Gabe Newell on Valve's past, gaming's present and Steam's future

Jump to

Valve co-founder Gabe Newell's back-to-back lectures given Wednesday to University of Texas students of the LBJ School of Public Affairs focused on a subject that was as important to the company when it was founded in 1996 as it is today: Productivity is one of the most valuable commodities in the world.

Newell came to this conclusion during his tenure at Microsoft, where he noticed that the metrics for measuring the penetration of software were entirely misleading. Microsoft — and nearly every other software producer — distributed their products through resellers, who would in turn report back to the company with their list of top-selling items. Those reports had major deficiencies (particularly for tracking international, or third-hand sales), which Microsoft learned after performing its own investigation into the sales and penetration of Windows.

Despite the fact that Windows was in 30 million homes, and was performing much better than Microsoft had assumed, it was still being outsold by a much smaller piece of software, from a much smaller company: id Software's Doom.

"Somehow, the largest software company in the world was being out distributed by a 12-person company in Mesquite, Texas," Newell said. "There was no conception, there was no model for how this made any sense. How could they possibly have done this, right? It took years to build up enough distribution strength to go from number seven to number six in the word processing category, much less to come out of nowhere with a product that would be on more desktops than the most important product that Microsoft had."

When Newell and collaborator Mike Harrington decided to found Valve, avoiding the kinds of structural sinkholes and lack of direct consumer input they'd experienced at Microsoft was the highest priority. As id Software was proving, it was an ineffective way to distribute products; but, more importantly, it was wildly unproductive.

"So what did that mean in terms of how you were designing a company," Newell asked. "This is the question that I was trying to answer in 1996. Because we'd go out and meet with all these companies; insurance companies, and airline companies and start-ups in Silicon Valley, and you had to come up with some way of thinking what the design was. And the thing that we became convinced of — Mike Harrington, who was the other founder, and myself — was that everybody was going in the wrong direction."


Putting the team together

Instead of outsourcing jobs, ensuring work was being done as inexpensively as possible, Valve would employ the "most expensive talent that was out there in the world," talent which Newell measured in productivity. To build that team, Newell and Harrington would meet with potential employees face-to-face, and pitch the company to them individually.

A lot of people said no.

"A lot of people we thought were going to join us in the beginning didn't," Newell said. "We were like 'Let's go, let's storm the castle!' And then it was like, 'Where'd everybody go?'"

However, the idea took hold in a group of influential individuals, and spread through social networking. In particular, Newell named Michael Abrash and John Carmack as industry figures who helped spread the word and build authenticity for the company.

We were like, 'Let's go, let's storm the castle!' And then it was like, 'Where'd everybody go?'

Valve also reached outside the typical hiring pool to find the founding developers who would help establish the company and create its first pieces of software.

"We were willing to take what other companies would say were risks," Newell said. "One of the first programmers at the company, his previous job was manager of a Waffle House. He was one of the most creative, and still is one of the most creative programmers in the industry."


In the interest of maximizing productivity, the founders decided that there would be no titles; the entire company would have a flat organizational structure. Valve has no Q&A department. It has no marketing department. Doug Lombardi, the vice president of marketing for Valve, has no employees working under him for the express purposes of marketing. He wasn't given the title for the company's sake, Newell explained, but to assuage the confusion of press and outside observers.

Employees of Valve are free to design their own office space to maximize their own efficiency, Newell explained. A pair of programmers are experimenting with paired programming, and are building a special desk designed to facilitate that process — not due to novelty, but because they've determined that building a desk (and not a game) is the best way to maximize productivity.

It is, as you might imagine, a process that's not devoid of problems. One of the most notable, Newell said, was how to introduce new human resources into the process, and recognizing when that process just wasn't working out.

"You have to be really aggressive about firing people," Newell said. "We haven't done a really good job with interns or new hires. That's kind of a sink or swim thing. People have to take it seriously, right? It's an engineering problem in the sense of, you have to make decisions, measure outcomes, and make changes as a result of it. I would have trouble working any other way now, I think most people at Valve would have trouble.

"There's something we somewhat unkindly call the beaten wife syndrome, where people come in from other industries and really struggle. The worst are people from the feature film industry, where they've been taught that any time they show initiative, that somebody's going to leap out and smite them for doing that. It usually takes about six to nine months for people to really sort of internalize the working model of the company."

The Multiplayer Divide

When Valve's most productive creators do take that initiative, interesting things can happen.

Some of the most unique advances in Valve's single-player games were discovered by the outside-the-box interpretation of simple heuristics — how players were immersed in Half-Life 1 because bullets left a permanent mark on the environment, or how Half-Life 2's characters were considered to be more relatable, simply because they were able to make eye contact with the player.

That kind of innovation, Newell said, becomes much more difficult when additional players are added to the mix.

"Single-player games are like a feature film where your lead actor is retarded and autistic, but you can think of it like a feature film," Newell said. "In a multiplayer game, these rules that you come up with don't work anymore. An example would be in Counter-Strike, we put the riot shield in, and our player number go up. We take the riot shield out, and our player numbers go ... up. How do you explain that?

"You start to think, after a while, that multiplayer games are all about externalities. They kind of look more like operating systems, or a sport. In terms of how they behave, they behave a lot more, and value is created a lot more like a spectator sport than a feature film."

Single-player games are like a feature film where your lead actor is retarded and autistic, but you can think of it like a feature film.

Measuring success through metrics becomes tricky when dealing with a body of interconnected players, Newell explained. However, it also opened up opportunities for Valve to incorporate newfound productivity into their corporate architecture: It allowed them to democratize content creation among Steam users, and it allowed them to create robust economic systems within Steam itself.


"To be really concrete, 10 times as much content comes from the user base for TF2 as comes from us," Newell said. "So we think we're super productive and kind of badass at making TF2 content, but even at this early stage, we cannot compete with our own customers in the creation of content for this environment. The only company we've ever met that kind of kicks our ass is our customers. We'll go up against Bungie, or Blizzard, or anybody but we won't try to compete with our own user base, because we already know we're going to lose.

"Once we start building the interfaces for users to start selling their content to each other, we start to see some surprising things," Newell added.

Those "surprising things" are all microcosms of phenomena usually found on a macroeconomic scale. Some content creators shot to the top of the sales charts, earning immense wealth — one industrious user made over $500,000 in a single year, Newell said. User-made currencies appeared, inflated and collapsed. The service was no stranger to economic crises, but it also hosted economic boon, as well; so much so, that Paypal began to question their motives.

"The first two weeks that we did this, we actually broke Paypal," Newell said. "They didn't have, I don't know what they were worried about, maybe drug-dealing, because nothing generates cash to our user base other than selling drugs like this. We had to work something out with them, and say, 'No, they're making hats.'"

"No, they're making hats."

Content creators that were employed at other video game developers were making more money than they were in their full-time jobs, Newell said. Steam had fostered the economy of a middle-sized country.

"I like to tease Yanis [Varoufakis, Valve's resident economist] about the fact that Steam is five times the size of Greece," Newell said, "and that we have less debt."

It may seem bizarre that the ecosystem grew as quickly as it did, given much of its success could be attributed to Team Fortress 2 hats, and digital vanity items of that ilk. According to Newell, the intangibility of the goods being sold doesn't really matter; even digital status symbols are relevant.


"I don't think the value for that is any different than the value of lots of other kinds of status goods," Newell said. "Why do people buy Porsches?"

The future of the Steam marketplace, Newell said, is to ensure that goods can be more permanent in a player's collection; that they can be transferrable and exchangeable between titles. He called to fault the MMO model of player progression: Characters level up, purchase new items, then when you play a new game, everything you worked for is gone. Game creators currently have a "whimsical notion" of player's property rights, Newell said.

"It's like, 'Hey, I'll sell you a house, and you can do a bunch of work, paint it and put furniture in it, and then, when you go to a new house, we're going to burn that one down,'" Newell said.

"The sets of lessons we're learning today in the video game space are probably going to be true of a much wider range of industries tomorrow."

It's important for Valve to pass on and systemize every improvement they can incorporate into to Steam, Newell said. It's the only way to improve the platform for its users and content creators, but it also eases the burden of certain processing tasks off of Valve. However, opening up the platform could lead to potential threats like hacking — though Newell said developers' current thinking on hackers is a little overreactive.

"One of the biggest mistakes game developers make is to mistake a genuinely entertaining action by a player as hacking," Newell said. "The example game developers tell each other is when Lord British was killed during his first speech in Ultima Online, they rolled the world back, rather than recognizing that it was the coolest thing that had ever happened at that time, and they should have let that instance turn into whatever the instance would have turned into after they killed Lord British."

Giving control to the internet and its users is the only logical choice that a modern corporation can take on, Newell said. Pre-internet thinking in terms of direct communication with consumers is harmful; the internet is much better at organizing individuals than a corporation will ever be. Open markets are what thrive in the digital age, and as technology improves — particularly 3D printing and other forms of digital-to-physical fabrication — that philosophy will become relevant far outside the Steam ecosystem.

"The sets of lessons we're learning today in the video game space are probably going to be true of a much wider range of industries tomorrow," Newell said.

That much was clear from the very beginning of Newell's speech to the graduate students of the LBJ School of Public Affairs. Before starting, Newell asked who was there because they were interested in gaming, and nearly half the class raised their hands. He then asked who was there because they were interested in economics, and nearly everyone raised their hands.

"Those two aren't mutually exclusive," a student in the second row responded. Gabe agreed.