Search
Choose a style
Dark
Light
Time to read: 6 min

Cascading is not free: When payment retries hurt more than they help

Alexander Mikhailovski, Co-Founder and Chief Product Officer at eComCharge.
Alexander Mikhailovski. Image credit: eCom Charge

Repeated authentication requests, declining transactions and duplicate charges can undermine the customer experience when PSPs use payment cascading to recover failed transactions.

Alexander Mikhailovski, Co-Founder and Chief Product Officer at eComCharge, examines the operational and commercial trade-offs behind additional payment attempts, and explains to Payment Expert why providers must assess whether each stage delivers enough value to justify its cost and the effort required from customers.


A customer enters a one-time password (OTP) to confirm a payment. Instead of a successful payment message, they encounter more redirects and another request for an OTP. They have already confirmed the purchase. Why are they being asked to do it again, and what happened to the first attempt?

Behind the screen, a payment service provider (PSP) has received a decline and sent the payment to another processing partner. These partners may be acquiring banks or other PSPs; I refer to them here as payment providers. The cascade offers another chance of approval, but the customer must complete another authentication challenge to pursue it.

PSPs using our eComCharge WLS platform frequently request cascading to increase conversion and strengthen merchant loyalty. In those discussions, additional approvals are the priority, and repeated 3D Secure (3DS) authentication is generally accepted as part of the trade-off. Whether that trade-off pays off depends on what each additional attempt actually contributes. In our experience, a few recurring problems deserve particular attention.

When the reason for failure is missing

The first difficulty is knowing what actually went wrong. A payment that failed because of a limitation in one provider’s channel may succeed elsewhere. If the cardholder simply has insufficient funds, changing providers does not put more money in the account.

Providers often give us no specific reason for a decline. In our integrations, we standardise response codes from different providers so PSPs can use them consistently in routing rules. However, if a provider returns only a generic decline, standardising that response cannot reveal the underlying reason for the failure.

That leaves the PSP with an awkward choice. Stopping at every generic decline would lose some recoverable payments. Continuing through the full cascade may mean paying for attempts that were unlikely to succeed. Without a reliable reason, neither decision is as straightforward as a routing diagram might suggest.

Authentication is part of the cost

From the customer’s side, each fresh challenge is another interruption. The platform may be trying a different route, but the customer is still trying to make one payment. They may leave before the cascade finishes. If insufficient funds caused the original failure, several confirmations can still end in a decline, leaving the PSP with attempt fees where applicable and the merchant without the payment.

Supporting external 3DS authentication does not automatically allow the same authentication result to be reused across successive authorisation attempts. A fallback may involve a different acquirer or merchant account, so reuse must be explicitly supported for that scenario. Where fresh authentication is required, the issuer may ask the customer to complete another challenge.

When a reported failure was actually a success

We have also encountered a more serious problem: a provider reports that an operation failed although it actually succeeded. The platform receives that failure and sends the payment to the next route. If both attempts complete the charge, the customer pays twice. Duplicate payouts can be harder to resolve: an extra card charge can be refunded, while recovering a completed payout may depend on the recipient’s cooperation and is not guaranteed.

The difficult part is that this is not an obvious timeout with an unknown outcome. The platform received what looked like a definitive failure. Even a rule that permits retries only after failure cannot prevent this on its own.

Keeping all attempts linked to the original payment helps make sense of what happened when late notifications or reconciliation reveal conflicting outcomes. If a provider repeatedly reports incorrect statuses, there is good reason to reconsider its place in an automatic cascade. These controls help manage the consequences, but cannot guarantee that the next attempt will be stopped in time.

What another attempt actually adds

The same caution applies to the length of the cascade. A first fallback that works well tells us little about a fourth or fifth route. By then, earlier providers have already failed to complete those payments. The remaining obstacles may be harder to solve by changing routes, and the customer has already spent more time waiting.

That is why returns can diminish while costs keep growing. A later provider may still recover payments thanks to a particular capability. Its value at this stage depends on how many payments it can complete after the earlier routes have failed. Its approval rate as the first choice does not tell us that.

A simple hypothetical example makes the trade-off visible. An extra stage receives 1,000 payments and recovers 20. Each attempt costs EUR 0.02. Suppose each recovered payment contributes EUR 0.60 to the PSP after other variable costs but before these attempt fees. The PSP earns EUR 12 from the recovered payments but pays EUR 20 in attempt fees. This extra stage therefore generates a loss of EUR 8 for the PSP, despite improving conversion. The figures will differ with each PSP’s tariffs and economics.

The merchant may see those 20 recovered payments differently, because a successful payment has commercial value beyond the PSP’s processing fee. That is a conversation worth having between them. The question is whether another attempt recovers enough payments to justify its cost and the extra effort required from the customer. A useful answer begins with the customer’s payment as a whole.

Knowing when to stop

It is easy to lose sight of that payment in a report full of individual attempts. Five declined authorisations may belong to one customer trying to pay once. Separating those two views helps us see whether the longer cascade actually completes more payments, and what it costs to do so.

For each stage, the useful comparison is between the payments it recovers and the cost of every attempt made there. A payment that succeeds twice still counts as one recovery. Waiting time, further challenges and any visible abandonment add the customer’s experience to that picture. This gives the merchant and PSP something concrete to discuss beyond an approval percentage.

Route sequence, issuer or BIN, amount and geography can all affect the answer, while unknown decline reasons still leave gaps. Where retries are permitted and outcomes reliable, controlled comparisons of shorter and longer cascades can help show what an extra stage adds. Historical recoveries alone cannot tell us whether a shorter checkout would have kept more customers.

Scheme and acquirer restrictions, including explicit instructions not to retry, still set the boundaries. Within them, a stage that adds too little value may be better removed or tried elsewhere in the sequence. That judgement can change as traffic and provider performance change.

As the developers of eComCharge WLS, we see these trade-offs in how our PSP clients use the platform. They raise a question for any PSP: does the next attempt offer enough value to justify its cost and the customer’s time? The customer entering another OTP is simply hoping that this time, the payment will go through.

Subscribe to our newsletter