A store expands into new countries by activating markets, translating the storefront, and announcing international shipping. Then the checkout data from those markets disappoints, and the layout gets blamed. In practice the completion gap on cross-border orders is usually set upstream: in how duties are presented, whether the currency survives to payment, and where the translation stops.
None of these are visible in a domestic test order, which is why they persist. The store tests checkout in its home market, in its home language, with its home tax logic, and ships a configuration that behaves differently for exactly the shoppers the expansion was meant to win.
Duties: decided in Markets, felt at payment
Whether duties and taxes are included in the price or calculated at checkout is a Markets level decision that lands on the shopper at the worst moment. A product page price that grows by a duty line at checkout reads as a price increase, and cross-border shoppers abandon over it at rates domestic shoppers never approach.
The alternative, duties included, raises the displayed price and can dent click through, but it converts consistently because the number never changes after the product page. Either model can work. What fails is the mismatch: marketing that implies one model and a checkout that performs the other. Audit the promise, not just the setting.
Currency must survive the whole journey
Markets can show local currency on the storefront while specific payment or discount configurations pull the checkout back to the store currency. Shoppers experience that as a bait and switch at the final step, with an exchange rate they did not choose and a bank fee they did not expect.
Test every active market the way its shoppers arrive: local currency displayed, a discount code applied, each payment method offered in that market checked to the confirmation. Currency that reverts at any step is a defect, and it tends to hide behind the least used payment method, which is often the one that market prefers.
Translation that stops at payment is a trust break
Storefront translations now cover themes well, but checkout fields, error messages, and the payment step itself draw from separate language settings. A shopper who browsed entirely in her language and meets English at the card fields experiences the switch exactly where trust matters most, and hesitation at that step is expensive.
Cover the full string set: field labels, validation errors, shipping rate names, and the confirmation. Error messages deserve special attention, because a declined card explained in the wrong language turns a recoverable moment into an abandonment. A native speaker pass over the checkout, not just the storefront, is the audit that catches it.
Every active market needs a working rate
The most embarrassing cross-border failure is a market that is active, translated, and advertised, with no shipping rate configured for it. The shopper completes every field and learns at the shipping step that delivery is impossible. Analytics rarely surface this, because the abandonment looks identical to price hesitation.
Keep a market matrix: for each active market, the rate options, the duty model, the currency behavior, and the checkout language, verified by a test order each time settings change. Markets multiply configuration quietly. The matrix is how a small team keeps thirty small checkouts behaving like one good one.
Related reading
These earlier guides go deeper on connected questions: One Page or Multi-Step? The Checkout Layout Question With No Universal Answer, Checkout A/B Tests Most Stores Run Too Short: Sample Size Without the Math Degree, and Checkout Session Timeouts: Recovering Shoppers Instead of Losing Them.
Frequently asked questions
Why do cross-border checkouts convert worse even with a good layout?
Because the friction is upstream: duty surprises, currency reverting to the store currency at payment, and translations that stop before the payment step. Layout cannot compensate for any of those.
Should duties be included in the price or added at checkout?
Either model works if it is consistent. The failure is a mismatch between what the product page implies and what checkout does.
How should a store audit its markets setup?
Place test orders as a shopper in each active market: local currency, a discount code, every payment method, and confirm shipping rates exist and the language holds through confirmation.
What is the most common hidden defect?
An active market with no shipping rate configured. Shoppers complete the whole form and only learn delivery is impossible at the shipping step.