Stripe or PayPal for a Small Business & Why Payments Get Held

Every few weeks somebody asks us the same two questions in the same breath. We're going to start taking card payments — is it Stripe or PayPal? And, usually a line later: I've read about payments being held, should I worry?

Both have better answers than the internet gives them.

They aren't quite the same kind of thing

Stripe is a processor. It charges cards and pays out to your bank every couple of days, so the money's resting place is your bank account.

PayPal is a processor and a wallet. Money lands in a balance that you then move, and it carries something Stripe doesn't: buyer familiarity. A meaningful number of people will finish a purchase with a PayPal button who would have abandoned a card form.

For most small businesses, use Stripe as the processor of record — website, virtual terminal and invoices all through one account — and add PayPal as a checkout button if you build a checkout. What you don't want is two full processors running in parallel, each with its own fees, dashboard, disputes queue and reconciliation. That's double the admin for one stream of revenue.

Note that the fee difference between them is rarely what decides it. The admin is.

Why payments get held

Worth answering properly, because most of what's written about this online is outrage without solutions.

Every payment company is carrying your risk. When a customer disputes a charge, the processor refunds it whether or not your money is still there. So they hold funds when their model can't yet predict you. Three situations trigger it.

You're new. PayPal holds a new seller's early payments for up to 21 days while it builds a picture of your business. It isn't personal and it goes away with history.

Your volume changes shape. This is the one that catches established businesses. You normally take $2,000 a month in small orders, then a single $14,000 order lands, and every fraud model in the industry flags it.

Your dispute rate climbs, at which point you may get a reserve: a percentage of each day's takings held back and released on a schedule.

Two things genuinely help. Add tracking numbers, because for physical goods PayPal usually releases held funds within a day of delivery confirmation. And fill in your business profile before you need it — what you sell, your average order, the largest order you expect this year, how long you take to ship. Underwriting is far more forgiving of a big order it was warned about than one that appears out of nowhere.

Holds also feel worse on PayPal because you sit and watch the balance, where Stripe just shows you a payout schedule. Same risk management, different psychology. If that anxiety is a real cost to you, that's a real reason to make Stripe primary.

Taking cards over the phone

If what you actually need is to key in a card while a customer is on the line, you can do that today with no website changes at all. Stripe has a virtual terminal in its own dashboard: log in, create a payment, type in the card. No plugin, no cart, no developer.

Keyed cards cost more. Stripe's US standard pricing is ~2.9% + 30¢ online and 3.4% + 30¢ for a card typed into the dashboard — the price of a card no device ever saw. You also carry the chargeback risk, because there's no chip, no signature and no 3-D Secure. If the cardholder later says it wasn't them, you will usually lose.

For regulars on the phone, none of that matters. For a first-time caller placing an unusually large order in a hurry, treat the urgency itself as the warning and send an invoice with a payment link instead. This is cheaper, better documented, and the customer authenticates themselves.

Before you build a checkout

One thing worth checking first: whether you need a store at all. Taking cards by phone, pointing people at your distributors, connecting an Amazon channel and sending invoices all get bundled into "let's add e-commerce," and only one of those four jobs actually needs a shopping cart. We've written up which is which, and the order to do them in.

Where we come in

If you do end up with a store, the failure that costs the most is the one that doesn't look like a failure: the homepage loads, the phone doesn't ring, and the orders quietly stop because a gateway key expired. On the stores we host we read the list of payment methods your own store is offering, every five minutes, and an empty list is an alert — whichever processor you settled on. Here's what we check, and what it does and doesn't prove.

And if you're still weighing it up, that conversation is free. Tell us what you sell and we'll tell you which part is worth doing first.