Real client work solving the moments that lose players: unclear first sessions, easy-to-ignore notifications, confusing missions, inconsistent terminology, and character voices that do not hold together.
These case studies feature anonymized client work. Product names, proprietary visuals, and confidential data have been removed, while the challenges, decisions, and contributions remain true to the work.
Players should not have to decode the game before they can enjoy it. These are the production problems I help teams solve:
Players quit in the first session before they understand what to do.
The English works, but it reads like it was translated, not written.
Placeholder text and unproofed error messages that quietly went live.
Copy that's been live since launch with nobody checking if it still works.
The Problem
The client had dozens of promotional events, each with a title translated individually into English. On paper, each translation was accurate. In practice, none of it worked — players scanning a promotions screen couldn't tell a VIP-exclusive offer from a weekend bonus from a deposit reward, because every title was worded like a one-off, even when the underlying event type repeated constantly.
How I Approached It
I looked at it the way a player actually would: fast, scanning, deciding in half a second whether something's worth tapping. Nobody reads a promo screen top to bottom — they scan for a pattern that tells them what kind of thing they're looking at. If the first few words don't instantly communicate the event type, the player's already lost interest.
So instead of translating title by title, I built a small, fixed set of category labels — VIP Only, Time-Limited, Exclusive Offer, and a few others — mapped to the type of event underneath the wording, not the literal source phrase. Every VIP-only offer read as VIP Only, every time, regardless of how creative the original title got. Once a player learns what a label means once, they never have to re-learn it.
Along the way I also caught something the client hadn't flagged: one recurring reward was tied to a payment method that only exists in one region. Kept as-is, it would've shipped to a Western audience referencing something they couldn't use. I flagged it, and we swapped it for something market-agnostic before it went live.
The Result
The result was a more consistent promotional taxonomy designed to reduce recognition effort for players. Because the labels were built around event type rather than one-off phrasing, the team also gained a reusable system for future promotions instead of solving the same naming problem every time.
Why it mattered: Notification copy across reward reminders, expiring items, and season-end warnings sounded equally urgent. When every message shouts, players lose the signal that tells them what actually deserves attention—and may begin ignoring all of it.
What I noticed: This was not a collection of isolated lines. It was a messaging system without an urgency hierarchy. Routine reminders, useful updates, and genuine deadlines needed different jobs, tones, and calls to action.
What I changed: I mapped tone to the real stakes of each trigger: playful for low-pressure reminders, direct for useful actions, and genuinely urgent only when time or rewards were actually at risk. I also audited the full set to reduce repeated hooks and phrasing.
What the team gained: A more intentional notification system that could preserve urgency instead of spending it everywhere. Following rollout, the client reported stronger engagement and longer sessions. I did not have access to the underlying analytics, so I’m intentionally not attaching an unsupported percentage.
Why it mattered: Mission text had been written like backend specifications. The conditions were technically present, but players had to stop, interpret the language, and work out what action would count—during a part of the game meant to create momentum.
What I noticed: Shorter copy alone would not fix the problem. Each line needed to answer three questions at a glance: What do I do? What counts? What do I get?
What I changed: I translated technical requirements into concise, action-led player language and checked every condition against the underlying mechanic. When a requirement was ambiguous, I sent it back to the client instead of making the wording sound confident and hoping it was correct.
What the team gained: Clearer mission instructions plus an extra accuracy check before implementation. One question surfaced a win condition the client’s own team had not fully specified, improving the final instruction and giving QA a clearer condition to verify.
Why it mattered: An existing tutorial needed a seasonal narrator with a completely different personality, but every instructional beat still had to teach the same mechanic in the same sequence. Too little personality would make the character feel cosmetic; too much could bury the action.
What I noticed: Character voice and instructional clarity were not competing goals—they needed separate boundaries. The mechanic had to remain fixed while rhythm, vocabulary, attitude, and emotional framing could change.
What I changed: I preserved the functional meaning of each step, then wrote multiple voice-led options to test how far the persona could stretch without obscuring what the player needed to do next.
What the team gained: A repeatable method for adapting tutorials across character voices without changing the underlying game logic—a useful foundation for future seasonal variants and cross-functional review.
Client need: A narrative game project needed characters who felt grounded in recognizable human experience—not assembled from familiar archetypes or reduced to identity-based shorthand.
My role: I developed character profiles using interviews, contextual research, and close attention to how real people describe their motivations, contradictions, relationships, and environments. I translated that research into practical voice, behaviour, and backstory guidance the wider production team could use.
Collaboration: The profiles were built to support more than dialogue. I worked across development and audio considerations so character choices could remain coherent through gameplay, pacing, sound, and music—not only on the page.
Outcome: The team gained a shared character foundation designed to keep writing and production decisions emotionally consistent, culturally respectful, and resistant to stereotype. Project details and original materials remain confidential.
Tell me what is confusing players, weakening the experience, or slowing down production. We can talk through it and see whether my approach fits your team.
Let’s See If We’re a Fit