Bird Island
The Long Night
This production encompasses the full game development lifecycle, from initial concept development through to delivery of a functional alpha build. The project includes all core aspects of game creation:
Concept
The Production Plan
The Long Night was built on a 6-month development cycle with 2-week sprints, a rhythm that balanced momentum with adaptability. As producer and project manager, I orchestrated this cadence, tracking progress in Jira, leading sprint planning, and ensuring no discipline lagged behind. But it wasn’t just about deadlines; it was about syncing creative risks with practical constraints.
Each sprint began with a clear goal: a combat prototype, NPC dialogue systems, or the darkness mechanic’s visual polish. My role was part strategist, part firefighter—reviewing daily stand-ups, adjusting task priorities when art dependencies blocked design, or stepping in to test builds mid-sprint. For example, when the darkness-closing system bugged during a playtest, I worked directly with engineering to reprioritise fixes without derailing the next sprint’s content goals.
Completion rates weren’t just metrics; they were a team health indicator. If a sprint underdelivered, we dissected why—was it scope creep in level design? Unforeseen tech debt?—then adjusted. By the final months, this rigour let us lock the prototype’s core pillars (combat, exploration, choice-driven horror) while leaving room for polish.
Production Goals
- Create the prototype: Ensure that the game’s flow feels cohesive, with clear hooks that extend to the full game.
- Iterate on Combat Clarity: Refine UI icons and tooltips to communicate stats/abilities intuitively.
- Optimise Narrative Pacing: Balance world-building dialogue with gameplay, avoiding excessive inner monologue.
Development
The Team
The team a hybrid of artists, programmers, designers, and sound specialists,all fluent in each other’s languages. Early on, we rejected silos. Concept art wasn’t tossed over the fence to modeling, artists sat with designers to ensure environments reinforced gameplay. For instance, the village’s claustrophobic sightlines were born from a designer’s need for tension and an artist’s mood boards of 18th-century decay.
Tech bridged the gaps. Engineers built tools for designers to tweak the darkness’s spread rate without code, while audio implemented dynamic stingers that reacted to player proximity—no scripting required. My role? To facilitate these collisions. Weekly “show-and-tells” forced cross-discipline feedback: a programmer might flag that a designer’s quest logic broke the save system, or an artist might realize their textures murdered performance.
This wasn’t just collaboration; it was shared ownership. When the sound designer prototyped “whispers” that intensified as darkness closed in, it became a core horror pillar—proof that the best ideas emerged when specialities blurred.
My planning system revolved around two parallel tracks:
Tech’s Feature List – A living document of engineering milestones, turn-based combat, dialogue branching, the darkness timer that are each broken into atomic tasks. For combat alone, this meant as an example:
- Grid movement
- Enemy AI (un-killable, pressure-based)
- Tile effects (darkness debuffs, light buffs)
Tasks were assigned by dependency; e.g., AI needed movement functionality first.
Content’s Avalanche
Meanwhile, art, design, and audio faced a mountain of bespoke tasks such as:
- 8-10 explorable POIs per outskirts zone
- 2-3 combat maps with unique tile effects
- 500+ lines of NPC barks (each requiring recording, implementation)
I worked with leads to batch content—e.g., all “grave site” assets due by Sprint 5—and flagged bottlenecks early (like VO delays). The key was visibility: a shared board showed how environment art lagged behind level design, so we could reallocate resources.
Team planning
My approach to planning revolved around two pillars: a feature-driven tech list and a content pipeline for art, design, and audio. For tech, we broke down systems like the darkness mechanic, combat triggers, and dialogue into clear tasks. Meanwhile, the rest of the team tackled a mountain of content—NPC barks, environmental storytelling, quest design—all prioritized to match our prototype’s scope. Flexibility was key; we adjusted as playtests revealed what resonated (and what didn’t).
Release
Final production
After 6+ months, the backlog ended into a 20-30 minute horror vignette. The final sprint was ruthless, added changed to various content and fixed major QA related issues, simplify and cuttet what was not needed, this looked something like this:
Milestone 1: Core Framework (Month 1-2)
We established the technical foundation with combat systems and a working darkness mechanic. The greybox village proved our level design could create tension through constrained spaces. Early playtests confirmed our core loops worked, though they lacked atmosphere.
Milestone 2: First Horror Playthrough (Month 3-4)
The Woodsman quest came to life with branching choices and grotesque creature designs. Combat gained teeth with the unkillable Nemesis enemy, while dynamic sound made the darkness feel alive. Players finally felt the dread we wanted to deliver.
Milestone 3: Full Demo Content (Month 5)
Every NPC found their voice through barks and quest dialogue. We locked all explorable areas and balanced combat encounters. Playtesters reacted strongly to corruption consequences, though pacing needed tweaks.
Milestone 4: Polished Experience (Month 6)
Visual effects and sound design upgrades transformed rough systems into palpable horror. Hardcoded story beats ensured shocking moments landed perfectly. The final prototype delivered 20 minutes of tense, choice-driven gameplay that left players unsettled.
