A customer requests a withdrawal at 11pm on a Friday. Your system approves it four minutes later. The money lands in their bank account on Tuesday afternoon. From where the customer sits, those four minutes count for nothing. They waited four days, and that is the number they repeat to everyone who asks.
Product teams spent the last decade sharpening the money coming in. Checkout dropped from six fields to one. Card vaults learned to retry failed charges at better hours. Apple Pay and Google Pay removed the keyboard from the equation entirely. Money going out kept running on the same overnight batch it ran on in 2012, because nobody in the room owned it.
That gap now shows up in churn numbers.
The rails stopped being an excuse
Real-time settlement is no longer a pilot programme in most of the markets that matter. The UK has run Faster Payments since 2008. SEPA Instant Credit Transfer went live across the euro area in 2017, and the EU’s Instant Payments Regulation made it compulsory: euro-area payment service providers had to be able to receive instant euro transfers from January 2025 and send them from October 2025. India’s UPI clears billions of transactions a month. The US Federal Reserve launched FedNow in July 2023, alongside The Clearing House’s RTP network.
Push-to-card payouts reach a debit card in under half an hour on the major schemes. Open banking payment initiation moves funds bank to bank without touching a card at all.
When a payout still takes three days, the delay lives inside the company’s own stack.
Where the days actually go
Four things usually eat the time, and none of them are the network.
Batch reconciliation. The finance team runs a payout file once a day at a fixed hour. A request submitted at 11pm on Friday misses Friday’s run, waits out the weekend, and joins Monday’s file.
Manual review queues. Risk flags a withdrawal, a human clears it, and that human works office hours in one timezone. Weekend requests sit in a queue until Monday morning.
PSP settlement cycles. Even when your provider supports instant payouts, your contract may settle on T+2 and prefund from your balance rather than theirs. Engineering built for speed; procurement bought a slower product.
Treasury float. Someone decided that holding funds for 48 hours improves the cash position. It does, until support tickets and one-star reviews start costing more than the interest earned.
Fraud checks have to move earlier
Speed changes the fraud model, and this is the part teams underestimate. A SEPA Instant transfer settles in seconds and offers no chargeback mechanism. Once it is gone, you are chasing the recipient bank, not reversing a transaction.
So the checks move to the front. Identity verification at signup rather than at first withdrawal. Device and behavioural signals collected across the whole session, not just the payout request. Sanctions and PEP screening run at onboarding and refreshed on a schedule. Risk scoring that runs in under 200 milliseconds against the payout decision itself.
Teams that get this right end up approving more payouts instantly, because they already know who the customer is by the time the request arrives. Teams that bolt instant payouts onto a review process designed for T+2 either slow everything down again or start writing off losses.
The markets that feel it first
Pressure lands hardest where users withdraw often and regulators demand verified identity. Trading and brokerage apps compete openly on how fast a sell order turns into money in a current account. Gig platforms discovered that same-day driver payouts affect supply on a Friday night. Insurers now advertise claim settlement in hours. Marketplaces fight for sellers on settlement terms.
Regulated online entertainment sits in the same bracket, and it is one of the few consumer categories where withdrawal times get published and compared directly. Comparison sites rank operators on payout speed alongside licensing and product range, which gives the metric a public scoreboard most industries do not have. Click here for an example of how those comparisons are laid out for consumers.
Any sector where a third party publishes your payout times will feel the commercial effect before your own analytics catch it.
What to measure instead
Most dashboards report approval time, which flatters the team and tells the customer nothing. Three changes fix that.
Measure time to cash: the gap between the customer pressing the button and the money being spendable. Report the 95th percentile rather than the average, because the average hides every weekend request and every manual review.
Then break the timeline into segments. Request to risk decision, risk decision to file submission, submission to settlement. Most companies find one segment holds 80% of the delay, and it is rarely the one they assumed.
Finally, publish a number. An SLA that says funds arrive within one hour for verified accounts forces the internal argument between finance, risk and engineering to happen once, properly, instead of every time a customer complains.
The competitive position
Instant payouts will stop being a differentiator at some point. Card acceptance did. Same-day delivery did. What remains is the cost of being the company that still takes three days after the rest of the market moved.
The infrastructure exists and is regulated into availability across most major markets. The remaining delay is a decision about batch windows, staffing and treasury policy. Customers cannot see any of that, and they read all of it as one thing: you are holding their money.
