because it reads your catalog live rather than holding a copy I have to update.

4 of the 12 platforms below let an outside page do that with a credential that is safe in a browser. 7 do not, and the reason is almost always the same: they require a secret on every request, which would mean me running a server holding a key that can read your orders. That is your customers' data and I do not hold it.

This is not a ranking and a "no" is not a mark against a platform. It is a fact about what I can build, not about how good they are at selling. Several of the ones that refuse are the best products on this page.

Platform	Can a page I build read it	What it needs	Why
Shopify	Yes	A public storefront token, which is not a secret	Their Storefront API lets an outside page read products, build a cart and hand the buyer to Shopify’s own checkout. It also exposes the price, whether an item is available and how many are left, so the page never holds a stale copy of your catalog.
WooCommerce	Yes	None at all	The Store API is explicitly unauthenticated and needs no key. The honest caveat is that its sessions are cookie-based and writes need a one-time token, which is straightforward on the same site and unproven from another — so nobody should promise this on Woo until it has been built against a real store.
BigCommerce	Yes	A storefront token BigCommerce itself calls safe in a browser	Locked to the addresses you name, up to two per token, so a staging site and a live site use both. There is a deprecation clock on tokens created without an address: from March 2027 they stop being accepted, so anything built here names the address from day one.
Wix	Yes	A client ID, with no secret	Wix Headless authenticates a visitor with an ID alone. A secret is required for anything administrative, so the safe half is visitor and member access — which is all a configurator needs. The cart shape was seen but not read endpoint by endpoint, so confirm it before building rather than assuming it matches Shopify’s.
Ecwid	With a caveat	A public token, for their widget rather than your page	Ecwid’s public tokens are safe in a storefront, and the model is that you embed THEIR storefront into your page and customize it — not that your page reads their catalog and builds its own. That is a different product from the Shopify pattern and it should not be sold as the same thing.
Square	No	A secret access token, server-side only	Square instructs outright that access tokens must not be exposed in client-side code. Building this would mean running a server holding a token that can read your orders, which is customer data I do not hold. Whether Square Online exposes an unauthenticated add-to-cart address is unsettled; community posts suggest one and Square does not document it.
Squarespace	No	There is no cart to hand anything to	A harder no than Square’s, and for a different reason. Square has a cart and forbids reaching it from a browser. Squarespace’s commerce interfaces manage a merchant’s own data and include no cart or checkout at all, so even with a server there is nothing to hand a configured item to.
Etsy	No	A shared secret travels on every request	That rules a browser out before any question about carts. Etsy is a marketplace and the buyer buys on Etsy — the cart interfaces that exist read a shopper’s own Etsy cart after she authorizes an app, which is not a checkout an outside site can send anybody to.
Big Cartel	No	A secret, and an application they have to approve	Three separate blocks and any one is enough. It needs a secret; its cart interface reports abandoned carts from the last seven days rather than building one, and only on the dearest plan; and Big Cartel approves a select number of developers at all.
Shift4Shop	No	A private key on every request	Their public key is not a browser credential — it is what a merchant uses to subscribe to an application. The private key is what authenticates the request, so this is server-side only.
TikTok Shop	No	The checkout happens inside the app	Purchases complete inside TikTok and there is no documented way for an outside page to build a cart against it. That rests on the absence of public developer documentation rather than on a statement by TikTok, and the sheet says so rather than dressing it up.
Facebook and Instagram Shops	No	There is no longer a checkout here to build against	Since September 2025 the buyer is sent to the seller’s own website to pay. So the question does not arise — and the page they are sent to is exactly the kind of thing this practice builds.

There is a second version of this, and it is worse for most people.

A configurator can also hold its own copy of your catalog instead of reading your store's. That version works on any platform, including all the ones above that say no — and it means every new thing you list, I have to add. If you want to add your own listings and have the page keep up, it has to read your store, and that is what the table answers.

Before you ask me for anything

What I build, what I will not, and what has to be quoted first.

The same three lists sit on every sheet I publish, because the answer does not change with the subject. The middle one is the one worth your time.

What I can build for you

All of it ends up in accounts opened in your name, and none of it needs me afterwards.

A website that is yours outright: The files, the domain, the hosting account and every login open in your name. There is no leaving, because you were never standing on anything of mine.
A tool built for the way your business actually works: Your rate card, your roster, your categories, your trade’s words. It arrives twice — as a page at your own web address, and as one file that opens with the internet switched off and needs no login and no subscription.
One payment link, connected to one page: Tested with a real payment before handover and refunded afterwards, on an account that is yours throughout. This is included in a build under §1.7 of the Scope, and I am never an owner on the account.
A form that reaches your inbox: Names, emails and phone numbers are ordinary work and always were. What a form collects and how long it is kept is written into the privacy page the build includes.
The arithmetic on your own numbers, printed where you can check it: No model, no forecast, no industry average and no suggested rate. Your figures through a sum you can follow, and a plain sentence wherever the sum cannot be done.
A read of what your Google listing shows the public: Checked against what your site says, so the two agree. This needs no access to anything of yours — it is the same page any customer sees.
A page that reads your own catalog and hands a configured item to your cart: Your customer builds the thing on your site and lands in your platform’s own checkout with it. You add a listing in your store and the page follows, with nothing new for you to learn.

What I will not build, and why

Two different reasons, kept apart on purpose. Some of these would make a promise you already hold in writing false. The rest I could build and will not, and the reason is the same each time — somebody else has to keep it correct for years and does.

Anything that sees, stores or transmits a card or bank number: The master agreement says outright that I do not see, store, transmit or process payment card data. A payment link hands your customer to your provider’s own page, on your provider’s own address, and nothing about the way I build changes that.
A seat on your Google Business Profile: Both Scopes say that nothing in them asks you to grant me any access to it, and Google’s own rules for third parties require the business to hold its profile at all times. The listing is yours to create, to verify and to keep.
A system that holds records carrying their own legal regime: Patient notes, medical history and the like. Section 3 of the Scope says I do not build systems that hold them. A practice holding such records can still have a website — the site simply does not touch that material.
Legal, tax, accounting, employment or financial advice: Named in section 3 of the Scope as outside every build. I will tell you who the merchant of record is, because that decides who owes the sales tax. Whether you owe any is a question for your accountant and the answer is worth paying for.
A way into your computer when you are not sitting at it: The data-handling note in your pack says I have no way to do it and do not want one. Every remote session needs you present and letting me in, and it ends when you close it.
Sales tax worked out by state: I could build this and I am not going to. Rates and rules change without notice and it has to stay correct for years after I have been paid. A company whose entire business is that maintains it, and I will name the one that does rather than sell you a version of it that rots.
Shipping worked out from a customer’s address: Same answer and the same reason. Carrier rates, zones and surcharges move on somebody else’s schedule, and a shipping calculator that is six months out of date overcharges your customers quietly.
Online booking, invoicing and payment chasing, review requests, staff scheduling, waiting lists: Every one of these is a category a platform you already pay for has taken, or will inside a year. Where you need one, I will say which product does it. That answer is worth more to you than a worse version of it with my name on.
A configurator on a platform that will not let a browser read the catalog: Some platforms require a secret key on every request, which would mean me running a server holding a token that can read your orders. That is customer data, and both the Scope and the data-handling note say I do not hold it. On those platforms the honest answer is that this cannot be built the clean way.

What has to be quoted before it starts

Real work, quoted in writing before anything starts, and never folded into a build price to win one.

An online store — a basket, a checkout and stock counts: A real store is a different build at a different price, quoted in writing before it starts. One payment link on one page is a much smaller thing and comes with a build.
Opening a payment account from scratch: This runs on your identity documents, your bank account and a verification process the provider controls, which neither of us can hurry. Where you already hold an account, connecting it is included.
Setting up or running your Google Business Profile: Separate work with its own price, and still in your name when it is done. Reading what the listing already shows is included; holding it is not.
Any account your own customers log in to: Logins are the largest single step in cost of anything on this page. A site that has them is a different build, and pretending otherwise at quoting time is how a build goes wrong in month two.
Connecting to another system you already run: Possible, often useful, and never assumed. It gets named in the Statement of Work with its own price, because the cost lives in the other system rather than in mine.
A second tool, a further phase, or anything else named in section 3: Quoted as its own fixed price. Small work outside a Scope runs at the hourly rate on the rate card, with thirty minutes the smallest unit invoiced.
The configurator itself, on a platform where it is possible: It runs on a login belonging to your store rather than on your own machine, so it is never one of the free fitted tools. It is quoted work, and the quote covers keeping it working when your platform changes.

What this is not

Not a recommendation, and not a claim that any of these are bad.

They are well built and they are run by companies with more engineers than this town has people. Nothing here argues otherwise. What the sheet does is put the fee beside the question nobody asks, which is what happens to your customer list if you ever want to leave.

It is also not complete, and two gaps are worth naming rather than hiding. TikTok Shop's referral fee sits behind a seller login that would not open, so it is not printed. And the commerce plan prices for Wix and Squarespace were not read on this pass — their website plans were, on the other sheet, and the plan that sells is a different one.

It is not tax advice either, and the line matters. Who owes sales tax on a sale depends on who the legal seller is, which differs between a marketplace and your own store. What nobody here will ever do is work out whether you owe any. That is a question for your accountant.

And it is not a list I keep current by remembering to. Every figure above is registered with the date it was read, it expires on its own, and this page carries that date in plain sight. One line on this site went stale exactly that way and the register is what caught it.

If you want the sum done on your own numbers

Send me the four numbers and I will send the table back.

A month of store sales, an average order, how many things you list, and roughly how many hours a month it takes you. No charge and no meeting. Tell me which platform you are on and the catalog answer comes back with it.

Send me the four numbers The same reading, for card processing And for websites

Every fee and every term on this page was read off those companies' own pages on 11 September 2026. They change them without announcing it, so if you are weighing this up, go and read them again yourself.