Sierra, LucasArts, and the two deaths of the adventure game

On-Line Systems' first commercial office — a back room behind Ponderosa Printing in Oakhurst, California. Photo by Guywelch2000, CC BY-SA 4.0.

By the late 1990s, Sierra On-Line's name sat on Half-Life, Homeworld, the SWAT and Tribes shooters, and the Caesar/Pharaoh city-builders — alongside the adventure canon it had built its reputation on: King's Quest, Police Quest, Gabriel Knight. On paper that is the catalog of a company poised to become one of the defining publishers of the next quarter-century.

It became nothing of the kind. "Sierra" is now a dormant trademark held, after a chain of acquisitions, inside Microsoft. The distance between that catalog and that outcome is the question worth answering, and the usual answer — that adventure games went out of fashion — is only half of it. Two separate forces brought the era's giants down: one financial, one structural. They are routinely collapsed into a single story, and they were not the same death.

Two studios, one rivalry

Sierra began in 1979 as On-Line Systems, founded by Ken and Roberta Williams in a back room behind a print shop in Oakhurst, California — a mountain town, not a tech hub. Roberta's Mystery House (1980) is generally credited as the first graphic adventure; King's Quest (1984) made the form commercially dominant for a decade. Its rival, Lucasfilm Games — founded 1982, renamed LucasArts in 1990 — produced Monkey Island, Day of the Tentacle, and Grim Fandango.

The two defined themselves in opposition. Sierra's games were punishing: they killed the player freely and allowed unwinnable states, so progress depended on constant saving. LucasArts built its house style on the inverse — its adventures generally refused to let the player die or reach a dead end, favoring comedy and craftsmanship over jeopardy. The disagreement was real, and both studios were, by any measure, foundational to the medium.

The financial death

The first collapse had nothing to do with games. In 1996, Sierra was acquired by CUC International — a consumer-services conglomerate — in an all-stock deal worth roughly a billion dollars. In a separate transaction the same year, CUC also bought Davidson & Associates, which happened to own Blizzard Entertainment. The two purchases made Sierra and Blizzard corporate siblings under one owner, not parent and child. In 1997, CUC merged with HFS Incorporated to form Cendant.

In April 1998, Cendant disclosed that CUC had been systematically inflating its income — by roughly half a billion dollars over the preceding three years. The stock fell from about $36 to $19 in a day, erasing some $14 billion in market value, in one of the largest accounting frauds of the era. Cendant declared its software holdings non-core and sold them off in late 1998 to the French media group Havas, soon absorbed into Vivendi. (Reported sale figures vary; the date is firmer than the price.)

The sequence put Sierra and Blizzard under one owner as siblings, then swept both out together — bundled, devalued, sold:

flowchart TD
    SIERRA["Sierra (1979)"] --> CUC["CUC buys both (1996)"]
    DAVIDSON["Davidson + Blizzard"] --> CUC
    CUC --> CENDANT["Cendant accounting fraud (1998)"]
    CENDANT --> VIVENDI["sold off to Vivendi (1998)"]

    classDef plain stroke:#7b88a1,stroke-width:2.5px
    class SIERRA,DAVIDSON,CUC,CENDANT,VIVENDI plain

Sierra had not lost its market. It was collateral in a securities fraud — bundled, devalued, and divested as part of a balance-sheet cleanup, then hollowed out by a new owner reorienting toward online platforms. Ken Williams had divested most of his holdings and left in November 1997, before the disclosure, and retired comfortably. The founders survived; the company did not.

What followed was a long descent through ownership — Sierra passing between conglomerates that had never shipped a game, each booking it as a line item rather than a studio. Vivendi merged its games division with Activision in 2008 to form Activision Blizzard — named for Blizzard, by then the most valuable asset in the portfolio, while Sierra had already been reduced to a label. In 2023, Microsoft acquired Activision Blizzard for $68.7 billion, a deal driven by Call of Duty and World of Warcraft; the dormant Sierra trademark rode along in the paperwork, an asset no one at the table was bidding for.

flowchart TD
    VIVENDI["Vivendi (1998)"] --> ATVI["Activision Blizzard (2008)"]
    ATVI --> MSFT["Microsoft buys it all (2023)"]
    MSFT --> BNOW["Blizzard: funded, World of Warcraft"]
    MSFT --> SNOW["Sierra: dormant trademark"]

    %% color = present-day status: green still shipping, red name-only
    classDef live stroke:#a3be8c,stroke-width:2.5px
    classDef dead stroke:#bf616a,stroke-width:2.5px
    classDef plain stroke:#7b88a1,stroke-width:2.5px
    class BNOW live
    class SNOW dead
    class VIVENDI,ATVI,MSFT plain

The sibling comparison is the tell. Sierra and Blizzard spent the Vivendi years under identical ownership; one was left autonomous and built World of Warcraft, the other was stripped to a trademark. The variable that separated them was not catalog quality or historical importance — both had those — but which asset the owner chose to fund. Corporate survival turned out to be orthogonal to how foundational a studio was.

LucasArts traveled a parallel road on a different timeline: Disney acquired its parent, Lucasfilm, in 2012, shut the studio's internal development the following year, and licensed the Star Wars games rights to Electronic Arts.

flowchart TD
    LUCAS["LucasArts (1982)"] --> DISNEY["Disney buys Lucasfilm (2012)"]
    DISNEY --> SHUT["internal dev shut (2013)"]
    SHUT --> EA["Star Wars rights licensed to EA"]

    %% color = present-day status: red name-only
    classDef dead stroke:#bf616a,stroke-width:2.5px
    classDef plain stroke:#7b88a1,stroke-width:2.5px
    class EA dead
    class LUCAS,DISNEY,SHUT plain

The structural death

The financial collapse explains why Sierra ended. It does not explain why the adventure game did, because the genre's decline was already underway and would have happened to an independent Sierra too. The timing separates the two clearly: LucasArts abandoned adventures around 2000 to concentrate on Star Wars, yet survived as a studio until 2013. The form died there more than a decade before the company did.

Three forces hollowed out the genre, and only the third is the one usually named:

• The shift to real-time 3D. Doom (1993), Quake (1996), and Half-Life (1998) relocated the medium's sense of spectacle to fast, immersive, real-time worlds. Against them, a static point-and-click screen read as obsolete.
• A design ceiling. Stretching a finite story into enough playtime pushed puzzle design toward arbitrary "moon logic," which limited how wide the audience could grow.
• Economics. This is the decisive one, and it is best seen as a contrast in revenue shape.

flowchart TD
    COST["High up-front cost: art, writing, voice, hand-built content"] --> FORK{"What shape is the revenue?"}
    FORK -->|single-player adventure| FIN["Sold once · ~15 hours · solved · finished"]
    FORK -->|competitive / live-service| INF["Played for years · ladders · community · updates"]
    FIN --> FINR["One-time revenue, then flat"]
    INF --> INFR["Recurring revenue that compounds"]
    FINR --> MKT["Capital chases the compounding model"]
    INFR --> MKT

    %% color = economic position, not quality: amber finite, green compounding
    classDef warm stroke:#ebcb8b,stroke-width:2.5px
    classDef good stroke:#a3be8c,stroke-width:2.5px
    classDef plain stroke:#7b88a1,stroke-width:2.5px
    class FIN,FINR warm
    class INF,INFR good
    class COST,FORK,MKT plain

An adventure game was expensive to produce — hand-authored art, writing, and voice — and was consumed once: roughly fifteen hours, solved, and shelved. Counter-Strike, StarCraft, and Diablo cost similarly to build but were played for years, sustained by competition, community, and a reason to keep returning. As broadband spread, the industry's capital followed the compounding model. A finished story could not match a standings table that never ends, and investment flowed accordingly.

This is the half of the popular explanation that holds: competitive and live games won the contest for engagement and the revenue that follows it.

What collapsed, and what survived

One thing collapsed, precisely: the point-and-click puzzle adventure as a mass-market commercial format. Not storytelling, not the studios' craft, not narrative ambition — that single format, and chiefly its mainstream footing. The engagement verdict is easy to overstate into something larger — that story-driven games lost — and that version is wrong. The narrative ambition Sierra and LucasArts pioneered migrated into genres with better economics and newer technology, and there it expanded enormously: The Last of Us, The Witcher 3, Red Dead Redemption 2, Disco Elysium, Baldur's Gate 3, and the entire walking-simulator lineage descend directly from what those studios were attempting. The original form did not vanish either. Telltale, founded by LucasArts veterans, revived episodic narrative adventures; crowdfunding produced spiritual successors; and in 2022 Ron Gilbert returned to make Return to Monkey Island.

The accurate version is narrower than the overstatement. Competitive games captured the engagement market; narrative design captured the prestige market and most of the largest productions in the medium; and the point-and-click adventure, caught between them, lost its mainstream economic footing without taking storytelling down with it.

What decides survival

The two deaths share a single lesson, and it is not "the best product survives." A studio's fate is set by the strategy of whoever owns it and by which revenue shape the market is rewarding at the time — variables largely independent of how good or how important its work is. Sierra and Blizzard proved the first half under one roof; the adventure genre and the live-service boom proved the second across the whole industry. The pattern isn't unique to games — technology markets routinely settle on economics and governance over raw merit.

Being early is no safer than being good. Sierra pioneered the graphic adventure, and the head start bought it nothing the market chose not to reward. The shape recurs: GitHub Copilot — owned, like the Sierra trademark, by Microsoft — was the first LLM coding assistant to reach mass adoption, a genuine first mover, and a wave of newer tools closed the gap within two years. The first mover clears the trail; it rarely owns the road.

What endured was neither the company nor the catalog but the people. The talent that built both studios dispersed and rebuilt the medium elsewhere — Double Fine, Telltale, the Valve lineage, and Blackbird Interactive, whose Homeworld 3 shipped in 2024. The franchises remain stranded as dormant entries in a portfolio assembled for other names entirely; the craft that made them did not stay in the vault. Being foundational secured a place in the medium's history. It secured nothing in its balance sheet.