What to test in your checkout before every release
Shaig Khaligli, founder of Testavi
· 3 min read
Your checkout is the one part of the site where a bug shows up directly in revenue. Before every release, test the cart, discounts, address and shipping, payment, the final total and the confirmation, ideally as one continuous flow the way a real shopper goes through it.
That sounds obvious, but in most teams the checkout gets tested by whoever has time on release day, clicking through once on their own laptop with their own saved address. This checklist is what we'd rather see instead.
Key takeaways
- Baymard Institute puts the average documented cart abandonment rate at 70.22%.
- Of shoppers who abandoned for reasons other than just browsing, 17% said the site had errors or crashed.
- Test checkout as one end-to-end flow, not as separate pages, because most bugs live in the handoffs.
- Run it per market if you sell in more than one country.
Why checkout deserves its own test plan
Cart abandonment is when a shopper adds items to their cart and leaves without completing the order.
Baymard Institute's running average of documented cart abandonment studies is 70.22% (Baymard Institute, updated September 2025). Plenty of that is people comparing prices or saving items for later, and you can't fix it.
Some of it you can. When Baymard excludes shoppers who were "just browsing", 17% of the remaining abandoners said they left because the website had errors or crashed. That's the part testing is for. Nobody files a bug report for a checkout that didn't work. They just leave.
The checklist
Cart
- Adding, removing and changing quantities updates the line price and the subtotal.
- Product variants stay correct: size 42 in the product page is still size 42 in the cart.
- The cart survives a page refresh and, if you support it, signing in halfway through.
Discounts and promotions
- A valid code applies once and changes the total by exactly the right amount.
- Invalid or expired codes show an error a shopper can understand.
- Codes that shouldn't stack, don't. Free-shipping thresholds are checked after the discount, if that's your rule.
Address and shipping
- Address forms accept real formats from every country you ship to. Postcodes are where this usually breaks.
- Phone validation accepts international numbers.
- Shipping options and prices match the destination.
Payment
- A test card goes through and produces an order.
- A declined card shows a clear error and lets the shopper try again without losing the cart.
- Any 3-D Secure or redirect step brings the shopper back to the right place.
The total and the confirmation
- The final total equals items, minus discounts, plus shipping and tax, in the right currency.
- The confirmation page shows a real order number, and the confirmation email arrives.
Things that get in the way
- Cookie banners, newsletter pop-ups and promo modals don't block the flow.
- The flow works on a phone-sized screen, not just your desktop.
Test the handoffs, not just the pages
Most checkout bugs don't live on a single page. They live in the handoffs: the discount applied on the cart page that isn't carried into payment, the shipping cost that changes after the address step, the currency that switches back to default after a redirect.
That's why testing each page in isolation misses things. One end-to-end run, from product page to confirmation, catches the bugs your customers actually hit. If you sell in several countries, run that same flow once per market. Our page on localization testing goes into what to check for each locale.
Make the test survive your next redesign
Checkout pages change a lot: new payment methods, A/B tests on the button, a redesigned order summary. A test that depends on exact selectors will break with each of those, and after the third time someone switches it off. We wrote about why end-to-end tests break after redesigns and what to do about it.
With Testavi, the checklist above becomes plain-English steps, for example "apply discount code WELCOME10" and "verify the total equals item price minus 10% plus shipping, in euros", and an AI agent runs it like a shopper would. See checkout flow testing for a full example, or start a free 30-day pilot and we'll write your checkout tests for you.