Changing a reload button from blue to green reduced £20 slot reloads placed between midnight and 04:00 by 16.3% across a six-week trial run on a UK-licensed casino's mobile web funnel. The effect held after controlling for day of week, payday proximity and the operator's own weekend bonus schedule, and it did not come with a matching drop in overall deposits. The finding lands awkwardly for an industry that has spent four years arguing about stake limits and affordability checks while quietly ignoring the colour of the thing punters actually tap.

The trial, run between 14 October and 24 November 2024 by a mid-sized operator with roughly 340,000 monthly active UK accounts, split traffic 50/50 at the point of the reload prompt. Control saw the existing blue primary action button. Variant saw an otherwise identical button in a green that sat at roughly #1DB954 on the hex scale — the same shade family several payment processors use for "confirm". Nothing else changed. Same copy ("Reload £20"), same placement, same friction, same two-tap Apple Pay and open banking flow behind it.

What the numbers actually showed

Across the full six weeks, the variant arm produced 16.3% fewer reload transactions initiated between 00:00 and 03:59, with a 95% confidence interval of 11.1% to 21.4%. That is a real effect, not noise. It is also narrower than it first sounds.

Total deposit volume per user fell 1.4% in the variant arm — statistically indistinguishable from zero at the sample size involved. Average deposit size rose 2.1%, because users who did reload after midnight tended to top up by £30 or £50 rather than the £20 default. The net revenue impact was negative but small: approximately £0.09 per user per month, which the operator's own analysts flagged as within the range they'd accept to reduce late-night intensity.

The interesting split is by user segment. Among accounts with a verified annual deposit under £2,000 — the bulk of the base — the reduction was 21.7%. Among higher-deposit accounts, the reduction was 6.2% and not statistically significant. In other words, the green button mostly deterred the people who were already closer to the edge.

Why midnight matters more than the button

Midnight to 04:00 is the period UK regulators have circled for years. GamCare's helpline data consistently shows calls peaking between 22:00 and 02:00. The Gambling Commission's own 2023 remote gambling research noted that late-night sessions are disproportionately associated with chasing behaviour and with deposits made after a loss. A £20 reload at 01:40 is not, in isolation, a problem. A £20 reload at 01:40 that follows three previous £20 reloads in the same session is a different animal.

The green button appears to interrupt exactly that loop. Session-level analysis showed the median number of reloads per late-night session fell from 2 to 1 in the variant arm. It didn't stop the first reload. It made the second and third slightly less likely.

The design question nobody wants to ask

Here is the uncomfortable part. If a colour change moves behaviour by 16%, then the existing design — the blue button, the pre-filled £20, the "Quick reload" microcopy — was already nudging people toward depositing. Operators have spent a decade A/B testing deposit flows to maximise conversion. They have gotten very good at it. The same toolkit that lifts conversion by 8% on a Tuesday afternoon lifts it by more at 1am, when the user is tilted.

That is not a conspiracy. It is just what optimisation does when the metric is deposits and the context is unconstrained. The green button trial is notable precisely because it inverts the objective function — it optimises for a reduction in a specific behaviour, at a specific time, for a specific segment.

The obvious follow-up: why green? The trial team's own hypothesis was that green reads as "confirm/go" in payment contexts, which should have increased conversions. It didn't. The post-hoc explanation they landed on was that green also reads as "safe/regulated/proceed with care", and that at 1am, in a session already in progress, a button that feels like a considered confirmation rather than a fast action introduces just enough hesitation to break the loop. That is a hypothesis, not a finding. It could equally be that the variant simply looked less like the rest of the app's dark UI and created a moment of unfamiliarity.

What would falsify this

Three things would kill the result. First, if a follow-up trial with a green button that matches the app's existing palette shows the same 16% drop, the colour-specific explanation is dead and we're looking at novelty effect. Second, if the effect decays over eight to twelve weeks — which is the standard pattern for UI-driven behaviour change — then we're looking at a temporary interruption, not a durable reduction. Third, if the same button change applied to first deposits of the day shows no effect, then the mechanism is specific to reload behaviour and not to deposit behaviour generally.

None of those tests have been run publicly. The operator has not committed to rolling the change out, and the trial was not pre-registered, which means the 16.3% figure should be treated as suggestive rather than definitive. It is one operator, one quarter, one segment.

Where this fits in the wider picture

The UK's stake limit regime for online slots — £5 per spin for adults, £2 for 18-24s, in force since September 2024 — was designed to cap the size of individual bets. It says nothing about the frequency of reloads, the design of deposit prompts, or the time of day. The green button sits in a gap the regulations don't cover and probably can't.

That gap is where most of the actual harm reduction happens, or doesn't. A £5 stake limit is easy to enforce and easy to verify. A button colour is neither. But the button colour moved behaviour in a direction the stake limit cannot, because it targets the moment before the session extends rather than the size of each bet within it.

There is a version of this that becomes standard practice: operators running pre-registered trials on deposit-flow design, publishing the results, and accepting small revenue hits for measurable reductions in late-night intensity. There is another version where the finding stays in a slide deck, gets mentioned in a panel at ICE, and nothing changes because the effect is small enough to ignore and the incentive to ignore it is large enough to matter.

The 16.3% figure is not going to reshape the industry. But it does raise a question the industry has been avoiding: if a colour change works, what else in the deposit flow is currently working the other way, and who is going to test it?