← Back

Splat Games

Making a sweepstakes casino's two-currency, buy-a-proxy model legible and auditable enough that a skeptical first-time visitor understands it, trusts the payout, and can see what they are spending.

Project overview

  • Type: Client · 14-week engagement · mobile-first (responsive web app)
  • Team: with 1 PM and 2 engineers, plus a legal/compliance stakeholder
  • Project type: iGaming / Sweepstakes · Mobile-First · Onboarding & Purchase · Trust & Transparency
  • Role: Lead Product Designer · Art Direction · Information Architecture · UX Writing
  • Methods: Comprehension test · 5-second test · usability testing (SUS) · A/B with guardrail metrics
  • Tools: Otter.ai · Adobe Suite · Maze · Figma + Dev Mode · Amplitude · Statsig
  • Case thesis: Designing a sweepstakes casino where a new player can see, in plain terms, what they pay for, what they play with, and what redeems for cash — plus proof they will be paid and clear visibility of what they are spending — so a model people find confusing becomes one a first-time visitor understands, trusts, and can engage with on informed terms.

The context

Online casinos start from deep skepticism: a new brand is assumed to be a scam or a site that will not pay out. Splat Games is a new brand on the sweepstakes model, where coins are received free with the purchase of another product rather than bought directly, and the platform runs two currencies. That structure adds confusion on top of the baseline distrust. The games are licensed from multiple providers and shown in a broad catalog, and unlike most online casinos that assume desktop, Splat was designed mobile-first to behave predictably on a phone.

It is worth stating plainly where this model sits: the dual-currency sweepstakes structure has always occupied a legal gray zone, and through 2025–2026 a growing number of US states moved to restrict or ban it. The design work here was scoped to comprehension, trust, and spend-transparency within that structure, not to settle its legal status — and the regulatory volatility is treated as a live risk, noted in the limitations.

The problem

The model is confusing and distrusted, and that blocks an informed first purchase. In a comprehension test, only 34% of new players could correctly say which currency redeems for cash after going through a standard sweepstakes flow, and a majority read "purchase a product, receive free coins" as a catch (behavioral + attitudinal). Purchase-flow completion sat at 41%, and cashout status and "what is this currency worth" questions made up 38% of support (operational). A player who cannot tell what they are buying cannot consent to it meaningfully — so legibility was both a conversion problem and a consumer-protection one.

The goal

Make the value exchange, the path to cashing out, and the player's own spending legible, so first-time visitors understand the model and engage on informed terms — measured by comprehension, purchase-flow completion, cashout-related support, and guardrail metrics that catch clarity turning into pressure, rather than by the size of the game catalog.

Empathize — Only 34% of new players could tell which currency redeems for cash, and most read "buy a product, get free coins" as a catch

In this section: Research foundation · Key insights

Research foundation (method)

  • Phase 1 — Interviews (n=18, ~40 min, online-casino players and skeptics, recruited via dscout, transcribed in Otter.ai): how they judge a new casino and what they make of sweepstakes.
  • Phase 2 — Comprehension test (n=28): participants went through a standard sweepstakes purchase and explained what they got and what redeemed for cash.
  • Phase 3 — 5-second test: first-impression reads of the purchase screen, testing whether the value exchange registered at a glance before any reading.
  • Phase 4 — Survey: ~1,081 invited; 200 analyzable responses (18.5% response rate); ~127 completed every item (12.7% completion rate). Attitudinal percentages are computed on the 200 analyzable responses; question types are labeled per question in the appendix.
  • Phase 5 — Usability testing (SUS): task-based sessions on the rebuilt purchase and redemption flows, scored with the System Usability Scale.
  • Phase 6 — Prototype pilot (Amplitude-instrumented, mobile-first, ~600 new players, ~300 per A/B arm, with guardrail metrics monitored): behavior on the rebuilt purchase and redemption flows.

Key insights

1. New players cannot tell which currency has real value. With two currencies and no plain explanation, players could not say what was for play and what redeemed for cash, so they froze at purchase and doubted redemption.

  • Behavioral: 34% correctly identified the redeemable currency after a standard flow.
  • Verbatim (P6, skeptical first-timer) — coded: Currency confusion: "I have two kinds of coins and I genuinely can't tell which one is real money and which is just for fun."

2. The buy-a-proxy indirection reads as a trick. "Purchase music or game coins and receive free casino coins" sounded like a catch unless the exchange was shown plainly, so a structure built for legal compliance created distrust.

  • Attitudinal: a majority of surveyed players assumed there was a hidden condition behind the free coins.
  • Verbatim (P11, lapsed player) — coded: Hidden-catch read: "Nothing is free, so if you're giving me free coins for buying something, I assume the catch is I'll never see a cent back."

3. Trust that you will be paid is the gate on the whole model. Players would not engage a model they already found confusing without proof that real people cash out, so payout evidence and an auditable history were prerequisites for engaging at all.

  • Verbatim (P2, online-casino regular) — coded: Payout doubt: "Show me people actually getting paid and a record of my own cashouts, or I'm not putting in a cent."

Dashboard — Where the sweepstakes model loses new players

Where the sweepstakes model loses new players
Scope: comprehension test (n=28) + survey (n=200)
Guiding question: Why don't new players complete a first purchase?

  Correctly identified what redeems for cash .. 34%
  Read "free coins with purchase" as a catch .. majority
  Purchase-flow completion .................... 41%
  Support that is cashout / currency questions  38%

Key Insight: Comprehension and trust are the block: players who can't tell
what they're buying or whether they'll be paid don't buy.

Define — A first-time visitor had to see plainly what they pay, what they play with, and what redeems for cash, with proof they would be paid and clear sight of their own spend

In this section: POV · How Might We · Principles · Insight→decision map

POV statement. A first-time visitor needs to understand, in plain terms, what they pay, what they can play with, and what redeems for cash, to see proof they will be paid, and to see what they are spending, because the sweepstakes two-currency, buy-a-proxy model reads as confusing and as a possible scam — and an informed choice is impossible without all four.

How Might We

  1. How might we show the value exchange plainly while staying within the model's required legal framing?
  2. How might we make the proxy purchase a clear, low-friction part of the loop instead of an off-site detour?
  3. How might we prove, from the first screen, that real people get paid?
  4. How might we keep spend visible and give players controls, so clarity helps them decide rather than pushing them?

Design principles (each traceable to an insight)

  • The value exchange is legible. What a player pays, what is playable, and what redeems for cash are shown in plain language. (Insight 1, 2)
  • Compliant and comprehensible together. The required purchase-of-a-product structure stays intact while the value is shown clearly. (Insight 2)
  • Proof of payout from first contact. Recent real winners, provider logos, reviews, and an auditable cashout history run throughout. (Insight 3)
  • The loop stays on-platform. The proxy purchase happens inside Splat, on-platform, so friction and off-site distrust stay low.
  • Legibility serves the player, not just the purchase. Spend is always visible, deposit/play limits and self-exclusion are reachable, and age verification gates entry, so making the model clear never tips into making it pushy. (Insight 1, 3)

Insight → decision map

Insight (from Empathize) Concrete design decision
34% knew which currency redeems for cash The purchase and redeem flows state plainly what is for play and what redeems for cash, with the two currencies clearly distinguished
The proxy purchase reads as a catch An in-house proxy game is built so buying its coins (and music) yields casino coins on-platform, with the exchange shown step by step
Players won't engage without payout proof Real-time recent winners, provider logos, reviews, a community, and Billing, Cashout, Partner-cashout, and Rewards history make payouts auditable
Clarity must not become pressure A persistent spend/coin header, reachable deposit and play limits, self-exclusion, and an age gate keep the player in control of an informed decision

Ideate & Craft — The purchase flow showed the value exchange plainly within the compliant structure, payout proof ran throughout, and spend stayed visible

In this section: Design execution · Before → after · Other deliverables

Design execution

  • Legible value exchange — the Purchase Entries and Redeem flows show, in plain language, what the player pays, what they receive to play, and what redeems for cash, with the two currencies visually distinct, all inside the required purchase-of-a-product structure.
  • On-platform proxy loop — an in-house horse game was built so that buying its coins, and music, yields Splat casino coins with minimal friction, keeping the store inside the casino rather than sending players to an external shop. (This proxy structure is also the legally contested core of the model; see Regulatory context.)
  • Proof of payout everywhere — real-time recent winners and amounts, provider logos, reviews, and an internal community appear from first contact, and a persistent header keeps the player's coins, current spend, and a withdraw button always visible.
  • Transparency modules — Billing history, Cashout history, Partner cashout requests, and Rewards history let a player audit their own money end to end.
  • Player-protection controls — deposit and play limits, a self-exclusion path, spend visibility in the persistent header, and an age-verification gate are built into the flow rather than buried, so an informed player stays in control.
  • Sequenced sign-up and KYC — registration and identity steps are ordered to introduce one thing at a time so the KYC requirements do not overwhelm a new user.
  • An original world — an optimistic-cyberpunk universe of characters, bright and approachable, gives a new and unknown casino the feel of a real, cared-for brand and a clear, readable interface.
  • A clean main menu — Home, Purchase Entries, Redeem, Store, Profile, Payment methods, Withdraw, Promo codes, Invite and earn, Referral programs, Support, and Log out, ordered to be easy to follow, with an authenticated lobby leaderboard of top players and latest big wins.

Before → after

Before (standard sweepstakes flow) After (Splat Games)
What the player is buying Legal wording, value obscured Plain "what you pay, play, and redeem"
The proxy purchase An off-site, confusing detour An on-platform loop shown step by step
Proof of payout A claim of legitimacy Recent winners and an auditable history
Spend & controls Buried or absent Visible spend, reachable limits, self-exclusion
Platform Desktop-first Mobile-first and predictable

Other deliverables

Built in Figma with Dev Mode handoff: the purchase and redeem value-exchange patterns, the two-currency model, the transparency-module set, the player-protection controls, the sequenced KYC sign-up, and the original illustrated character universe and its interface system.

Dashboard — Plain framing lifts comprehension and purchase

Plain framing lifts comprehension and purchase
Scope: Last 30 days · mobile-first pilot (~600 new players)
Guiding question: Did the value-clear framing help players understand and buy?

  Correctly identified what redeems for cash .. 34% → 82%
  Purchase-flow completion .................... 41% → 68%
  Cashout / currency questions in support ..... 38% → 15%

Key Insight: Showing the value plainly, inside the compliant structure,
let players understand what they were buying and complete the purchase.

Prototype / Test — Leading with the legally cautious wording was compliant and left players unable to understand the purchase; a value-clear, equally compliant framing fixed comprehension without raising harm signals

In this section: The experiment · What it taught

The sweepstakes model requires framing a purchase as buying a product with coins received free. The first build led with that wording at its most cautious ("Purchase music. Receive complimentary coins."), on the assumption that foregrounding the legal structure was the safe choice. It was A/B tested against a value-clear framing that keeps the same legal structure in Statsig across the pilot. Because the value-clear framing was designed to raise purchase completion, the experiment carried guardrail metrics to check that improved clarity was not simply pushing people into spending they would regret.

The failed variant. The legally cautious framing was compliant and tested as trustworthy in tone, but it obscured the value: only 34% of players could say what they would get to play or redeem, and purchase-flow completion sat at 41%. (This legal-cautious variant is the same condition as the standard-flow baseline in the comprehension test, which is why both read 34% / 41%.) Leading with the disclaimer optimized for caution and left players unable to understand what they were buying.

The cautious wording is compliant and unreadable
Scope: Statsig A/B · mobile-first pilot · ~300 players / arm
Guiding question: Which framing lets players understand and complete a purchase — without raising harm signals?

  Variant A — Legal-cautious wording first
    Comprehension (what redeems for cash) .. 34%
    Purchase-flow completion ............... 41%

  Variant B — Value-clear, same legal structure
    Comprehension (what redeems for cash) .. 82%
    Purchase-flow completion ............... 68%

  Guardrail metrics (Variant B vs A) — monitored, not optimized
    Chargeback / refund requests ........... no material increase
    Limit-setting / self-exclusion usage ... available and used, not suppressed
    Repeat-purchase within 1 hour .......... flat (no impulsive-spend spike)

Read with care: ~300 players per arm; comprehension and completion gaps are
large and consistent. The day-30 retention figure (in Outcomes) is directional.

Key Insight: Both variants were compliant, so the difference was legibility;
showing what the player gets and what redeems for cash is what let them buy —
and the guardrails showed clarity did not translate into impulsive spend.

What it taught. Compliance framing and comprehension are one design problem: a layout that satisfies the legal structure and shows the value plainly is what lets players act, while leading with cautionary wording leaves them unable to. Pairing the conversion metric with guardrails is what kept "clearer" from quietly meaning "pushier." The value-clear, compliant framing shipped.

Outcomes & reflections

In this section: Causal chain · Limitations · Regulatory context · Competitive context · Reflections

Causal chain (mobile-first pilot, ~600 new players)

The value-clear, compliant framing made the exchange legible, so new players could state what they were paying for and what redeems for cash (comprehension 34% → 82%), which lifted purchase-flow completion from 41% → 68%; the payout proof at first contact and the cashout, rewards, and billing history made redemption auditable, cutting cashout and currency questions in support from 38% → 15% of volume and getting more first-time players through a first redemption, which lifted day-30 retention from 43% → 61% (the 43% baseline from the legal-cautious cohort). Guardrail metrics over the same period showed no material rise in refunds or impulsive repeat purchases, so the comprehension gain did not come at the player's expense.

Metric Baseline Result What it compares
Comprehension (what redeems for cash) 34% (standard / legal-cautious) 82% (value-clear) real-world & rejected variant vs shipped
Purchase-flow completion 41% (standard / legal-cautious) 68% (value-clear) real-world & rejected variant vs shipped
Cashout / currency questions in support 38% (real-world) 15% (tool) real-world vs shipped tool
Day-30 retention (directional) 43% (legal-cautious cohort) 61% (value-clear cohort) rejected variant vs shipped
Guardrails (refunds, impulsive repeats) monitored no material increase safety check, not a win metric

Reading note: the standard sweepstakes flow and the legal-cautious A/B variant are the same condition, which is why comprehension and completion share one baseline (34% / 41%) across the research and the experiment.

Scale note: a sweepstakes model only works if players trust the redemption, so comprehension and payout proof are the gate the whole catalog sits behind.

Limitations (stated, because a portfolio claim is only as strong as what it concedes)

  • Pilot scale and horizon. ~600 new players, ~300 per A/B arm. Comprehension and completion effects are large and consistent; day-30 retention is directional.
  • Self-report in part. The "catch" perception and trust signals are attitudinal; the comprehension, completion, and support figures are behavioral and operational.
  • Guardrails are necessary, not sufficient. The monitored guardrails catch gross harm signals over a short window; they are not a full responsible-gambling audit, which a licensed framework would require.
  • The proxy loop is the contested mechanism. The on-platform proxy purchase is what makes the model work and is also the part regulators challenge; the design's clarity does not change its legal standing.

Regulatory context (named, because ignoring it would date the case)

The dual-currency sweepstakes model sits in a legal gray zone that, through 2025–2026, a wave of US states moved to close: California's AB 831 (effective January 2026) and bans or restrictions in states including Montana, Connecticut, New Jersey, New York, Indiana, Iowa, Oklahoma, and others, with many more considering legislation. Regulators' stated concern is that these platforms offer gambling-like products without the consumer protections required of licensed operators. This case is a comprehension-and-transparency redesign within that structure, completed while it was operative; it does not assert the model's legality, and the volatility above is exactly why the player-protection and spend-transparency work is the part that ages well — clarity and auditability are defensible regardless of where the structure lands.

Competitive context

Sweepstakes operators like Chumba, McLuck, and Stake.us compete largely on catalog size and promotional offers, with the dual-currency mechanics presented in much the same legally cautious way across the category. Splat's differentiation was not the catalog; it was the legibility of the value exchange, the payout-proof and audit trail surfaced from first contact, the on-platform proxy loop, and a mobile-first build in a desktop-assuming field — i.e., competing on comprehension and trust rather than on bonus size.

Reflections (transferable principles)

  • When a model carries a required legal framing, compliance and comprehension are one design problem: the layout that satisfies the legal structure while showing the value plainly is what lets users act, where cautionary-first wording leaves them unable to.
  • Trust in a new platform is built with proof a user can audit; recent real payouts, a visible cashout history, and traceable rewards carry more weight than any claim of legitimacy.
  • In a model under scrutiny, the honest version of "make it clearer" includes making spend visible and controls reachable, and pairing any conversion metric with guardrails — so legibility serves the player's informed choice and not only the purchase.
  • Bringing an indirect monetization loop on-platform, through an in-house proxy product, removes the friction and the off-site distrust of sending users elsewhere — while remaining the mechanism a designer should name honestly as the model's legal and ethical pressure point.