Chapter 34: The Game Industry — Studios, Roles, Pipelines, and How Games Actually Get Made

There is a particular look on the face of a student designer the first time they learn what a "Vertical Slice" actually requires. It is the look of someone who has just been told that the magic trick they admired their whole life is not magic — it is six people in a basement, a spreadsheet, and a publisher demanding playable proof that the climbing animation will not look broken at scale. The game they love did not get made the way they thought it did. No game does.

This chapter is the practitioner's tour of the apparatus. We are going to walk you through the buildings — first-party studios, third-party studios, publisher-owned studios, mid-tier outfits, indies, mobile shops, the invisible work-for-hire houses. We are going to walk you through the roles inside those buildings — designers, programmers, artists, animators, audio leads, producers, QA, community, marketing, legal, localization. We are going to walk you through the pipeline that converts a concept document into a product on a storefront — concept, pitch, greenlight, pre-production, production, alpha, beta, certification, launch, and the long tail of post-launch operations. We are going to look at money: how games get funded, how revenue is split with platforms, what an indie game actually earns, and what the median Steam release brings home (the answer is sobering). And we are going to talk about the human cost — careers, layoffs, burnout, unionization, and the question of whether a games career is worth having at all.

You should care about all of this even if you intend to spend your career as a one-person indie. The structures are the medium. They shape what gets made. The Last of Us Part II could only have been made inside Naughty Dog's particular structure of authority and craft, with Sony's particular willingness to fund a dark, slow, single-player project that would not move 200 million copies. Stardew Valley could only have been made by a single person with no employer, no deadline, and an indifference to whether it sold at all. The studio is the room the game is dreamed in. If you do not understand the rooms, you will not understand why some games are possible and others are not.

I will warn you up front: this is the chapter where naïveté goes to die. If you arrived believing the games industry is a meritocracy where the best ideas win, you will leave understanding that publisher relationships, capital structure, platform fees, and luck are at least as important as craft. If you arrived believing indie success is a 24-month sprint to a Stardew-shaped exit, you will leave understanding what survivorship bias looks like in practice. This is not pessimism. It is calibration. You cannot navigate a system you do not see clearly. Let us look at it clearly.

The Industry Is Not One Industry

The first thing to internalize is that "the game industry" is a misleading singular. It is a dozen industries in a trench coat, sharing only the noun "game." The economics, schedules, hiring practices, design constraints, and ethics differ so wildly between segments that a senior designer at King (mobile match-three) and a senior designer at FromSoftware (action RPG) have less in common professionally than a baker and a chemist. Both work with flour, in some sense. They are not in the same trade.

AAA console/PC. Budgets from $50 million to over $300 million. Teams of 200 to 2,000+ across multiple studios. Cycles of five to seven years (sometimes more — Red Dead Redemption 2 was seven, GTA VI over a decade by ship). Risk tolerance: low. Pitches are evaluated against a model of recouping a huge advance through unit sales, season passes, microtransactions, or all three. The role of "designer" here is highly specialized — combat designer, encounter designer, mission designer, systems designer, narrative designer. You will rarely touch the whole game. You will become an expert in one slice.

AA / mid-budget. Budgets from $5 million to $30 million. Teams of 30 to 150. Cycles of three to five years. Remedy's Control. Arkane's Dishonored. Obsidian's The Outer Worlds. Asobo's A Plague Tale. The endangered middle of the industry — squeezed from above by AAA marketing budgets and from below by indie discoverability and price expectations. Many AA studios survive only because a publisher (often Microsoft or Sony first-party) absorbs them.

Indie. Budgets from your savings to maybe $2 million. Teams of one to twelve. Cycles of one to three years (or longer — Stardew Valley was four, Caves of Qud has been in development since 2007). The roles collapse: an indie "lead designer" is also the producer, the writer, the QA tester, and often the marketer. Risk profile: catastrophic for the individual, low for any single project (no studio to fold, just a hard year and a different job).

Mobile free-to-play. Budgets from $200,000 (hyper-casual) to $50 million+ (RPG gachas like Genshin Impact, Honkai Star Rail). Teams of 5 (hyper-casual) to 400+ (live-ops gacha). Cycles: launch in 8 to 18 months, then operate forever. The core role is "LiveOps Designer" — a job that did not exist as a title in 2010. The economics are about LTV (lifetime value per player), CAC (cost to acquire customer), and ARPDAU (average revenue per daily active user). If you do not enjoy spreadsheets, you will hate this segment.

Live service (console/PC). Destiny 2, Fortnite, League of Legends, Final Fantasy XIV, Path of Exile, Warframe. Budgets indistinguishable from AAA, but the cycle never ends — you are not shipping a game, you are running a perpetual operation. A "season" of Fortnite requires the content output of an entire small game, every twelve weeks. Burnout in this segment is structural.

Cloud gaming. Xbox Cloud Gaming, GeForce Now, PS Plus Premium streaming. Not really a development segment yet — more a distribution layer that affects how games are designed (input lag tolerance, session length, save-state assumptions). Stadia's collapse in 2022 cooled the gold rush, but the layer persists.

VR/AR. Meta Quest drives the market. Budgets typically $1 million to $20 million. The designer's job is fundamentally different — you are designing for a body in a room, not a player in a chair. Locomotion, comfort, scale, IK puzzles. Half the questions you ask in flat-screen design do not apply; new ones replace them. Half-Life: Alyx is the high-water mark; very few studios can match it.

Browser/HTML5. Long thought dead, alive again because of itch.io and instant-play platforms. Tiny budgets, tiny teams, often one weekend (game jam scale). The skill ceiling is in the design, not the tech.

Esports. Adjacent industry: tournament organizers (ESL, BLAST), team organizations, broadcasters, the publishers who run their own leagues (Riot, Valve, Blizzard). The designer's role here is "competitive integrity" — patch design, balance, anti-cheat, spectator UX.

Creator-platform. Roblox, Fortnite UEFN, Core, Dreams (RIP), Manticore. You are not making a game; you are making a game inside someone else's game, monetized through the platform's currency and revenue-share rules. Roblox's top creators earn millions per year; the median creator earns under $100. The economics resemble YouTube more than they resemble traditional games. The platform owns the audience, the payment rails, the discovery algorithm, and ultimately your livelihood. Read Chapter 33 again on dependency before you build a career here.

Each of these segments has a different ethical landscape, too. The dark patterns Chapter 33 surveyed are concentrated in mobile F2P and live service; AAA single-player and indie carry different temptations (crunch, false advertising, scope-driven layoffs). The career paths diverge sharply. A senior designer in mobile may have shipped twelve games; a senior designer in AAA may have shipped two. They are not interchangeable.

When someone asks "should I get into the games industry," your first counter-question is which one. The answer changes everything that follows.

Studio Types

The buildings — and the legal and financial structures inside them — shape what can be made. Five rough categories cover the majority.

First-party studios are owned by a platform holder. Sony Interactive Entertainment's PlayStation Studios (Naughty Dog, Insomniac, Sucker Punch, Santa Monica, Guerrilla, Bend, Media Molecule, Bungie). Nintendo Entertainment Planning & Development (EPD), with internal divisions led by long-tenured leads (Aonuma, Koizumi, Sakurai). Microsoft Xbox Game Studios (343 Industries before its 2024 reshuffle, The Coalition, Turn 10, Rare, Ninja Theory, Obsidian, inXile, Double Fine, Compulsion, Mojang, Bethesda after the ZeniMax acquisition, Activision Blizzard after that acquisition). First-party studios make the games that sell their parent's hardware or service. They generally have more creative latitude than third-party studios — Sony lets Naughty Dog spend seven years on a single-player narrative game because The Last of Us sells PlayStations — but they are not free. They serve the platform's strategy. Nintendo will not make a Mario game on Xbox; Naughty Dog will not make Uncharted on Switch. The trade is: stable funding, long horizons, deep platform integration, in exchange for the platform owns the IP and the studio.

Third-party AAA studios are independent or publicly traded but not owned by a platform holder. Rockstar Games (under Take-Two but operationally autonomous), FromSoftware (under Kadokawa, who in turn has Sony as a major shareholder), Capcom, Square Enix, Bandai Namco, Sega. They ship across PlayStation, Xbox, PC, and sometimes Switch. The advantage: bigger addressable market. The disadvantage: no first-party support cushion, so a flop hurts more.

Publisher-owned studios are the bulk of the industry. Ubisoft owns Massive (The Division, Avatar), Montreal (Assassin's Creed), Toronto (Far Cry), Quebec, Bordeaux, Annecy, Pune, and a dozen more. Electronic Arts owns DICE (Battlefield), BioWare, Respawn (Apex, Titanfall, Star Wars Jedi), Maxis, EA Sports' internal studios. Activision Blizzard owns Infinity Ward, Treyarch, Sledgehammer, Beenox, High Moon, Toys for Bob, Raven, Vicarious Visions (now folded into Blizzard), Blizzard, King — now all under Microsoft after the 2023 acquisition. Take-Two owns Rockstar, 2K, Firaxis, Hangar 13, Cloud Chamber, Cat Daddy, Zynga (after the 2022 acquisition). The pattern: publishers buy studios for IP, talent, and pipeline capacity. Studios chafe under publishers when corporate priorities override creative ones — see BioWare under EA in our case study.

Mid-tier / AA studios are the endangered middle. Remedy Entertainment (Helsinki), Arkane (Lyon and Austin, both Microsoft-owned now), Obsidian (Microsoft-owned), Larian (until Baldur's Gate 3 made them effectively AAA), Supermassive (Until Dawn, The Quarry), Asobo (Microsoft Flight Simulator, A Plague Tale), Housemarque (Sony-owned after Returnal). They make games that can punch above the indie ceiling without the AAA risk profile, but the segment has been thinning for fifteen years as production budgets balloon and publishers consolidate. Several of the studios in this list survive because a first-party publisher has decided the segment is strategically valuable — Microsoft especially has been shoring up the AA layer.

Indie studios. Supergiant Games (Bastion, Transistor, Pyre, Hades, Hades II) — 15-20 people, self-published, in Oakland. Mossmouth (Spelunky, Spelunky 2) — Derek Yu and a small rotating team, distributed, contract-based. Daniel Mullins Games — one designer plus contract collaborators (Pony Island, The Hex, Inscryption). Motion Twin (Dead Cells) — a flat-hierarchy worker cooperative in Bordeaux. Maddy Makes Games (Celeste) — small Vancouver team. House House (Untitled Goose Game) — Melbourne, four people. ConcernedApe (Stardew Valley) — Eric Barone, alone. The defining trait is creative control. The defining risk is that the next project might not work, and there is no parent company to absorb the cost.

Mobile studios. King (Candy Crush, owned by Activision/Microsoft). Supercell (Clash of Clans, Clash Royale, Brawl Stars) — famously small cell-based teams in Helsinki, owned majority by Tencent. miHoYo / HoYoverse (Genshin Impact, Honkai Star Rail, Zenless Zone Zero) — Shanghai, the new juggernaut. Niantic (Pokémon Go, Ingress). Playrix (Gardenscapes, Homescapes, Township). Scopely, Jam City, Zynga. The work culture and design disciplines here look more like web/SaaS than like traditional game development. Live-ops cadence rules everything.

Work-for-hire and outsourcing. This is the invisible industry. Virtuos (Singapore, with offices across Asia and Europe — they did massive uncredited work on Elden Ring, Horizon Forbidden West, dozens of other AAA titles). Iron Galaxy (Chicago — co-developed games in the Killer Instinct and Diablo series, and is the studio behind many "definitive edition" ports). Nixxes (Netherlands, Sony-owned now — they do PC ports of Spider-Man, Horizon, Ratchet & Clank). Sperasoft, Streamline, Keywords Studios. These studios build levels, ports, animations, and sometimes whole games on contract. They are the reason a 2,000-person AAA project actually has 4,000 to 6,000 contributors total. The credits on a modern AAA title routinely run forty minutes; most of those names are the invisible industry. Their working conditions, pay, and stability vary hugely by region — and the practice of leaving outsourced studios uncredited for major contributions is one of the long-running ethical scandals of the field.

When you think about your career, you are choosing which of these buildings you want to live in. The buildings are not interchangeable.

The Roles

A modern game is a coordinated act of about a dozen disciplines. On a small indie team, one person wears five hats. On an AAA team, each hat is its own department with sub-specialties. Here is the rough taxonomy.

Designer. Decides what the player experiences. Sub-roles include:

• Game designer (general, often used as the entry-level title).
• Level designer — builds and tunes the spaces the player moves through.
• Systems designer — owns mechanics interactions, balance spreadsheets, the math of progression.
• Narrative designer — writes the story integrated with gameplay (different from a writer; see Ch 20).
• UX designer — owns menus, HUD, accessibility, onboarding.
• Combat designer — specializes in encounter pacing, enemy variety, damage tuning, hitboxes.
• AI designer — writes the behavior trees, FSMs, utility curves; often half-designer, half-programmer (see Ch 27).
• Economy designer — owns currencies, drop rates, monetization (in F2P, this is one of the most senior roles).

Programmer / engineer. Writes the code. Sub-roles:

• Gameplay programmer — implements features in close collaboration with designers.
• Engine programmer — works on the underlying tech (rendering, physics, memory).
• Tools programmer — builds the editors and pipelines other devs use; often the highest-leverage role on a team.
• Networking programmer — owns multiplayer, replication, anti-cheat (see Ch 28).
• Graphics programmer — shaders, lighting, post-processing.
• Audio programmer — DSP, mixing, occlusion, integration with FMOD/Wwise.
• AI programmer — pathfinding, navmesh, decision systems.
• Build / DevOps engineer — manages CI, daily builds, automated testing.

Artist. Visual design and execution.

• Concept artist — paints the early vision.
• 3D artist — models, often subdivided into character and environment specialists.
• 2D artist — sprites, textures, UI graphics.
• Character artist — hero and NPC modeling and texturing.
• Environment artist — terrain, props, set dressing.
• VFX artist — particles, shaders, gameplay-feedback effects (see Ch 8).
• Technical artist — the bridge between art and tech; rigs, shaders, optimization.
• UI artist — HUDs, menus, fonts, iconography (see Ch 29).

Animator. Movement and acting.

• Character animator — keyframed and motion-captured character motion.
• Rigger — builds the skeletons animators work with.
• Cinematic animator — cutscene work (see Ch 23).
• Tech animator — the bridge between animation and tech; runtime animation systems.

Audio. The half of the experience players notice last and miss first (Ch 30).

• Sound designer — SFX, ambient, foley.
• Composer — score.
• VO director — direction of voice actors in the booth.
• Audio implementer / technical sound designer — wires the audio into the game in Wwise/FMOD.

Producer. Schedules, budgets, removes blockers.

• Project producer — owns the schedule and the team's velocity.
• Associate producer — junior version, often runs a single discipline's planning.
• Executive producer — owns the project's strategic direction; in some studios this is a VP role.
• Live producer — runs an operating live game, distinct from a project producer.

QA (Quality Assurance). Tests the game. Often the entry door to the industry.

• Functional QA — finds bugs.
• Compliance QA — ensures the game passes platform certification (Sony's TRC, Microsoft's TCR, Nintendo's lotcheck).
• Compatibility QA — tests across hardware variants (a hellish job in PC and Android).

Community / PR / Marketing. Owns the relationship between the game and its audience.

• Community manager — runs Discord, social channels, forums.
• PR — runs press relationships and review embargoes.
• Brand / marketing manager — owns campaigns, trailers, store pages (see Ch 38).

Localization. Translates and culturalizes the game for non-source-language regions. Larger than you think — a major AAA game is localized into 15-25 languages, each requiring its own VO and QA pass.

Legal and business affairs. Negotiates contracts with platforms, licensors, publishers, voice actors, music rights holders. Reviews terms of service. Increasingly important as games incorporate user-generated content and AI tools.

A small indie of three people might have one designer-coder, one artist-animator, one audio-and-everything-else. A 2,000-person AAA team has every role above plus internal sub-specializations — a single environment art team can have separate leads for foliage, architecture, props, and cinematic set dressing. As you grow more senior, you specialize and then re-generalize: a senior designer often has a deep specialty plus enough fluency across disciplines to lead.

Roles That Don't Exist Yet

The taxonomy above looks settled. It is not. Half the roles on it are inventions of the last fifteen years. The job titles at a 2026 studio would have confused a hiring manager in 2008.

LiveOps Designer did not exist as a title in 2010. Genshin Impact and Fortnite and Apex Legends invented the discipline by needing it.

UX Researcher — embedded researcher who runs structured player studies, eye-tracking sessions, biometric playtests — was rare outside Microsoft Research and a few Sony studios until around 2015. Now it is standard practice in any AAA outfit serious about onboarding.

Accessibility Designer — a dedicated lead for the accessibility feature suite — barely existed before The Last of Us Part II's 2020 release made the role visible. Now any first-party studio takes it seriously.

Inclusive Design Specialist — focused on representation, cultural review, sensitivity reading — emerged in the late 2010s as a result of repeated public scandals around tone-deaf content.

AI Ethicist / Generative AI Lead — emerging now, in 2025-2026, as studios figure out how to use generative tools for concept art, animation cleanup, asset generation, and procedural content without violating IP, displacing workers, or angering audiences. By 2030 this will be a standard role.

ML Engineer — for studios using neural-net upscaling (DLSS, FSR), animation cleanup, or procedural content generation. Adjacent to engine programmer but distinct.

Game Director / Creative Director Coach — a few large studios now have senior advisors whose job is to mentor first-time directors through their first major project. The realization that "creative director" is a specific learnable skill, not a promotion you fall into, is recent.

The lesson is not just historical. The lesson is that your career will include roles that do not exist yet. The student designer who graduates in 2027 will, in 2035, hold a title that is not on this page. Build the underlying skills — design thinking, systems literacy, communication, taste — and the role names will sort themselves out.

The Project Pipeline

A modern game moves through roughly nine stages from idea to player. The names vary by studio, but the shape is consistent.

1. Concept. A pitch, a one-page document, a prototype. The team or the lead designer is convinced an idea is worth building. This may be one person on a weekend or a dedicated incubation team for months. Output: a vision statement, design pillars (Ch 16, Ch 24), reference games, target audience, rough scope estimate.

2. Pitch. Convincing someone with money to fund the next stage. Internal pitch (to studio leadership) or external pitch (to a publisher). Output: greenlight or rejection. We will spend more time on this in the funding section below.

3. Greenlight. The project is officially funded for pre-production. A team is assembled. A budget is allocated. A high-level milestone schedule is set. In many publisher arrangements, greenlight comes with kill clauses — the publisher can cancel at the next gate.

4. Pre-production. The most important stage and the one most often underinvested. Goal: prove the game is worth making and figure out how to make it. Outputs: a vertical slice (one polished, representative slice of gameplay), technical prototypes for risky systems, the design pillars hardened into mechanical detail, the production plan for the next stage. Length: 6-12 months for AA, 12-24+ months for AAA.

5. Production. Build the actual game. Content creation at scale. Most of the team's time is here. Length: 18-48 months for a typical AAA project. The longest, most expensive phase.

6. Alpha. Feature-complete. All systems exist. The game can be played end-to-end, but is unbalanced, ugly in places, and full of bugs. From here, no new features are added; the team transitions to polish and bug-fixing.

7. Beta. Content-complete and feature-locked. All assets are in place, all levels exist, all cutscenes are in. The work shifts to bug-fixing, balance, performance optimization, and localization QA.

8. Certification (cert). Platform holders (Sony, Microsoft, Nintendo, Apple, Google, Steam) review the build against their technical requirements (TRC, TCR, lotcheck). Failures must be fixed and resubmitted. Failing cert delays release. A second-cert failure can blow the launch window entirely.

9. Launch. The game ships. The team takes a long weekend, then comes back to fix the day-one patch, monitor reviews, monitor sales, and start planning the post-launch roadmap.

10. Post-launch / live operations. Patches, DLC, expansions, seasons, community management. For a live-service game, this stage never ends and is structurally more important than launch itself.

Typical timelines by segment:

• Indie (1-12 person team): 1-3 years from concept to ship. Sometimes much longer when a single person is building everything.
• AA (30-150 people): 3-5 years.
• AAA (200-2,000+): 5-7 years, sometimes longer. Red Dead Redemption 2 was seven; GTA VI over a decade. Star Citizen is, as of writing, fourteen years and counting.
• Live service: launch in 2-3 years; operate for 5-15+.
• Mobile F2P: 8-18 months to soft-launch; operate for years.

What can go wrong at each stage is its own catalog. Concept stage: the idea is not actually a game, it is a pitch. Pre-production: the team rushes through and "starts production" before knowing what they are making (this is the Anthem and Cyberpunk 2077 failure mode). Production: scope creep eats the schedule. Alpha: features that should have been cut at pre-pro are still there, half-built. Beta: localization runs late because nobody scheduled the VO recording sessions. Cert: an edge case crashes the platform's automated test suite. Launch: a server bug nobody anticipated takes the live game offline for the first 36 hours.

Chapter 37 will go deeper into the production craft — sprints, kanban boards, content reviews, milestone gates. For now, the high-level shape: every project moves through these stages, and most of them fail or struggle for predictable reasons that earlier stages were supposed to prevent.

Pre-Production Is Where Games Succeed or Fail

If you take one piece of practical wisdom from this chapter, take this: the quality of a finished game is determined more by the discipline of pre-production than by the talent of the production team. A great team in production cannot rescue a project that did not figure out what it was during pre-pro. This is the single most under-appreciated truth in the industry, and it is the single most common reason high-profile projects ship in trouble.

What does great pre-production look like? It looks like a vertical slice — a small, end-to-end, fully polished slab of the game that proves the experience works. Not a tech demo. Not a sizzle reel. A slice the player could actually play, that contains a representative sample of every system the full game will need: combat, exploration, dialogue, inventory, save, audio, UI. If the slice is not fun, the full game will not be fun. If the slice is fun, the bet is reasonable.

It looks like risk-burning prototypes — small, focused experiments that prove the hardest technical or design problems are solvable. If your game depends on procedural cave generation, you build the cave generator in pre-pro and you prove it works at scale. If your game depends on enemies that climb arbitrary geometry, you prototype the climb before you commit to the design. The expensive version of finding out a system does not work is finding out in production, when twenty levels have already been built around it. The cheap version is finding out in pre-pro, when there is still time to redesign or cut.

It looks like design pillars with veto power. Blizzard's pre-production culture, in its strongest era, gave the design pillars actual authority — if a feature did not serve the pillars, it was cut, regardless of how much someone loved it. The pillars were not aspirational marketing copy; they were the gate every feature had to pass. A pillar is useless if it does not have the power to kill. (We covered this in Ch 16 and Ch 24, and will return to it in Ch 37.)

It looks like honest scope assessment — comparing the production plan against the team's capacity, the schedule, and a realistic estimate of how much content gets built per month per discipline. The production plan is where pre-pro lands. If pre-pro produces a vertical slice but no plan for getting from slice to full game, pre-pro has only done half its job.

What does failed pre-production look like? Look at Anthem (BioWare, 2019). Jason Schreier's Kotaku reporting, later expanded in Press Reset, documented years of pre-production paralysis. The team could not agree on what Anthem was — flying-shooter, looter-shooter, world-of-something? — and the question was never resolved before production began. The leadership kept reversing course. The Frostbite engine, designed for Battlefield, had no native support for the systems Anthem needed (saves, inventory, RPG elements). When production finally started, the team had eighteen months to build a game whose design had not been settled. The vertical slice was not built until late. Anthem shipped, sold poorly, was abandoned in 2021. The post-mortem lesson was unanimous: the project's death was decided in pre-production, not in production.

Cyberpunk 2077 is a more complex story — CD Projekt Red's launch disaster was overdetermined, with engine problems, last-gen console targets, scope creep, and crunch all contributing. But the through-line, again, was pre-production discipline. Reporting from Bloomberg and elsewhere indicated that the game's core systems (cyberware, AI, vehicle handling) were still being prototyped years into what was nominally "production." When you are still prototyping in production, you are still in pre-production — you just lied about it.

The teams that get this right tend to be quiet about it. FromSoftware's pre-production for Elden Ring reportedly involved years of small open-world experimentation built on top of the Dark Souls tech, with George R.R. Martin's worldbuilding done early. Nintendo EPD treats pre-production with religious seriousness — Breath of the Wild spent over a year in a "2D Zelda" prototype phase that Aonuma's team used to validate the chemistry-of-systems design before committing to 3D. The cost of long pre-production is real (the team is being paid while not producing shippable content), but it is the cheapest insurance any project can buy.

The temptation to short-change pre-production is structural. Pre-pro looks like nothing is happening. There are no shippable assets, no impressive trailers, no satisfying milestones. Executives who do not understand the craft will pressure the team to "get into production." Resist. Or, if you are the executive, protect pre-production from the pressure to skip it. The four worst words in production are: "we'll figure that out later."

💡 Intuition: A vertical slice is not a demo. A demo proves "we can render this." A vertical slice proves "this is what the game feels like to play, and we are sure." If your slice does not change your mind about anything, it is not a slice — it is a sales pitch.

Production — The Long Slog

Once pre-pro hardens, you enter the longest phase: building the actual game. The shape of production has converged across most of the industry on some variant of agile, but with significant adaptations for the realities of game dev.

Sprints / Scrum. Two-week iterations are the dominant cadence. A team plans the sprint, executes, demos, and retrospects. The sprint backlog is a mix of features, content, and bug-fixes. Many studios run "Scrum-but" — they took the parts that work (sprint reviews, retrospectives, demoable increments) and dropped the parts that do not (rigid story-pointing, daily standups that drag, religious adherence to the methodology). The discipline matters more than the dogma.

Kanban. Boards visualizing work in progress, with WIP limits. Often used for content pipelines (asset production) where there is a continuous flow rather than feature increments.

Content pipelines. A pipeline is the process by which an asset moves from concept to in-game. For a character: concept art → model → rig → animation → texture → in-engine → review → revision → final. Each stage may be a different person, a different discipline, sometimes a different studio. Content pipelines are the production engineer's craft. A bad pipeline can double the cost of every asset; a good pipeline is invisible.

Asset reviews. Cross-discipline meetings where art, animation, and design leads approve work-in-progress. Healthy review culture is supportive and specific; toxic review culture is performative authority and design-by-committee. The leads' job is to maintain the former and prevent the latter.

Daily builds and CI. Modern AAA projects compile and test the game multiple times per day. A build that breaks blocks the team. The build engineer's role is to catch regressions before they cost a team-day.

Milestone definitions. Most projects mark major gates with formal milestones — Alpha, Beta, Release Candidate. Each milestone has a definition (what must be true to claim that milestone). The hardest discipline in production is honesty about milestones. A team that calls Alpha when it is really pre-Alpha is lying to itself and to the publisher, and the lie will be paid for in the next phase. Feature-locked means no new features added; content-locked means no new content added; code-locked means no new code; the order varies by studio.

Scope management. The single hardest production discipline. The temptation to add features, content, polish — to do "just one more thing" — is universal. Scope creep is the silent killer. Chapter 37 will dig into the craft of scoping (must-have, nice-to-have, cut list), but the production-time form of the discipline is: every change request gets evaluated against the milestone schedule, and the team is honest about whether the change can be absorbed. Most cannot. Most should be cut, deferred, or have something else cut to make room.

What production looks like, day to day, is unglamorous. You wake up, you check your task board, you work on whatever sprint task you have. You hit a blocker, you ask in Slack, someone helps. You merge to the build, the build breaks, you fix it. You go to a code review or an art review. You go home. You do that for three years. Sometimes there is a great demo and the team feels alive. Sometimes there is a brutal milestone and people cry. Mostly there is just the slog of converting a plan into a thing.

The Funding Landscape

Money is the substrate. No money, no game. The funding sources for game development have changed significantly over the past fifteen years, and the changes shape what is currently possible.

Self-funded. You pay yourself out of your savings, day-job income, or partner's income. The cleanest path because there are no obligations to anyone — but also the riskiest, because the runway is whatever you can personally sustain. Most successful indies started this way. Eric Barone (Stardew) lived on his savings and his partner's income for years. Lucas Pope (Papers, Please, Return of the Obra Dinn) self-funded.

Bootstrapped. You take small contracts to fund your own work. Your "real" game is what you build between client gigs. The advantage: you are not bleeding savings. The disadvantage: contract work always swallows more time than planned. Many small studios live this way for years.

Grants. The cleanest non-savings money. Indie Fund (Jonathan Blow, Ron Carmel, others) provides recoupable advances to indie projects. Epic MegaGrants (Unreal-only) gives non-recoupable grants — meaning, no royalties owed back. Platform first-party grants — Sony's PlayStation initiative for indies, Microsoft's ID@Xbox, Nintendo's developer programs — provide funding plus platform exposure in exchange for varying degrees of timed exclusivity. Government grants exist in Canada, the UK, Quebec specifically, France, Germany, Korea, and elsewhere, varying in generosity.

Publisher advance. A publisher gives you a recoupable advance against future royalties. You build the game; the publisher publishes and markets; you split revenue, with the advance recouped against your share before you see royalties. Typical revenue splits: 70/30 publisher/dev is common for first-time devs; experienced devs negotiate up to 50/50 or beyond. Devolver Digital, Annapurna Interactive, Raw Fury, Playtonic Friends, Hooded Horse are well-regarded indie-friendly publishers; the larger AA publishers (505 Games, Focus Entertainment, THQ Nordic) operate similarly with bigger budgets and tighter constraints. The catch: publishers approve key creative decisions, control the marketing calendar, and own the relationship with platforms. A bad publisher deal can cripple a good game (see "common pitfalls" below).

Venture-funded. Took several years of frenzy in the 2020-2022 window — VC money flowing into web3 games, crypto-adjacent studios, and "metaverse" plays. Mostly catastrophic. The crypto/web3 model collapsed in late 2022 along with the broader crypto market, taking many studios with it. As of writing, VC interest in games is sharply down. A few long-running games-focused VCs (Bitkraft, Galaxy Interactive, Andreessen Horowitz's games practice) remain, but the bar is higher.

Angel investor. A wealthy individual writes a check, often as a friend-of-team-member. Less common in games than in tech generally, but occurs.

Crowdfunding. Kickstarter's 2012-2015 golden era funded Broken Age, Pillars of Eternity, Shovel Knight, Yooka-Laylee, Bloodstained, Shenmue III, and dozens more. The model has cooled significantly. Backers became wary after high-profile failures (Mighty No. 9, Star Citizen shipping never). Modern crowdfunding works for projects with a built-in audience (sequels, established creators) but is hard for unknown teams. Backerkit and Fig (now defunct) extended the model with equity and post-campaign pre-orders. Patreon funds ongoing development for some indie creators (Caves of Qud, Dwarf Fortress before Steam release).

Platform investment funds. Sony, Microsoft, Nintendo, Tencent, NetEase, Sea Group, miHoYo, and Embracer Group have all run investment funds that take equity stakes in studios in exchange for capital and platform support. The trade is partial loss of independence for capital and distribution.

In 2026, the most common indie path is some combination: self-funded for the early prototype, then a publisher advance for production, then platform first-party support for marketing. The pure self-funded indie ship-to-Steam model still works — Vampire Survivors was one developer with a tiny Unity hobby project — but the discoverability problem has gotten worse, and most indie successes have publisher support behind them.

Pitching to a Publisher

If you are pursuing a publisher advance, the pitch is the gate. A pitch is a meeting (often on a video call) plus a deck plus a playable build.

The deck typically covers:

• Hook. One sentence. "It's Dark Souls with the flow of Devil May Cry." "It's Stardew Valley with explicit anti-capitalist politics." "It's Inside if Inside was a roguelike." Publishers pattern-match against successful games. You will hate this; it is unavoidable.
• Game pillars. Three to five sentences of design intent. What is this game about?
• Target audience. Who plays this. Comparable games and their sales numbers if you have them.
• Scope. Hours of content, content categories, number of levels, number of characters, etc. Realistic. Publishers can smell scope inflation.
• Team. Who you are, what you have shipped, why you are credible.
• Schedule and budget. Roughly when you ship, roughly what it costs. Both should be conservatively estimated and presented with a buffer.
• Prototype / vertical slice. The actual playable thing. This is where most pitches succeed or fail.
• Marketing / community plan. How players find this. A publisher wants to see you have thought about the audience, even though they will ultimately drive marketing.
• Risks and mitigations. Acknowledge the hard parts. The pitches that pretend nothing is risky read as naive.

What publishers actually want, despite their own marketing, is de-risked upside. They want a game that has a clear comparable they can model revenue against, a recognizable hook, a credible team, and a budget they can afford to lose. They do not want pure innovation; they want familiar enough to underwrite, novel enough to stand out. The pitches that succeed thread that needle. The pitches that fail are usually too weird (publisher cannot model the revenue) or too generic (publisher cannot model the differentiator).

The power dynamic is asymmetric. The publisher has many pitches; you have one game. The publisher has lawyers; you may not. The publisher has standard contract terms that favor them; you have to negotiate every favorable term in. Get a games-industry lawyer before you sign anything. Do not sign on the strength of a verbal promise. Do not give away your IP to fund development unless you have to and you understand exactly what you are giving up. We cover the contract side in more detail in Chapter 38.

Revenue Splits and Platform Economics

Once you ship, the platform takes its cut. The platform fees are the silent partner in every game's revenue model.

Steam. 30% on the first $10 million of lifetime revenue per title; 25% on revenue between $10M and $50M; 20% on revenue above $50M. The thresholds make Steam more attractive for hits and roughly equivalent for everyone else. Valve's argument for the cut is the platform — Big Picture, the Workshop, social features, refunds, regional pricing, server CDN, payment processing in 30+ currencies. Critics argue the cut is rent extraction enabled by Steam's near-monopoly on PC discoverability.

Epic Games Store. 12% (with payment processing included) or 0% if you use Epic's free stipulated terms with the Unreal engine in certain conditions. Epic launched in 2018 to undercut Steam and has invested heavily in exclusivity deals (paying upfront for timed Epic-only releases). Discoverability on Epic remains weaker than on Steam; the lower fee partly compensates.

Console (Sony, Microsoft, Nintendo). 30% standard. Console parity requirements — the unwritten rule that you cannot launch on PlayStation late if you launched on Xbox first, etc. — have softened over time but still affect deal-making. Publishers and large indies sometimes negotiate better terms; small indies pay sticker.

Apple App Store / Google Play Store. 30% standard, 15% for the first $1 million of annual developer revenue under the Apple Small Business Program (introduced 2021) and the Google equivalent. The 30% rate has been the subject of extensive litigation (Epic v. Apple, Epic v. Google) and EU regulation (Digital Markets Act). Alternate app stores on Android have always existed; on iOS they emerged in the EU in 2024 under DMA. Outside the EU, the 30% / 15% structure remains dominant.

AppLovin / mediation networks. For ad-supported mobile, the ad mediation layer takes its own cut, typically 10-20%. UA (user acquisition) costs eat the rest. The math of mobile F2P is brutal: you pay $5 to acquire a player who will spend on average $3 over their lifetime, and you make money only on the long tail of whales (the top 1-2% of players who spend hundreds or thousands).

Subscription services. Game Pass, PlayStation Plus, Apple Arcade, GeForce Now subscription. Revenue model varies: some pay flat upfront fees for inclusion, some pay per-hour-played, some pay revenue-share against subscriber engagement. The economics for developers are still being worked out and are studio-specific.

The 30% platform tax is the single largest line item in any game's P&L after development cost. It is not negotiable for small developers. It is a structural reality you must price into your scope.

The Indie Reality

Now, the hard numbers.

The median game on Steam earns under $5,000 in lifetime gross revenue. Multiple studies (VG Insights, GameDiscoverCo's Simon Carless, Steam Spy when it still worked) have triangulated this number. After Valve's cut, after taxes, after development costs, the median Steam release is a financial loss for its developer.

This is not because indie developers are bad. It is because the supply of games has exploded — over 10,000 games released on Steam in 2023 — and the number of buyers has not kept pace. Discoverability is a brutal funnel.

Success rates by genre vary, but every category has the same shape: the top 1% earns 90%+ of the money. Stardew Valley sold over 30 million copies. Vampire Survivors sold 5+ million. Hades sold 5+ million. Balatro sold 5+ million in its first year. The middle of the indie distribution — games that sell 10,000 to 100,000 copies — is shrinking. The bottom — games that sell under 1,000 copies — is growing.

What does "indie success" actually look like, statistically? It looks like one of:

1. The breakout hit (1-in-1000 or worse): you are a household name, you can fund the next project from this one alone, you may never need to work for someone else again.
2. The sustainable success (maybe 1-in-100): the game pays back development plus enough for the team to start the next project. Not life-changing money, but a viable career.
3. The break-even modest hit (maybe 1-in-50): the game pays back development cost. The team can keep going if they cut their take or raise more money.
4. The financial loss but career win: the game does not pay back, but you got a portfolio piece, made connections, learned the craft. Many "first indie games" are this.
5. The total loss: the game does not earn enough to cover anything. You return to a day job.

The Stardew Valley story is the most quoted, and the most misleading. Eric Barone spent four years working alone, sustained by his partner's income. He had near-zero marketing. The game caught on through word-of-mouth, enabled by the fact that the design was brilliant and the timing was right. The story is real. It is also survivorship bias. For every Eric Barone there are tens of thousands of solo developers whose four-year projects sold 800 copies. Their stories do not get told because there is no GDC talk to give about them. Calibrate accordingly.

The actionable advice is: if you are going indie, plan for the modest case. Have a runway. Have a side income or a partner who can carry you. Have a Plan B. Set a financial floor — if the game has not earned $X by month 6, you will return to client work or a job. Indie is a viable path; indie is not a lottery ticket.

Work-for-Hire and Outsourcing

The credits on Elden Ring run for over thirty minutes. FromSoftware's internal team is not enough to fill that runtime. Virtuos, Streamline, and a host of other outsourcing studios contributed massive quantities of asset work — environment art, character animation, level construction — without their contributions being individually credited in many cases.

This is the invisible industry. It is not new. Western AAA studios have outsourced art and engineering work to Eastern European, Chinese, and Southeast Asian studios since the late 2000s. The economics are straightforward: a senior artist in Singapore or Krakow or Manila costs less than a senior artist in San Francisco or London or Tokyo. The work is just as good. The pipeline lives on Perforce and Slack regardless of where the artist sits.

The problems are also straightforward. Credits practices vary; some publishers credit outsourcing studios, many do not. Working conditions and pay vary widely by region; some outsourcing studios pay well by local standards, others run brutal crunch on contract work for hardly any margin. Quality varies; the best outsourcing studios produce work indistinguishable from internal teams, others produce work that needs significant rework. The IGDA and various worker advocacy groups have pushed for credits transparency, and the practice has improved in the 2020s, but it is still common for outsourced work to go partially or fully uncredited.

If you spend a career in this industry, you will likely work with outsourcing studios at some point — even from an indie position, you may contract concept art or audio work to a small studio overseas. Treat the people you work with as collaborators, credit them in your game, and understand that the relationship is reciprocal. The "invisible" industry is the largest part of the workforce.

The Human Side — Careers, Stability, Burnout

Now the hardest section.

Games careers are boom-bust. You ride the production cycle of your studio's project — pre-pro is fine, mid-production is intense, alpha-to-launch is brutal, post-launch is recovery. Then your studio either greenlights a new project, or restructures, or gets acquired, or gets shut down. The 2023-2024 wave of layoffs is the most recent and the most severe in industry history: over 15,000 jobs lost across the industry in 2023, and another 10,000+ in 2024, by the IGDA's tracking. Studios closed: Mimimi (Shadow Tactics, Shadow Gambit), Volition (Saints Row), Tango Gameworks (Hi-Fi Rush, briefly — later spun out), Arkane Austin (Redfall), Roll7, Possibility Space, Smilegate Barcelona. Massive layoffs at Microsoft, Embracer Group, Unity, Sony, Riot, Bungie, Take-Two, ZeniMax. Whole studios staffed by veterans with twenty years of experience were dissolved.

The cyclicality is structural. Games are project-funded, project-staffed, and the cycle does not respect anyone's mortgage. When a project ends, the team that built it is often laid off, regardless of how successful the project was. Mimimi shut down after Shadow Gambit underperformed despite excellent reviews. Tango Gameworks was closed after Hi-Fi Rush was a critical success but a commercial soft-launch (the studio was later partially revived under a new owner). The talent of the people involved had no bearing on whether they kept their jobs.

This is the part of the industry no one tells you in a GDC talk. The reality of a games career, statistically, is: you will work at three to seven studios over a fifteen-year career; you will be laid off at least once; you will be asked to crunch at least once; you will burn out at least once; you will leave the industry and come back at least once.

The mitigations are real but partial:

• Build a portfolio that is not tied to one studio. Side projects, jam games, contributions to community tools, public talks — anything that demonstrates capability outside your employer.
• Build a network that is wider than your studio. When the layoff comes, you find your next job through people who already know your work.
• Have a financial buffer. Six months of expenses minimum, more if you can. The job search after a layoff often takes 3-9 months.
• Specialize in a transferable craft. Engine programmers can move between any studio using their engine. Tool programmers are valuable everywhere. UX designers are increasingly cross-industry. Combat designers are more locked into one type of game.
• Know when to leave the industry temporarily. The web/SaaS industries pay better, treat workers better, and are open to game-developer skill sets. Some of the best veterans of the games industry have rotated into and out of game work multiple times.

The "3-year rule" — quit your day job and go indie when you have 3 years of savings — has been stress-tested in the 2020s and is still roughly correct. If you have less than 18 months of runway, going indie is a bet you cannot afford to lose. If you have 36 months and a clear project plan, the bet is a gamble worth making.

The job is worth doing. The medium is worth contributing to. But you should know what you are signing up for. Calibrate, plan, and protect yourself.

Unions and Worker Power

For decades, the games industry was the largest unorganized sector of the entertainment industry. Hollywood is unionized. Television is unionized. Music has guilds. Games — until recently — had nothing. The IGDA (International Game Developers Association) is a professional advocacy group, not a union; it cannot collectively bargain.

The change began in 2022. Activision QA Albany (then called Raven Software's QA department in Albany) voted to unionize with the Communication Workers of America (CWA), becoming the first AAA games union in the United States. Raven QA Wisconsin followed. ZeniMax QA (across multiple studios) unionized with CWA in early 2023, the largest games union win to date. Sega of America QA unionized in 2023. Bethesda Game Studios (the Maryland design and dev team) voted to unionize in 2024 — the first non-QA AAA dev team in the US to organize.

Internationally, the game-worker organizing landscape varies. In France, Game Workers Unite and the STJV union have organized at multiple studios. In the UK, IWGB Game Workers has organized successfully at Activision Blizzard King and other studios. In Sweden, collective bargaining is normalized; major studios like Paradox have works councils. In Korea, Smilegate and Nexon workers have organized through tech-worker unions. In Canada, Keywords Studios employees in Edmonton unionized in 2022.

The publisher anti-union playbook, especially in the US, has involved: "captive audience" meetings (mandatory anti-union briefings), studio relocations to union-hostile geographies, mass layoffs in organized or organizing departments, and legal challenges to bargaining unit definitions. Microsoft has been a partial outlier — having recognized CWA's right to organize at Activision Blizzard during the 2023 acquisition negotiations, the company has not fought subsequent unionization efforts as aggressively as the historical norm. (Whether this remains true through future business cycles is uncertain.)

IATSE (the International Alliance of Theatrical Stage Employees), which already represents many cinematics and motion-capture workers, has signaled interest in broader games organizing. SAG-AFTRA's 2023-2024 strike against major game publishers over AI voice rights and likeness usage was the first industry-wide labor action in games and ended with a partial agreement.

If you work in games and your studio is organizing, you have a legal right to participate without retaliation (in most jurisdictions; specifics vary). If you are management, your obligations under labor law are specific and worth understanding. If you are a designer choosing where to work, the labor culture of a studio is a legitimate input to your decision.

Progressive Project Update

Ch 34: Write a 1-page studio pitch document for your action-adventure project.

You are going to write the document you would hand to a publisher, an investor, or a platform first-party fund. One page. One side of a sheet. The discipline of compression is the point.

Required sections:

• Studio name and team. Even if it is just you, give yourself a studio name. Note your team size and key roles (e.g., "Solo dev with contracted audio").
• Project name and tagline. One sentence.
• Hook. "It's [game] meets [game] with [twist]." Two sentences max.
• Core gameplay loop. Three sentences. What does the player do moment-to-moment? What is the hour-by-hour engagement?
• Target platform(s) and audience. Where does this ship? Who plays it? What comparable games does this audience already love?
• Scope. Hours of content, levels, characters. Be specific and conservative.
• Schedule. When does the vertical slice complete? When does production wrap? When does it ship? Use months from now, not absolute dates.
• Budget estimate. Total dollars to ship. Break down into salary, audio, art outsourcing, marketing, tools, contingency. Even rough numbers.
• What I/we are asking for. Are you self-funded? Looking for a publisher? Applying for a grant? State the ask clearly.
• Why us. What gives you credibility? Past projects, jam wins, demo reception, technical demonstrations.

Constraint: it must fit on one printed page. You will be tempted to make the font small. Resist. If it does not fit at 11pt with normal margins, you have not compressed enough.

This is a real document you would actually use. Save it. The next time you talk to anyone about your game — a publisher, a friend who can fund a small grant, an Indie Fund application — you have something to send. Most indies wait until they need a pitch document to write one. The ones who write theirs early have an advantage.

Common Pitfalls

Believing the indie dream. "I'll make my Stardew Valley in 18 months." No, you will not. The Stardews are 1-in-10,000. Plan for the modest case.

Signing a bad publisher deal. Publishers send you a contract that includes a "key person" clause locking you to their company personally, a perpetual IP grant, an unspecified marketing-spend cap, and a recoupment formula that puts you 2-3 games deep before you see royalties. You sign because you are excited and tired. You spend the next decade wishing you had not. Get an entertainment lawyer who specializes in games. Pay them. It is the cheapest insurance you will ever buy.

Equity without a contract. "We're co-founders, we'll figure out equity later." Then the project starts working, and "later" arrives, and one of you is now claiming a much larger share than the other believed was the deal. Founders' agreements at the start of a project have saved more friendships and studios than any other piece of paperwork. Use a template, fill it in, sign it, store it. Address: equity split, vesting, what happens if someone leaves, what happens if the project is sold, who owns the IP if you split up.

No written agreements with collaborators. A composer scores your game on a handshake. The game ships. The composer wants royalties. You believed it was a flat-fee deal. Now you have a dispute. Always: write down the deal. Even an email confirming "to be clear, this is a flat $5,000 fee for full rights to the music, with credit as composer" is enforceable.

Skipping pre-production. Discussed above. The single most expensive mistake in production. You cannot rescue a project in production that did not figure out what it was in pre-pro.

Underbudgeting marketing. Indies routinely budget for development and forget that a game which no one knows about does not sell. A reasonable indie marketing budget is 25-40% of the development budget, even for a "Steam page and tweets" approach (because that approach requires hiring a community manager, paying for trailers, attending events, paying influencers, etc.). If your budget has zero for marketing, you have not finished budgeting.

Summary

The game industry is not one industry; it is a dozen industries with different economics, schedules, ethics, and career paths. The studios are the rooms in which games are dreamed; the role taxonomy is more diverse than any single studio reveals; the project pipeline has nine to ten stages, of which pre-production matters most and is most often shorted. Funding ranges from your savings to publisher advances to grants, each with tradeoffs. Platforms take 30% (typically), and that fee is a structural line item in every game's P&L. The median Steam game makes under $5,000; the indie success stories are real and rare. Work-for-hire and outsourcing studios are the invisible majority of the workforce. Games careers are boom-bust, with massive layoff waves in 2023-2024 affecting 25,000+ workers. Unions are emerging — first US AAA games unions in 2022-2024 — and changing the worker-power dynamic for the first time in decades.

The next chapter (Ch 35) looks at genre as a business category as much as a design category — which is its own kind of industry literacy. Chapter 37 returns to scope and production, the craft of converting a plan into a shipped artifact. Chapter 38 covers publishing and marketing — the work of getting your finished game in front of players. The thread connecting all of these is the same: design does not happen in a vacuum. It happens inside structures. The structures are the medium.

Now write your one-page pitch.