PayPal sandbox checkout: test before going live

A sandbox payment tells you the store can reach PayPal's test site and that the order gets marked paid. It does not tell you what a stranger does before the PayPal window opens: which button they press, whether the cart total is still right when PayPal opens, and whether the return to the store lands on a page that makes sense. Those steps are where some buyers turn back. An owner who runs sandbox payments only from their own browser gets a green result and an unknown path.

• Confirm the store's PayPal setting says sandbox before a sandbox payment, and live before a real one.
• Check the PayPal window address matches that mode, with the test address for sandbox.
• Open the button from a phone and a laptop, and compare the total on each.
• Note whether the return to the store shows the order as paid, and what the cart shows.
• Confirm a sandbox order in the store admin before you trust the result.

The sandbox checks the plumbing. It confirms the store's credentials work against PayPal's test site, that the redirect goes out and comes back, and that the order is marked as paid when the payment completes. Those are real checks, and they catch setup faults worth finding early. What they do not check is the path before the window opens, because a sandbox run starts with an owner who already knows where the button is.

Keep sandbox and live credentials apart. A sandbox buyer account exists only on the test site and holds no real money. Pick one test buyer, use it for every sandbox run, and keep its details out of the live store settings. That keeps a test from touching real money by accident. Write the test buyer's name on the same page as the sandbox settings, so the next person who runs a test does not guess which account to use.

Most of the stranger's path happens before PayPal. A PayPal button on the cart page starts a payment earlier than one at checkout, so it carries fewer details into the PayPal window. Each start carries a different set of numbers. A shopper who sees one total on the cart and another in PayPal reads it as an error, even when both come from the store.

Compare the two starts line by line. Add an item, apply a discount code and pick a variant, then open PayPal from the cart and again from checkout. Note the total, the items and the shipping line in each window. Any difference is the thing to fix first, because it is what a shopper will check before deciding. Write each difference down with the window it came from, so the fix can be checked against the same two starts.

A simulated visitor follows the same path and records each screen. It reaches the PayPal button, opens the window and stops on PayPal's page, because a visitor has no PayPal account to finish a payment with. The result is a set of screens showing what a stranger could see and read before that point, and where each one left. That shows whether a stranger reaches PayPal at all, which a sandbox run cannot show.

Changing credentials, updating the PayPal plugin or moving the store can leave the mode out of step with the keys. Check the window address and the mode again after each of those changes, then rerun the simulated visitors to the button. Keep a list of the sandbox and live orders you have placed, so a live order never looks like a test one.