Home › Separate Confirmation Screens Cut Duplicate Deposits 33%

Separate Confirmation Screens Cut Duplicate Deposits 33%

Separate Confirmation Screens Cut Duplicate Deposits 33%

A 14-week split test across three licensed operators found that adding a dedicated confirmation screen before a deposit is finalised reduced duplicate deposits by 33.4%. The effect held across card, e-wallet, and open banking rails, and it persisted after controlling for deposit size and player tenure. The result matters less for its headline number than for what it implies: a meaningful share of duplicate deposits are not player error in any meaningful sense, but interface artefacts produced by ambiguous system states.

The study, run between March and June 2024, covered 412,000 deposit attempts from 61,800 active accounts across operators licensed in Malta, Denmark, and Ontario. Duplicate deposits were defined narrowly — a second deposit from the same account within 90 seconds of the first, at the same amount, on the same payment instrument — to exclude deliberate repeat deposits and scheduled top-ups. The baseline rate in the control cohort was 1.87 per 1,000 attempts. In the treatment cohort, which inserted a confirmation screen showing amount, payment method, and a 15-second cancellation window before the payment was submitted, the rate fell to 1.245 per 1,000. That 0.625-per-1,000 absolute reduction translates to the 33.4% relative drop in the title.

Why duplicate deposits happen at all

The industry's default assumption has been that duplicate deposits are a user-side problem: a player taps twice, or refreshes a slow page, or forgets they already deposited. That framing drives the usual remedies — debouncing buttons client-side, disabling the submit control after the first tap, and warning text in the cashier. These are cheap and they help. But they address the tap, not the uncertainty that produces the second tap.

Session recordings from the control cohort tell a different story. In 71% of duplicate events, the first deposit had entered a pending state — authorisation requested but not confirmed — for longer than 4 seconds. In 58% of cases, the player had navigated away from the cashier and returned before the second attempt. The second deposit was, in most of these sessions, an attempt to complete a transaction the player believed had failed. The interface had not told them otherwise.

This is a state-visibility problem. A cashier that shows a spinner and nothing else, with no distinction between "we have not yet contacted the processor" and "the processor has authorised and we are waiting on the acquirer," invites the player to resolve ambiguity themselves. The cheapest resolution available to them is to try again.

The confirmation screen as a state contract

The treatment screen was not elaborate. It displayed three fields — deposit amount, payment instrument (masked), and the operator's own internal reference — plus a single primary action and a 15-second countdown during which the player could cancel without the payment being submitted. After the window elapsed, the screen transitioned to an explicit "submitted, awaiting confirmation" state with a reference number the player could quote to support.

The design intent was to make the system's state legible at the moment of highest uncertainty. The countdown is doing specific work: it gives the player a bounded interval in which to notice an error — wrong amount, wrong card — before commitment, which is why the same screen also reduced mis-keyed amounts by 12.1% in a secondary analysis. But the larger effect came from the post-submission state, which removed the ambiguity that had driven most return-and-retry behaviour.

What the numbers do and do not show

The 33.4% reduction is robust to the obvious confounds. Deposit size, time of day, player tenure, and payment rail were all balanced across cohorts by randomisation, and the effect survived stratification on each. The three operators ran the test independently with a shared protocol, which limits single-operator artefacts. A four-week washout period before measurement reduced novelty effects, though it cannot eliminate them entirely.

What the numbers do not show is whether the reduction is permanent. Duplicate deposits are a low-frequency event, and 14 weeks gives adequate power for the aggregate but not for detecting drift. It is plausible that players habituate to the confirmation screen and that some of the effect decays — the mechanism is attention, and attention is finite. A 12-month follow-up would settle this; none has been published.

There is also a selection question the study cannot answer. The treatment screen added roughly 2.1 seconds to the median deposit flow. Players who abandon at that screen are excluded from the duplicate-deposit denominator, and if the screen disproportionately deters the kind of impatient player most likely to double-tap, part of the measured effect is compositional rather than behavioural. The authors report a 3.8% reduction in completed deposits per session in the treatment arm, which is consistent with either explanation. Whether that 3.8% represents friction doing useful work or revenue leakage is, at present, an open empirical question.

The regulatory angle is narrower than it looks

It is tempting to read this as a responsible-gambling finding, and in part it is. Duplicate deposits are a known precursor to unintended spend, and any interface change that reduces them without restricting access is preferable to deposit limits applied bluntly across a whole population. Several regulators have begun to treat ambiguous transaction states as a consumer-protection matter rather than a pure UX concern, on the reasoning that a player who cannot tell whether a payment succeeded cannot exercise informed control over their spending.

But the effect size here is modest in absolute terms. A reduction from 1.87 to 1.245 per 1,000 attempts removes roughly 0.6 duplicate deposits per thousand — meaningful at the scale of a large operator, trivial at the level of an individual player. The confirmation screen is not a harm-reduction intervention in the sense that a deposit limit is. It is a hygiene measure, and it should be evaluated as one.

Where the finding travels

The result is likely to generalise across jurisdictions more readily than most cashier findings, because the underlying mechanism — uncertainty about transaction state — is a property of payment infrastructure rather than of regulation. Markets with slower acquirer settlement, such as those routing through local banking rails with batch processing, should see larger effects, not smaller. Markets with instant settlement and strong push notifications may see less. The study's Ontario cohort, which included a high proportion of Interac transactions, showed a 36.2% reduction against the pooled 33.4%, consistent with this prediction, though the subsample is too small to be conclusive.

The more interesting question is whether the same logic applies upstream. If ambiguous state drives duplicate deposits, it plausibly drives duplicate bets, duplicate withdrawals, and duplicate account-verification submissions — all of which carry their own operational and player-protection costs. None of these has been tested with the same rigour, and the industry's default response to each remains the disabled button.

An open question for operators

The practical implication is not that every cashier should add a confirmation screen. It is that the 33.4% figure reframes where the cost of duplicate deposits sits. If most duplicates are interface artefacts rather than player mistakes, then charging them back to player behaviour — through warnings, cool-downs, or support interventions — misallocates the remedy. The fix belongs in the payment flow, and it is cheap.

What remains unresolved is the trade-off the study measured but did not explain: the 3.8% drop in completed deposits. An operator considering the change has to decide whether a small reduction in conversion is a fair price for a 33.4% reduction in duplicates, and the honest answer is that nobody has published the cohort-level economics that would settle it. Until someone does, the finding stands as a strong signal about mechanism and a weak one about policy — which is, for an industry that has spent a decade blaming the player, an uncomfortable place to be.