Three numbers to run before you rent a checkout.

One. The fee, as a percentage

$29 a month is a flat number, which is why it reads as harmless. It is flat against the calendar and nothing else. Divide it by what you actually took last month.

Sales in the month	Flat fee	What the fee really is
$200	$29	14.5%
$400	$29	7.25%
$1,000	$29	2.9%
$5,000	$29	0.58%
$0	$29	still $29

That is a platform rate sitting on top of whatever your processor already takes, and it is charged in full during the months you sell nothing. The rate falls as you grow. It falls to something reasonable at exactly the point where a monthly bill stopped mattering to you.

So the fee is heaviest in the year you can least carry it and lightest in the year you would not notice it. Find your own row in that table before you read anybody's plan comparison.

Two. Thirty-six months

Nobody changes checkout every season. Whatever you sign up for this week is likely still taking your orders in three years, so price it over three years rather than one month.

$29 a month is $1,044 over 36 months. There is no discount for loyalty in that sum and no interest working in your favour. It is simple multiplication, and it is the honest sticker price of the decision you are making today.

Then ask the second half of the question, which is what you are holding on the last day of month 36. An account. An account is a permission, and permissions are revocable. The pages, the theme, the discount logic and the subscription plumbing all stay on somebody else's servers, licensed to you for exactly as long as the payment clears.

Thirty-six payments buy thirty-six months. At the end of them you own thirty-six receipts.

Three. The price of the door

The exit is the number nobody prints, because nobody bills it. It is paid in work, and it comes due at the worst possible time, which is the moment you have already decided to go.

The customer list. Export it today, before you need it, and look at what actually comes out of the file. Email addresses and names are the easy part. Order history, refund state and subscription dates are where exports get thin.

The saved payment methods. This deserves an accurate answer rather than a scary one, and the only accurate answer is the one your processor gives you in writing. Ask what happens to cards on file when a shop moves, and ask while you are still a customer in good standing. Whatever comes back, it will not rebuild your subscription state, reconcile billing dates or repoint your webhooks. That is real integration work on both ends, and every subscriber whose renewal fails while you do it is a cancellation you have paid for twice.

The links you already sent. Every checkout URL you printed on a card, read out in a video, pasted into an email sequence or handed to an affiliate is a promise about a domain. If those links live on your own domain, moving is a deploy. If they live on theirs, moving means either paying the old bill purely to keep redirects alive or accepting that some of your best links quietly stop working.

None of this is an argument against renting. Work the figure out before you owe it.

When renting is the right answer

Renting is the correct decision far more often than a page on a site like this one usually admits. Three cases where it plainly is.

You want somebody to call. A support desk is a real product with real salaries behind it. If a broken checkout at two in the morning is a business emergency for you, $29 buys you a number to ring, and that is cheap.

You do not want to hold a database. A customer list on your own Postgres is genuinely yours. So are the backups, the upgrades, the retention policy and the breach. If holding that makes you uneasy, the unease is telling you something true.

You would rather the whole problem belonged to someone else. That is a legitimate thing to buy. Paying $1,044 over three years to never think about webhook signatures again may be the best value in your entire stack.

What owning costs

Self-hosting has no licence fee. It is not free.

Run Charming Seal and you are the operator. You keep Node and the dependencies current. You watch the webhooks. You hold the database and its backups. You read the security policy, and you are the one awake when something breaks. There is no desk behind you, because the absence of that desk is precisely why there is no bill. That is the trade, stated plainly: you keep the fee and the data, and you take the shift.

The three sums will not tell you which side of that trade to stand on. They only tell you the price of the side you pick. The trouble is that most people sign up having run none of them, and find the third number out years later, on the day they try to leave.