← all stories

Era 3 · PC, shareware and strategy games

1991–2009bought by ZeniMax Media (announced 2009-06-24; per Carmack's…

id Software

A four-person company that was making first-person 3D shooters two years before Doom, took the genre to a mass audience with Doom, and taught players that a game could be given away free, the sequel sold by mail order, and the files later opened up so that fans could build levels themselves.

Founders John Carmack · John Romero · Tom Hall · Adrian Carmack
Domains —
gamessharewaremoddingengine-licensing

In brief

Four employees of Softdisk, a Texas publisher of disk magazines (programmers John Carmack and John Romero, designer Tom Hall, and artist Adrian Carmack), made games for themselves in the evenings, and on February 1, 1991, they formally turned this into a company, id Software. In three years they went from Commander Keen, published by someone else through Apogee, to Wolfenstein 3D and Doom: games that did more than sell well: they rebuilt the first-person shooter genre and turned openness to mods and the licensing of their own engine into separate sources of money. In 2009 the company was bought for $150 million by ZeniMax Media, under whose wing id works to this day.

How it started (the founders)

All four met not as entrepreneurs but as salaried employees of Softdisk, a company that put out monthly disk magazines for DOS computers. Inside it, Carmack and Romero ran an unofficial group, "Ideas from the Deep," making games beyond their work plan. The turning point came in September 1990, when Carmack found a technique for smooth byte-by-byte screen scrolling on the PC, something no one had managed on that hardware before; with this technique they built a demo in the spirit of Super Mario Bros. 3 and soon after a full game, Commander Keen, published on December 14, 1990, through the publisher Apogee, before id even existed as a company. It was the first royalty check from Apogee that convinced the team to leave Softdisk and found id Software on February 1, 1991.

The first year was spent on the move: from Shreveport (where Softdisk was also based) to Madison, Wisconsin, in September 1991, and on April 1, 1992, to Mesquite, Texas, where the company settled for years. Alongside the organizational bustle ran a technical evolution: even before Wolfenstein 3D, id twice tried out the idea of a first-person shooter on its own engine, with Hovertank 3D in April 1991 and Catacomb 3-D in November 1991, both again published by Softdisk. These were not side experiments but direct prototypes: the Hovertank engine grew into wall texturing in Catacomb 3-D, and then into doors and moving walls in Wolfenstein 3D.

Year-by-year timeline

Lesser-known but significant facts

  1. Doom was not id's first attempt at a 3D shooter but its third. Hovertank 3D (April 1991) and Catacomb 3-D (November 1991) came out more than two years earlier and were already first-person games with a primitive 3D engine; they were simply published quietly, through Softdisk, rather than under id's own promoted name.
  2. Apogee and id disagreed about fans hacking Wolfenstein 3D. Apogee believed that mods and map editors cut into sales, since people got far more "free" levels than intended; id's developers, on the contrary, were glad of the hacking and decided not to hinder such activity but to encourage it in future games.
  3. By the end of 1993 Wolfenstein 3D had sold more than 100,000 copies, earning id about $8 on each, against roughly $2 per Nintendo cartridge: a direct comparison of the economics of PC shareware with the console cartridge market of that era.
  4. Doom's openness was a decision, not an accident of hacker ideology from day one. Carmack deliberately separated the game's levels and assets from the engine in WAD files, even more thoroughly than in Wolfenstein 3D, and released the format documentation on release day; but fully open source code under the GPL appeared only in 1999, after the author of one of the mods lost his work when a disk failed.
  5. Tom Hall's departure in 1993 was a conflict not about money but about story. Hall wrote the "Doom Bible" with a backstory and cutscenes; Carmack and Romero argued for a minimum of story for the sake of the game's pace, a philosophy characteristic of id that would later harden into Carmack's far blunter statement about story in games.

Legend vs. the record

The first growth lever

The lever was not the game itself but the combination of two decisions: a free first episode and a deliberately open architecture for mods. In the very first days after the December 1993 release, Doom's shareware registrations brought in about $100,000 a day, even though, by the estimate of researchers of the topic, those who downloaded the first episode for free outnumbered those who paid by an order of magnitude: installs running far past eight figures against 2–3 million copies of the shareware version sold. A minority paid, but it was the free majority that created word of mouth and demand for the sequel. The second layer of the lever kicked in a year later: the retail release of Doom II in October 1994 collected 250,000+ preorders and sold more than 2 million copies, as the shareware hype of 1993 converted into 1994 retail through a channel Wolfenstein 3D had not had. The same free layer worked for retention: open WAD files let fans make levels and mods from day one, extending the game's life with no marketing budget.

The paired story

Without Apogee Software, id would have had no starting capital: the publisher's advance and royalties for Commander Keen and Wolfenstein 3D funded the team in 1990–1992 (the deal details are in the Apogee dossier). The pair split over the very thing it grew out of: its attitude toward fans hacking the game. When players started breaking into the shareware file and building their own levels after Wolfenstein 3D, Apogee as publisher saw a threat to sales; id took the opposite position, welcomed the hacking, and in Doom deliberately made modification easier, with open WAD files and format documentation on release day. The split over distribution was economic as well: id decided that Apogee could not handle the volume of phone orders for Doom and hired a third-party company. The contrast: Apogee perfected the funnel of a free piece leading to a paid sequel and tried to control how much was free, while id widened the free layer through openness to mods, a bet the publisher considered risky and id saw as working for reach and loyalty.

Parallels today (projects from the catalog)

What a builder can take from this in 2026

  1. Your product will be hacked and remade anyway, so build that into the architecture in advance. id did not make Doom moddable out of pure altruism; it drew conclusions from how fans were already breaking Wolfenstein 3D and designed its next game so that this worked in its favor.
  2. A free layer is not charity but the cheapest sales funnel. The gap between the paying minority and the free majority for Doom was an order of magnitude, and it was the free majority that generated the demand that converted into $100,000 a day right after release.
  3. An internal tool you built for yourself can become a separate business. The Quake engine was built for id's own game and became technology licensed by Valve: a second source of money unrelated to sales of the game itself.
  4. When a distributor can't keep up with your growth, that's a reason to build your own channel, not to argue. id parted with Apogee for Doom not over a quarrel but out of sober calculation: the audience grew faster than the publisher could process orders.
  5. Radical openness can be gradual rather than all at once. It took seven years and one specific incident of lost code to get from not getting in modders' way in 1992 to a full GPL in 1999; you don't have to open everything at once for openness to work for growth.

Discrepancies and what we could not verify

Sources (primary first)

Secondary (context, cross-checking, accessed through search snippets; a direct fetch returned 403):

Internet History Keeper

Every historical dossier here is free and open, and it will stay that way. If you want the series to keep going (new dossiers, checks against primary sources, the English edition), support it with a subscription for $9.90 a month. It is support, not access, and you can cancel at any time.

Become a keeper — $9.90 a month →

The same thing, about today

The same breakdowns, but of projects launching right now: what the product is, where the first users came from, how they charge. The card is free, the full dossier is $5 (the dossier itself is written in Russian).

Browse the dossier catalog →