A £10 cashout fee, displayed at the point a withdrawal is requested rather than buried in terms, cut completed withdrawal requests placed between 1am and 4am by 29% in a controlled trial run across three UK-licensed operators last autumn. The drop was not matched in a control group shown the same fee only in the summary terms, where late-night requests fell 4%. The finding matters because it reframes a familiar compliance argument: the fee itself was never the deterrent — its timing was.

The trial, run over eight weeks from 16 September to 11 November 2024 and covering 41,600 withdrawal sessions, is one of the few pieces of published evidence on how interface design at the cashier screen affects player behaviour rather than just operator margin. It also lands awkwardly for an industry that has spent a decade treating cashout friction as either a revenue line or a regulatory nuisance, rarely as a design decision with measurable consequences.

What the trial actually measured

Three operators — a mid-market sportsbook with casino attached, a slots-led casino, and a smaller poker room — agreed to randomise withdrawal sessions into two arms. In the treatment arm, a £10 fee appeared as a line item on the confirmation screen, with the net amount and the fee shown separately before the player pressed confirm. In the control arm, the same £10 fee applied but was disclosed only in the general terms and the transaction history after the fact. Both arms faced identical wagering checks and identical processing times.

The headline result was the 29% reduction in 1am–4am completions in the treatment arm. But the breakdown is more interesting than the headline:

• Requests abandoned at confirmation: 18.4% in the treatment arm versus 6.1% in control, concentrated almost entirely between midnight and 5am.
• Requests reduced in size to fall under a fee-free threshold: 11.2% of treatment-arm sessions, where operators offered a £50 fee-free band. Median withdrawal in those sessions fell from £140 to £48.
• Requests delayed to the following morning: 7.9% of treatment-arm sessions, with a median delay of 9 hours 20 minutes.
• Complaints per 1,000 withdrawals: 2.3 in treatment versus 4.7 in control — the upfront fee generated fewer disputes, not more.

That last figure is the one operators should sit with. The intuition that showing a fee upfront inflames players is, on this data, backwards. What players resent is discovering the fee after the fact, usually at 2am, usually when they have already mentally spent the money.

The late-night effect is really a decision-quality effect

The 1am clustering is not incidental. The trial's authors note that late-night sessions show measurably different characteristics: higher average stake per spin, shorter intervals between bets, and a greater likelihood of the session having started after a losing run in the sportsbook. The treatment arm's fee prompt functioned as a circuit break at exactly the point where a player's judgement is most likely to be compromised.

This is the responsible gambling angle that operators tend to skip. A cashout fee shown at confirmation is not merely a pricing disclosure; it is a deliberate interruption inserted into a flow designed to be frictionless. The 18.4% abandonment rate at confirmation is, from a harm-reduction standpoint, a feature. From a short-term revenue standpoint, it is a cost. The two are not reconcilable, and pretending otherwise is where most operator comms fall down.

Why the fee is usually hidden

Cashout fees in the UK market are typically £1 to £5 for standard bank transfers, rising to £10–£15 for same-day or instant methods. The Competition and Markets Authority's 2018 enforcement work on unfair terms pushed most operators to disclose fees more clearly, but disclosure in terms and conditions is not the same as disclosure at the point of decision. The trial's control arm is, in effect, what most of the market still does: compliant on paper, invisible in practice.

There are three reasons the industry has resisted point-of-decision fee display, and only one of them is defensible.

The defensible one is processing cost. Instant withdrawals genuinely cost operators more, and a fee that reflects that is legitimate. The indefensible ones are margin extraction from players who do not notice, and the related hope that friction at cashout keeps balances on account longer, where they are more likely to be re-staked. The trial suggests the second motive is self-defeating: delayed withdrawals in the treatment arm were re-staked within 48 hours in 34% of cases, and those re-stakes had a median value 2.6 times the original withdrawal. Friction at the cashier does not protect the balance; it just relocates the decision.

The £50 threshold problem

The 11.2% of players who reduced their withdrawal to stay under the fee-free band are the group worth watching. A player who wanted £140 and takes £48 has left £92 on account, and the trial's re-staking data suggests a meaningful share of that gets played rather than withdrawn later. Fee-free thresholds are a legitimate product design, but they interact with fee display in ways that can quietly increase play. Any operator adopting upfront fee display should model this effect before assuming the change is harm-neutral.

What this means for operators and for the regulator

The practical implication is narrow and testable. Displaying the fee as a line item on the confirmation screen, with the net amount shown separately, is cheap to implement, reduces complaints, and appears to reduce late-night withdrawal completion. The cost is the 29% of late-night requests that do not complete, plus whatever share of the reduced-withdrawal cohort would otherwise have withdrawn in full.

For the Gambling Commission, the trial raises a question it has so far avoided: whether point-of-decision disclosure should be a licence condition rather than a matter of operator discretion. The Commission's focus on financial risk checks has, reasonably, centred on deposit-side interventions. The withdrawal side has been treated as a consumer-protection matter rather than a harm-reduction one. This data suggests the distinction is artificial. A player at 2am trying to cash out a £300 balance after a session that has gone badly is making a decision about stopping, and the interface around that decision is a harm-reduction lever whether or not anyone labels it one.

The obvious counter-argument is that the trial is small — 41,600 sessions across three operators is not the whole market — and that the late-night cohort may be atypical. That is fair, and the effect sizes should be treated as indicative rather than settled. But the direction is consistent across all three operators, and the complaints figure is hard to explain away.

The open question is whether the 29% figure holds when the fee is smaller. A £2 fee displayed upfront may not produce the same interruption; it may simply be absorbed. If the effect is driven by the size of the fee rather than its visibility, then the policy lever is fee caps, not disclosure, and the industry's current spread of £1 to £15 becomes the thing to argue about. Nobody has run that trial yet. Given how cheap it would be to run, that is the more interesting omission.