By Artjoms Dozorcevs, CEO of Finsei
Many online merchants set up card acceptance early and only revisit it when something starts going wrong. By the time they sell in five countries, that early decision can cost more than their ad budget wins back, argues Artjoms Dozorcevs.
An online retailer turning over €2 million a year usually knows its customer acquisition cost to the cent. Ask the same company what share of attempted card payments gets approved and the room tends to go quiet. Someone thinks it is “around 90%”. Nobody is sure whether that is good.
It matters more than most leadership teams assume. Checkout.com and Oxford Economics estimated in 2023 that merchants lose between 1% and 2.1% of revenue to false declines, meaning legitimate purchases rejected by fraud screening. For a business on single-digit net margins that is a large slice of profit, and it is lost after the marketing money has already been spent getting the customer to the pay button.
I run a payments company, so discount my view accordingly. But the pattern I keep seeing is this: e-commerce businesses rarely stall because demand ran out. They stall because the payment setup they chose at €200,000 in revenue is still the one they have at €20 million.
Where the ceiling shows up
Approval rates first. A card issued in Spain, used on a UK site and processed by an acquirer in a third country looks riskier to the issuing bank than a domestic purchase. The bank is not being irrational. European Central Bank data on card fraud in SEPA shows cross-border transactions making up 10 to 11% of value and 63 to 65% of fraud losses. So issuers decline more of them, and honest customers get caught. The merchant sees a failed payment and, usually, no explanation.
Then cash. Card revenue does not arrive when the customer pays. It arrives when the acquirer settles, often days later and sometimes with a reserve held back. Then it moves to a bank account, and then it gets converted if a supplier wants another currency. Each hop takes time and a margin. A retailer growing 40% a year has to buy stock ahead of sales. If its money spends a week in transit, that growth is financed by an overdraft or it does not happen.
The last one is harder to spot: nobody sees the whole picture. The gateway has one dashboard, the acquirer sends a statement, the bank account sits somewhere else and FX somewhere else again. Finance stitches them together in a spreadsheet at month end. When approvals drop in Germany on a Tuesday, weeks can pass before anyone notices, because the data that would show it is split across four vendors.
How companies end up here
Nobody designs this. A founder picks whichever checkout plugin gets the shop live by Friday, which is the right call at that stage. Later a second provider is added for a market the first covers badly, then a third for a local payment method. In its 2025 merchant study, 451 Research found that 90% of merchants work with two or more payment providers, and 62% say they prefer it that way, up from 50%.
I don’t think having several providers is the problem. Redundancy is sensible, and no serious merchant should depend on a single route to the card schemes. The trouble is that each provider was bought as a tool, by whoever had a local problem that week, and the sum was never treated as infrastructure. Payments ends up owned by an engineer who integrated an API in 2021 and a finance manager who inherited the reconciliation.
Compare that with how the same company treats logistics. Warehousing, carriers and delivery times get a head of operations, a board slide and a yearly renegotiation. Payments moves the same money and gets a Slack channel.
The rails underneath do not help. Much of card acquiring still settles in batches, a design from the era when a merchant was a shop with one terminal and one currency. Cross-border selling was layered on top of that. So was e-commerce itself, come to think of it.
Four questions for the next board meeting
I would rather offer questions than a solution, because the right setup for a fashion retailer selling into the Gulf is wrong for a subscription software firm in the Nordics.
- What is our approval rate by country and card type, and who looks at it every week? If the answer is one blended figure, or silence, start there.
- How many days pass between a customer paying and that money being usable to pay a supplier? Count honestly: weekends, reserves and conversions included.
- What does it really cost us to accept a sale? Add up processing fees, scheme fees, FX spread, chargebacks and the hours finance spends matching reports.
- If our main provider suspended our account tomorrow, how long before we could take payments again? Plenty of merchants learn the answer the hard way.
A caveat. Moving everything to one provider is not automatically the fix, and I say that as someone who sells consolidation. In September Finsei added card acquiring to the accounts our business clients already use for transfers and FX, so settlement lands in the same place the money is spent from. That suits a mid-sized merchant who is tired of reconciling four systems. A large retailer with its own payments team may do better with several acquirers and its own routing rules. What matters is that somebody chose the arrangement on purpose.
The cheapest growth available
Most growth levers left to an established online merchant are expensive. Ad costs rise, new markets need localisation, new products need inventory. Customers who already tried to pay you and failed are about the cheapest revenue there is.
Recovering them is dull work. It means reading decline codes and settlement files, and it will never headline a strategy offsite. I suspect that is why it stays unfixed in so many companies for so long, and why the ones that do fix it rarely talk about it.
