Why foreign cards get declined, and what local acquiring changes
When a customer in São Paulo or Jakarta fails to pay for your software, the usual assumption is that their card was short of funds. Far more often, the issuer saw a charge arriving from an unfamiliar country and said no on principle.
Decline rates on cross-border card payments are consistently worse than domestic ones, and the gap widens in exactly the markets most software businesses want to expand into. The reasons are structural, which is good news: structural problems have structural fixes.
What the issuer actually sees
When you charge a card, the request reaches the customer's issuing bank carrying more than an amount. It carries the acquirer's country, a merchant category code, a descriptor, and a set of flags describing how the payment was made.
If your acquirer sits in Singapore and your customer's bank is in Colombia, the issuer sees a foreign-acquired charge against a local card. Its fraud model weighs that. Combined with anything else slightly unusual — a first-time merchant, a round amount, an unfamiliar descriptor, a late hour — the risk score crosses a threshold and the charge is refused.
The customer sees a generic failure. They rarely try again. You see a decline code that tells you almost nothing.
The things that compound
Soft declines that look hard
Issuers distinguish between "never retry" responses and ones meaning "not right now". Most checkouts treat both as failure. A meaningful share of cross-border declines are retryable, and a retry a few hours later on a different route often succeeds. If you are not separating these, you are discarding revenue.
Local authentication rules
Several markets require a step your foreign acquirer may not support properly. India has its own authentication and mandate framework for recurring payments. Europe requires strong customer authentication with specific exemption handling. If that step is handled badly, the payment fails at the exact moment the customer was ready to pay.
Subscription decay
A one-off decline is annoying. On a subscription it compounds — each monthly renewal is another chance for the same issuer to refuse the same foreign charge. Involuntary churn in international cohorts is often larger than voluntary churn, and it rarely appears in product metrics because it looks like a billing problem rather than a retention one.
The customer's own mental model
Someone who sees an unfamiliar foreign descriptor on their statement is more likely to call their bank and dispute it. Foreign-looking charges generate disputes that domestic ones do not, independent of whether anything was wrong.
A diagnostic worth running: split your authorisation rate by the customer's country and compare each against your home market. If a market is ten or more points below, the problem is almost certainly routing, not demand. That gap is a revenue number you can put in front of a finance team.
What local acquiring changes
Local acquiring means the payment is processed by an entity licensed in your customer's country, with a direct relationship to the local card schemes. The charge reaches the issuer as a domestic transaction.
Three things follow. The fraud model no longer flags it as foreign. Local authentication is handled the way the issuer expects. And the descriptor on the statement can be one the customer recognises, in their own language where that applies.
The practical obstacle is that getting licensed in each country, maintaining scheme memberships, and keeping up with local rules is a company-sized undertaking per market. Which is why most businesses reach it through a provider that has already done it, rather than building it.
You are not changing the customer's willingness to pay. You are removing a signal that made their bank nervous.
And the payments that were never cards
There is a limit to how far card optimisation goes. In some markets a large share of people who would happily buy your product do not have a card they can use online at all. No amount of routing improves a payment that was never going to be attempted.
In those markets the fix is offering what people actually use — Pix, UPI, a wallet, a bank transfer — alongside cards. We wrote about those methods here.
Where to start
- Measure authorisation rate by customer country, not in aggregate. Aggregates hide the markets that are broken.
- Separate soft and hard declines, and retry the soft ones on a schedule.
- Check what your statement descriptor looks like to a customer abroad.
- For your worst-performing markets, find out whether your provider acquires locally there or routes cross-border.
- Track involuntary churn separately from voluntary churn on international subscriptions.
The first two cost nothing but engineering time and usually pay for themselves. The third is a fifteen-minute check. The fourth decides whether you need a different provider for those markets.
Next: the methods people reach for when they do not use a card.
Pix, UPI and GCash