Troika's Flat Hierarchy Experiment

tim-cain troika arcanum management game-development flat-hierarchy studio-culture

Abstract

Problem: Can a game studio operate with a completely flat hierarchy — no titles, no roles, equal pay, equal work?

Approach: When Tim Cain, Leonard Boyarsky, and Jason Anderson founded Troika Games in 1998, they ran the entire development of Arcanum (1998–2001) as a flat hierarchy experiment: only senior hires, identical salaries and benefits, no defined roles, equal work distribution, and alphabetical credits with no titles.

Findings: The experiment failed on multiple axes. People don't work at the same rate, specialists can't simply swap hats, ego clashes arise when peers must judge each other's work, publishers and press demand specific points of contact, and external entities (publishers, review sites, legal bodies) simply don't accommodate flat structures.

Key insight: Structure exists for the same reason it exists in code — it provides clarity about who can say yes or no, reduces hurt feelings, and lets things flow faster. Troika abandoned the flat model after Arcanum and adopted traditional hierarchies for Temple of Elemental Evil and Vampire: The Masquerade – Bloodlines, which worked noticeably better.

Background

When Troika Games was founded in 1998, the three co-owners — Tim Cain, Leonard Boyarsky, and Jason Anderson — decided to run the company as flat as legally possible. Externally, they needed designated owners to sign paperwork with lawyers, the City of Irvine, Orange County, the State of California, and the federal government. But internally, the structure was radically egalitarian.

The Rules of the Experiment

Equal Pay and Benefits

Every person at Troika — owners included — received the same salary, the same insurance, the same time off, and the same bonuses. When Arcanum royalties came in, 50% went to fund the next project and 50% was split equally among everyone who worked on the game.

This created immediate financial strain. The unexpected costs of opening a new company meant the three owners had to drastically underpay themselves for the first year, to the point where they struggled with personal bills and mortgages. They eventually stopped subsidizing and instead cut company expenses — fewer snacks, reduced electric bills — to make it work.

Senior-Only Hiring

Troika hired only senior developers to ensure everyone was on a level playing field with comparable experience and capability. In practice, this didn't work as intended — even among seniors, people had vastly different abilities, specializations, and work speeds.

Equal Work Distribution

Work was estimated collectively, then divided into equal slices for each person. This failed quickly. Some people finished a week's work in three days; others weren't half done by Friday. When one programmer complained his workload was too heavy, Tim swapped assignments with him — Tim finished the programmer's work, but the programmer couldn't finish Tim's. The objective reality was clear: people don't work at the same rate, no matter how equal you declare them.

No Defined Roles

Anyone could wear any hat — write dialogue, design quests, do scripting, whatever they wanted. The catch was that the group collectively judged the quality of the output.

This revealed two painful truths:

  • Many people wanted to wear lots of hats but were only good at one. Tim himself learned he wasn't a good writer and already knew he couldn't do art well (he doesn't even see colors well). His interface suggestions were politely dismissed.
  • People resisted accepting they were bad at something. They'd ask to redo work, claim the group was being picky, or insist their work was good and everyone else was wrong. With no authority figure to defer to, these ego clashes were difficult to manage.

Highly specialized work — engine programming in C, 3D modeling, rigging, animation — proved impossible to just "put a hat on" and do. People tried. People failed.

External Forces That Broke the Model

Publishers Wanted Specific People

Sierra (and later Activision) didn't care about Troika's internal flat structure. They wanted to deal with one specific person. For Sierra and Arcanum, that was Tim. For Activision, it was Leonard. When Sierra demanded all three owners fly to Seattle for a meeting and Tim (who dislikes flying) wasn't going to attend, Sierra insisted — they needed him specifically. But when Jason later couldn't make a rescheduled date, Sierra said that was fine. The asymmetry was impossible to control.

Press Wanted Spokespeople

Journalists visiting the studio wanted to talk to certain people. Some team members were better at giving interviews and sound bites than others. In a flat hierarchy, telling someone "you're not good at talking to press" required the whole group to essentially vote on it, which felt far harsher than simply having a designated spokesperson role.

Credits and Review Sites

Arcanum's credits simply read "Arcanum was made by" followed by approximately 14 names in alphabetical order — no titles, no roles. Review sites of the early 2000s didn't know what to do with this. MobyGames reportedly listed Tim Cain as "Lead Animator," randomly slotting people into roles when they couldn't find actual titles.

Contracts, invoices, and legal documents required designated signatories. The three owners tried to be interchangeable — any of them could sign anything, with a rule that they'd wait 24–48 hours to try to get all three present before proceeding. But externally, this interchangeability was never truly respected.

The Aftermath

After Arcanum shipped, Troika deemed the flat hierarchy experiment a failure on multiple axes. When the studio split into two teams for Temple of Elemental Evil and Vampire: The Masquerade – Bloodlines, both adopted traditional hierarchies with lead designers and lead programmers.

The results were immediate: things flowed better, people could get definitive answers, there were fewer hurt feelings, and the studio could hire across the experience spectrum — including juniors and fresh graduates — instead of being limited to expensive senior-only hiring.

Tim's Takeaway

Tim compares organizational structure to code structure: "I don't want a bunch of random modules shooting events at each other. I want a structured code base." Similarly, he prefers knowing who can tell him yes or no, and who he can tell yes or no to.

He acknowledges that others may find ways to make flat hierarchies work, but advises being prepared for the problems Troika encountered. His experience was clear: hierarchy provides structure, and structure makes things work better.

Source: Tim Cain — "Troika's Flat Hierarchy Experiment"

References