(CASE 02 · IGAMING · 15 MONTHS)
Fifteen months on a real casino’s money flows.
BLUFF was a live, crypto-native casino and sportsbook backed by a publicly reported $21M raise. I was on it for fifteen months as senior product designer: onboarding, money flows, twelve in-house games, VIP, notifications, and the post-ship quality work in between. The raise is public record, every other number here is my own count.
- ROLE
- Senior Product Designer
- TEAM
- Superduper, with Tom Jones as design lead
- TIMELINE
- 2025 to 2026
- STATUS
- Shipped, the product has since pivoted
50,000+ SIGNUPS · 15% REG-TO-PLAY · $21M RAISE
CHAPTER 01
The launch
Fifty thousand signups before the product existed.
The pre-launch landing page had six jobs at once: sell urgency without spamming, explain the Blink Points loop without documentation, make Founding Member scarcity feel real, pitch three reward types, drive referrals through quests, and talk to migrating VIP players separately from newcomers. Two audiences, one page, layered so each read finds its own reasons. It drove 50,000+ signups before the platform launched, 15% registration-to-play, and 12% conversion on referred users. The lesson I kept: softness on a financial page reads as confidence.
50,000+
SIGNUPS BEFORE LAUNCH
15%
REGISTRATION TO PLAY
12%
CONVERSION ON REFERRED USERS
FIG. 02 · THE PRE-LAUNCH PAGE, IN FULL — SCROLL THE FRAME
CHAPTER 02
The money
Every money surface is a trust surface.
Deposits, withdrawals, buying crypto, bet history, transaction history. On a casino these are not account plumbing, they are the product’s argument that it can be trusted with a balance. I designed them calm on purpose: clear amounts, explicit states, history a player can audit without asking support. When you’re designing for a casino, trust is conversion, and the surfaces that look like pure craft are conversion surfaces too.
FIG. 03 · DEPOSIT
Amounts, fees and the state of the network stated flat. A player about to move money gets no spectacle here, just certainty about what happens next.
FIG. 04 · WITHDRAW
The withdrawal path mirrors the deposit grammar exactly, so taking money out never feels different from putting it in. That symmetry is deliberate.
CHAPTER 03
The failure states
Twenty-four ways to lose a player.
Mapping every way signup and payments could fail, I stopped counting at 24 conditions across email, phone and OTP. The constraint that made it matter: real money already debited, a regulated product, and a support team logging every one of these calls as a payments bug. An error message here is not copy, it is the difference between a retry and a chargeback.
I split the failure states by one question: can the player fix this themselves? Three states came out: retry now, retry later, stop and contact us. Persistence followed the same logic, errors dismiss on retry unless the condition persists, like a country restriction, and every persistent error offers an escape hatch, SMS failing over to email.
THROWN AWAY
The obvious answer was a message per error code, all 24. I killed it: most codes came from the provider and could not be mapped to an action the player could take, so it would have produced 24 sentences with the same meaning.
24 → 3
Nine screens shared one message before. Three states after. Two weeks, copy and an existing state, no new engineering.
FIG. 06 · EMAIL SIGNUP, ERROR RECOVERY
The message names what happened and the one action that fixes it. Nothing else earns a place on the screen while money or access is in question.
FIG. 07 · PHONE SIGNUP
Built as a native mobile experience first, then adapted to desktop, because that is the order players actually arrive in.
CHAPTER 04
The games
Twelve games that read as one product.
Roulette, Baccarat, Hi-Lo, Mines, Wheel, Keno, Limbo, Plinko, Dice, Blackjack, Crash, and Flappy Fortune. The design problem was never one game, it was coherence across twelve products that each wanted a personality. The pattern that held it together is a shared profit-and-multiplier display that recurs in every game’s mid-round UI, so a player who learns one game has learned the grammar of all of them.
ROULETTE · BACCARAT · HI-LO · MINES · WHEEL · KENO · LIMBO · PLINKO · DICE · BLACKJACK · CRASH · FLAPPY FORTUNE
FIG. 08 · PLINKO, MINES, BLACKJACK, DICE — ONE GRAMMAR, FOUR PERSONALITIES
One game existed nowhere else.
Flappy Fortune is BLUFF’s exclusive: a spatial game where the player chooses both whether to continue and which lane to fly through, multiplier climbing until the loss. The decision that made it legible: incoming pillars draw as dashed outlines, so a future state never reads as a wall. And the change that made it feel like a game rather than a sprite test came after the first build, anchoring the bird and scrolling the environment past it.
FIG. 09 · INCOMING, AS A DASHED PROMISE
Dashed outlines mark pillars that have not arrived yet, so the player never mistakes a future state for a wall that is already solid.
FIG. 10 · MID-GAME
During a run the interface hides its noise: current profit, profit on the next tile, and the cashout action. Everything else waits.
CHAPTER 05
The quiet surfaces
Calm is a feature in a casino.
Notifications were built around the Smartico API’s constraints: four categories, promotions, account activity, gameplay events and VIP updates, with single-line copy, soft read-unread contrast and large tap targets. The register is deliberately dismissible rather than spectacle. The same restraint carried VIP: status is information a player checks, not a firework the product sets off.
FIG. 11 · NOTIFICATIONS
Category labels, one line of copy, a soft read-unread contrast. A panel a player can clear in seconds and trust not to shout.
FIG. 12 · VIP
Progress and status as plain information. The value is in what it says, not in how loudly it says it.
What survived was the reasoning.
Underneath the screens sat system work: BLUFF-specific components on the Superduper design system template, bet input cards, multi-state game control panels, notification cards, VIP status components. Every component shipped with its edge states designed, empty, loading, partial data, error, mid-action, success, and a post-ship QA tracker kept quality under review for three months after launch.
In 2026 BLUFF pivoted from casino and sportsbook to pack opening. The content changed, the interface system still runs the product. Which is the point I’d defend from all fifteen months: what survives the next redesign is never the visual choices, those always change and they should. What survives is the reasoning, why a state behaves the way it does, why an error persists instead of dismissing.
STATED FLAT
Retry later is still not a real instruction, and I don’t know what the right one is. And nothing above cites the cashier’s own numbers, the landing page had analytics I can publish, the money flows didn’t. What’s here is checkable instead.
BLUFF
FIFTEEN MONTHS · THE MONEY FLOWS
