Ask a player to invent a nickname before they can put £10 down and you will lose roughly one in five of them. That is the headline from an A/B test run across four UK-licensed casino brands in Q1 2025, where adding a mandatory "display name" field to the registration flow reduced completed first deposits by 18.4% during the 21:00–22:00 hour. The same field cost only 6.1% of conversions at 10:00, which tells you something more interesting than "forms are friction": the cost of asking depends on when you ask.

The test is worth taking apart because the industry keeps reaching for the same lever — collect more data at sign-up, personalise earlier, build a profile before the player has even seen a game — and keeps being surprised when the deposit numbers wobble. What follows is what the data actually showed, where the effect concentrates, and why the 9pm cliff should worry operators more than the average suggests.

What the test actually did

Four brands, mid-market UK-facing, roughly 90,000 new registration sessions split 50/50. Control saw the standard flow: email, password, date of birth, address, deposit. Variant inserted a single required field between password and date of birth — "Choose a nickname (this is what other players will see)" — with a 3–16 character limit, a uniqueness check, and no skip option. Everything downstream was identical: same welcome offer, same payment page, same KYC triggers.

The headline result was an 18.4% relative drop in first-time depositors (FTDs) in the 21:00–22:00 window, against a 6.1% drop at 10:00–11:00. Overall, across all hours, the variant lost 11.2% of FTDs. That overall figure is the one most operators would have benchmarked against, and it is the one that hides the story.

Two things to note about methodology before anyone quotes these numbers in a board deck. First, the nickname was genuinely mandatory — no "decide later" escape hatch, which is how most real implementations behave. Second, the uniqueness check added a variable latency of 200–900ms depending on load, and we did not isolate that from the field itself. Some of the 9pm effect is almost certainly the rejection loop when a player's first three choices are taken, not the act of typing.

The rejection loop is the real villain

Of players who hit the nickname step at 9pm, 14.7% had their first choice rejected as already taken. Of those, 31% abandoned within the next 90 seconds. At 10am the rejection rate was 9.2% and the post-rejection abandonment was 19%. Nicknames are more contested in the evening because that is when the largest cohort of players is registering — so the pool of available names is thinner exactly when the queue is longest.

That is a design problem masquerading as a data problem. A uniqueness check against a live database, surfaced as a red error under a text field, is the worst possible way to ask for a name. It punishes the player for a collision they cannot see coming.

Why 9pm behaves differently from 10am

The hour-by-hour split is where the test earns its keep. FTD conversion in the variant, indexed to control at 100:

Hour	Indexed FTDs (control = 100)
10:00	93.9
14:00	91.2
18:00	88.7
21:00	81.6
23:00	84.3

The decline is monotonic into the evening and recovers slightly after 23:00. Three plausible mechanisms, and the test cannot fully separate them:

Intent quality. Evening registrants skew toward impulse. They arrived from a notification, an ad, or a friend's screenshot, and their patience for a form is measured in seconds. Morning registrants are more likely to be deliberate — comparing offers, reading terms — and a deliberate player will fill in a nickname without complaint because they have already decided to deposit.

Device and context. Evening sessions are more mobile-heavy (78% versus 64% at 10am in this sample) and more likely to be interrupted. A mandatory field with a validation round-trip is a bigger relative cost on a phone with a flaky connection than on a desktop.

Cognitive load at the point of payment. By 9pm, a player has usually made several small decisions already that day. Adding one more creative decision — invent a name — immediately before the deposit is a poor sequencing choice. The morning player has more decision budget left.

If the third mechanism is doing the heavy lifting, the fix is not to remove the field but to move it. Collect the nickname after the first deposit, when the player has already committed and the name is being requested for a purpose they can see (a leaderboard, a chat handle, a tournament alias). The field is not inherently expensive; it is expensive at that moment.

What this means for onboarding design

The instinct after a result like this is to strip the flow back to email and password and worry about profile data later. That is broadly right, but the test suggests three more specific rules.

Never put a uniqueness check in the critical path to deposit. If you need a unique handle, generate one — player_8472, or a random adjective-noun pair — and let the player change it later. A default that can be edited is worth more than a blank field that must be filled.

Sequence creative decisions after commitment, not before. Nicknames, avatars, favourite teams, preferred game categories: all of these convert better as post-deposit prompts, and the data you collect is more accurate because the player is now motivated to complete a profile rather than tolerate an obstacle.

Treat hour-of-day as a segmentation variable, not noise. An 18.4% loss at 9pm and a 6.1% loss at 10am average to a number that describes neither. If your funnel analytics collapse to daily totals, you are blind to the exact window where most of your acquisition spend lands.

There is a compliance angle worth flagging too. Nicknames are player-visible, which means they are a moderation surface — abuse reports, impersonation, and the occasional attempt to use a handle as a de facto advert for an unlicensed operator. Collecting them later does not remove that obligation, but it does mean you are moderating a smaller, more engaged population rather than every registration that bounces off the form. Under UK Advertising Code expectations around socially responsible messaging, a visible handle is also something you have to monitor, and a generated default is easier to police than a free-text field.

The question the test leaves open

The uncomfortable implication is that the 18.4% figure is not a property of the nickname field. It is a property of asking for something creative, at 9pm, on a phone, with a validation round-trip, from a player who has not yet decided whether they trust you. Change any one of those and the number moves.

Which raises the question operators should be asking their own funnels: how many of your mandatory fields are doing real work at the moment they appear, and how many are just the accumulated sediment of decisions made years ago by someone who was not measuring abandonment by the hour? The nickname test is a clean example because the field is obviously optional in spirit. The same audit applied to date of birth, address lookup, or the second line of a phone number verification would likely produce a less comfortable answer — and nobody has run that test at 9pm yet.