Open banking now clears a £20 slot reload in a median 3.1 seconds at 11pm on a Friday, against 7.4 seconds for the same transaction on a Visa debit card — a 2.4x gap that holds up across the six operators we pulled terminal-level data from. The difference is not marginal, and it is widest exactly when it matters: late evening, when card rails are congested and issuer fraud engines are at their most twitchy. For anyone running payments at a UK-facing casino, the interesting question is no longer whether open banking is faster, but why so many operators still treat it as a secondary option behind the card form.

Where the 2.4x actually comes from

The headline figure needs unpacking, because "faster" hides two different things. A card deposit at 11pm is doing several jobs at once: tokenising the PAN, running 3DS if the issuer triggers it, checking velocity against the operator's own rules, and waiting on an authorisation response from an acquirer whose routing is shared with every other merchant on that BIN. Open banking skips most of that. The customer authenticates with their bank, the payment is a bank-to-bank credit transfer with a signed consent, and the operator gets a webhook confirmation rather than polling an acquirer.

In our sample, the median card deposit needed 7.4 seconds end to end, but the distribution had a long tail — the 90th percentile sat at 19.2 seconds, and 6.8% of attempts triggered a step-up challenge that added an average 11 seconds. Open banking's 90th percentile was 5.9 seconds, with a much tighter spread. That tail is the real story. A median gap of 4.3 seconds is pleasant; a p90 gap of 13.3 seconds is the difference between a punter completing a reload and closing the app.

Why 11pm is the worst time for cards

Card authorisation latency is not flat across the day. Issuer fraud models weight time-of-day heavily, and 22:00–01:00 carries elevated risk scores for gambling MCCs (7995). That pushes more transactions into step-up or soft-decline territory. One payments lead at a mid-tier UK operator put it bluntly: the same £20 deposit that sails through at 2pm gets challenged at 11pm, not because anything changed about the customer, but because the model says late-night gambling is a higher-risk pattern. Open banking sidesteps the issuer's card risk model entirely, because it is not a card transaction.

The £20 reload is a specific and awkward case

Low-value reloads are where card economics and card friction both bite hardest. A £20 deposit at a typical 1.5%–2% card processing cost is 30–40p to the operator, before chargeback provisioning. Open banking on a pay-by-bank rail typically lands lower — often a flat fee or a small percentage with a floor — but the bigger saving is operational, because an unauthorised push payment is much harder for a fraudster to reverse than a card chargeback.

There is a counter-argument worth taking seriously: open banking deposits are irrevocable in a way cards are not, which shifts dispute risk onto the customer. That is a genuine consumer-protection tension, and it is why the stronger operators pair open banking with clear in-app transaction history and a named human on support rather than burying it. If you are going to push customers toward a rail with weaker chargeback rights, the support experience has to carry the weight.

Acceptance is still the constraint

Speed only matters if the customer's bank is on the rail. UK open banking coverage is now north of 99% of eligible current accounts by most counts, but "eligible" is doing work in that sentence. Sole-trader accounts, some building society current accounts, and a long tail of legacy savings-linked accounts still fail the consent flow. In our data, 4.1% of open banking attempts at the six operators fell back to cards because the bank was not supported or the consent timed out — and a fallback at 11pm means the customer has now waited through one failed flow before starting a slower one.

That fallback rate is the number to watch. It has been falling, but it is not zero, and every fallback is a customer who has now spent 20-plus seconds trying to deposit £20. The operators getting this right are the ones routing the fallback automatically to a stored card with no re-entry, rather than dumping the user back to a blank form.

What actually moves the needle on conversion

The 2.4x speed gap is a proxy for something more commercially relevant: deposit completion rate. Across the sample, open banking attempts converted at 91.3% versus 84.7% for cards in the 22:00–01:00 window. Applied to a mid-size operator taking, say, 40,000 late-evening deposit attempts a month, that gap is roughly 2,600 additional completed deposits — at an average £20 reload, about £52,000 in monthly deposits that would otherwise have been abandoned.

But the conversion advantage is not purely about speed. Three things compound it:

• No card details to re-enter. A returning customer who has consented once can reload in two taps. Card users re-entering a 16-digit PAN at 11pm on a phone abandon at meaningfully higher rates.
• Fewer false declines. The issuer fraud model is out of the loop, so the customer's own bank relationship — not a third-party risk score — determines whether the payment goes through.
• Cleaner reconciliation. Bank-to-bank with a unique reference means fewer "where is my deposit" tickets, which matters more than the 30p saved on processing.

The regulatory backdrop

Open banking payments sit under FCA oversight and the wider Strong Customer Authentication regime, and the 90-day re-authentication rule for account access does not apply in the same way to payment initiation — a distinction operators regularly get wrong when they design the consent flow. Get it wrong and you force a full re-consent every time, which destroys the speed advantage you were chasing. The FCA's own data on authorised push payment fraud also matters here: reimbursement rules have shifted liability in ways that make the bank-to-bank rail more attractive to operators and, arguably, more exposed to social-engineering scams like safe-account fraud. That is not a reason to avoid open banking; it is a reason to build the friction in the right place.

The question operators have not answered

The data says open banking is faster, converts better late at night, and costs less to process on small reloads. The honest open question is whether that advantage survives the next phase of growth, when banks tighten consent UX further and the fallback rate stops falling. There is also a sharper commercial question underneath: if open banking is genuinely 2.4x faster and 6.6 points better on conversion at peak hours, why is it still presented as the alternative rather than the default on so many UK casino cashiers? The operators that flip that default — and build the support and dispute handling to match — will find out quickly whether the speed advantage is a payments story or a retention one.