Remarkable evidence leaked from Valve Software over the last few weeks, portraying the company as a game developer’s paradise—awash in wealth, brimming with talent, and providing employees with remarkable trust and freedom.
First came an article by programming guru Michael Abrash praising Valve as a workplace and urging others to apply. Then came a copy of the internal Valve Employee Handbook, apparently intended to orient new employees, which outlines the company’s organizational philosophy.
Both pieces describe Valve as an employee’s wonderland. It is a company without managers, without hierarchy or bureaucracy, where no one tells anyone else what to do. Each person decides individually what he or she wants to work on. This freedom doesn’t just apply to choosing day to day tasks, but even in selecting and forming teams and projects. The desks at Valve are all on wheels, enabling workers to move to new projects as they see fit. Valve’s decisions aren’t made at the top and pushed down to underlings. Rather, the people of Valve themselves are their own managers, and they steer the company as a collective.
Valve views itself as a company of experts, “seniors” [Handbook, 38], where every employee is the best of the best. When everyone is competent and generally trustworthy, who needs managers? For game developers outside of Valve who care deeply about games, have strong opinions about how they should be made, and some skill in actually making and shipping them, this level of freedom, camaraderie, and trust holds strong appeal.
The Handbook hints that Valve has an annual revenue in the billions of dollars [Handbook, 42]. This remarkable wealth (Valve currently employs only about 300 people) enables them to take an annual, week-long, all-expenses paid holiday to faraway locations like Cancun and Hawaii. Employees are welcome to bring their families along. A remarkable section of the handbook encourages new employees not to fret about the extravagance of the office luxuries. “If you find yourself walking down the hall one morning with a bowl of fresh fruit and Stumptown-roasted espresso, dropping off your laundry to be washed, and heading into one of the massage rooms, don’t freak out. All these things are here for you to actually use. And don’t worry that somebody’s going to judge you for taking advantage of it—relax!” [Handbook, 19]. Employee compensation is, naturally, higher than industry norms. This generosity is fueled by the company’s unorthodox belief that “Valve does not win if you’re paid less than the value you create” [Handbook, 27–29].
Overall we get the impression of complete individual freedom, brilliant coworkers, and high pay—not to mention the chance to work on some of the greatest games and most important technologies in the world.
Valve sounds like the perfect company.
So we ask the obvious question. Is it all too good to be true?
Having worked in game development since 1995—the year before Valve was formed—I tend to be cynical about claims to workplace greatness. I’ve seen many an “ideal” company—Ion Storm springs to mind—turn out not to be as wonderful as they believed themselves to be. All companies tend to puff themselves up in their own internal dialogue. Managers remind employees that they’re all the best of the best, their past products were unquestionably awesome and their future is a never-ending path of gold. In some respects the praise may be true and well-earned; at other times, it is fantasy or—worse—a cover-up.
Is the glitter that shimmers over Valve’s handbook true gold, or the same old fairy dust?
As I read through the Handbook and Abrash’s column, I searched carefully between the lines for the reality behind the hype. And I came to a surprising conclusion—surprising to me, anyway.
On the whole, Valve is probably telling the truth. On the whole, mind you.
I don’t think their self-image is pure fantasy. I suspect there are dark spots they are not telling us about—perhaps that they’re not even aware of. But for the most part, yes, I buy Valve’s story. I suspect it really is a pretty awesome place to work.
Let me show you my reasons for thinking that 1) Valve probably is really pretty awesome and 2) there are probably some serious dark spots.
Why am I, the battle-worn cynic, willing to believe that Valve is as awesome as—or nearly as awesome as—they portray themselves to be?
The first part of my answer is easy. Just look at the products.
If Valve had only made Half-Life it would have been a remarkable achievement. But the games that have followed have been equally stunning, ground-breaking, and—most importantly—lucrative. They have made smart decision after smart decision. They bought the rights to Team Fortress and made it—albeit with much chicanery and long delays—into one of the great multiplayer franchises. Counter-Strike—need I say more? When Half-Life 2 was in development, the cynic in me said they could never top the first one. They did. When Steam was in development, no one but Valve believed that customers would be willing to pay for a game they could not hold in their hands, lend to their friends, or sell on eBay. They were.
Now something is enabling Valve to be smart when many other game developers and publishers continue to struggle and fall. If Valve says that it’s their flat organizational structure—”No managers!” “Everyone’s a leader!”—who am I to argue? It makes sense that a group of carefully chosen, intelligent, engaged, openly communicating, motivated people would tend to reach smart decisions—smarter, anyway, than a company where one or two or three bosses push through what they believe in. James Surowiecki’s book The Wisdom of Crowds lends some credence to this idea.
According to Surowiecki, the “openly communicating” part is crucial. It’s the part that enables a crowd to act with collective intelligence. When thoughtful people form their own opinions, then share their opinions in an egalitarian, open forum, they produce great ideas far beyond what any one of them could come up with on their own.
Most companies do not offer open communication, and the resulting decisions are often dumb—even tragic. In every company I’ve ever worked for, the employees in general—as a collective—knew better what the company strategy should be than the high-level managers themselves did. The employees knew what was wrong with the company, what was right with it, and in which direction it should go. If management had followed the advice of their workers, the company would generally have done better than it did by following the managers’ strategy.
Let me give you an example. I’ll have to be a bit vague about the details—you understand.
At one company, the management hired an outside firm—at a moment of crisis—to administer an anonymous survey to all employees. The employees came through beautifully, offering thoughtful and constructive suggestions about what projects to cut, what projects to advance, how to fix the compensation system, how to improve relationships and morale. But the feedback didn’t fit with management’s direction, and so all but a token few suggestions were ignored. The company’s flagship project—a deeply troubled effort—thrust ahead, while other, more promising projects languished.
So what happened in the end? Well, the flagship project eventually launched, after many years of painful delay. This vast effort by hard-working employees was met with a collective “meh…?” from critics and customers, and the company took a multi-million dollar hit. Later, one of the backburnered projects that employees had favored was given a proper budget and team and allowed to push forward. It—as the company at large had predicted—shipped on time, under budget, and was a remarkable, unexpected hit that sold millions upon millions of copies.
If these smart, well-intentioned, thoughtful employees had had their way, the company would have spared itself an embarrassing and expensive loss, and would have gained more quickly the wealth and kudos of a hit game.
So I, for one, am willing to believe that “The Valve Way” could really work—that it is working for Valve, enabling them to make some of the smartest decisions and biggest profits in the industry.
As one final piece of evidence, take a look at the Valve Handbook itself. Name one company that has a more cleverly-packaged, well-designed, well-written, illustrated, helpful employee handbook. And yet this production was created by employees working on their own initiative, under their own direction, simply because they believed it was important for the company’s overall health.
Or so we are led to believe.
But why shouldn’t we believe it? What is more likely, that a document like this emerged from an HR department, “tasked” by some VP, or from a group of motivated, creative people following their own collective gut?
So I believe that Valve, on the whole, is a great place to work. That’s the good news. What—and where—is the bad news?
You can find the bad news in the same place: between the lines of the Valve Employee Handbook.
How does anyone ever get fired from Valve? It’s a simple question, an obvious and important one, but the Handbook never addresses it. One section is titled “What if I screw up?”, but its primary message is that you won’t get fired for making bold mistakes. The chapter “How Am I Doing?” describes Valve’s peer review process. Once a year, Valve employees rank each other according to who they think is most valuable. The details of this process is absent from this “accidentally leaked” public document, but it’s clear that employee compensation is tied to overall rank. Makes sense.
Is firing also tied to overall rank? The Handbook doesn’t say. So the question remains. In a collective organization without managers, how do people get fired?
Let’s speculate. Perhaps each year, the bottom 10% of employees are fired outright. The next 15% up from those are put “on notice”. Anyone appearing in the bottom 25% three years in a row is fired outright, even if they were never in the bottom 10%. I don’t know—I’m just making this up. But it has to be something like this, right?
If it is, it sounds pretty dark. It’s similar to the approach used by glorious Enron, where it produced endless politics, backbiting, and skullduggery.
How would it feel to be fired not by a person but by a collective? Who would you complain to? Who would you appeal to, to explain yourself to? Who would you go to for support and sympathy? If Valve were to fire employees like this, it would be no paradise for many.
And what if everyone at Valve is sincerely and truly good. Do you still chop off the bottom 10% of your organization every year? Maybe you adjust it each year. Then who decides what that percentage should be? If Gabe decides, and decides to reduce it to 5%, isn’t that basically Gabe firing a certain set of people?
Then again, maybe 10% is too low. Maybe 25% of Valve employees are sitting back playing DOTA 2 all day while receiving hot stone massages and sipping high-end espresso. Maybe the chop bar should be 25% or 30%.
Okay, but all of this is speculation. Maybe Valve doesn’t use percentages at all. But then, I ask again: how do people get fired?
Because, you see, as nasty a business as firing is, it really is a fact of life in any healthy organization.
Companies sometimes hire poor employees. Sometimes good employees undergo personal changes—changes in motivation, changes in chemical dependency, nonstop World of Warcraft addiction (I’ve seen it all)—and become poor employees. Poor employees weaken the whole company. It doesn’t do anyone any good to allow a poor performer to languish in a job where they’re failing, hurting themselves and others. A healthy company has to let people go sometimes.
Or to look at it from the opposite direction, poor employees often improve, growing into strong contributors. They get help, they study up, they kick themselves in the pants, they change their attitude—who knows how it happens? People tend to grow, especially when motivated. Sometimes a “firing decision” is really a “saving decision”—the decision to give someone who’s struggling another chance to succeed.
How does Valve make these decisions? Even their internal documents ain’t telling.
Maybe Valve simply doesn’t fire anyone, or does so so rarely that they have no real process or policy. The emphasis on careful hiring that appears throughout the Handbook suggests that firing is difficult—perhaps unheard of.
Well that would be nice in one way, but I can promise you it would be no paradise. I’ve been on teams with plenty of deadwood—not naming names, don’t worry—and life is no picnic for anyone in those situations.
So there’s one dark spot. I worry that Valve’s self-image as a company of “seniors”, coupled with the need to make awkward, decisive, sharp decisions about firing leads to a company where firing is either coldly calculated and ultimately political, or else simply doesn’t happen, leading to a soft underbelly of lazy, incompetent deadwood.
Here’s another suspicious thing about the Handbook. Valve portrays itself as an open, freely communicating company, where employees are welcome to discuss anything with others. And yet the Handbook itself is remarkably tight-lipped. There is no mention of future projects—not even on the “company timeline” [p. 19]. The juicy details about employee rankings are hidden away. If the company is so open, why is the Handbook so formal, so careful, and controlled?
I suspect the answer to this question has more to do with Valve’s intention for the Handbook than its own internal culture. Come on: this Handbook was made to be leaked. It was intended for public viewing as much as for employees, for potential hires as much as for new ones.
Still, if everyone at Valve is a leader, involved in decisions at every level of the organization, it’s remarkable that they don’t leak more often. You don’t see Valve employees getting out and giving interviews much. You don’t see them blogging like Michael Abrash did about life within the company. His was the exception that proves the rule. For such an open company, they are remarkably closed to outsiders. And that makes me wonder if their internal dialogue is really as loose and uncontrolled as they portray. It just piques my skepticism a little.
I read an article one time—I don’t remember where—that said that the jokes that people tell within a company are the best hint to the truth about the company. I have found this to be true. The Valve Handbook ends with a set of jokes. What do they tell us about life at Valve?
“Gabe Newell—Of all the people at this company who aren’t your boss, Gabe is the MOST not your boss, if you get what we’re saying.”
I think we get what you’re saying. You’re saying that Gabe really is the boss—he’s the one holding the purse strings—but he’s careful not to impose his authority because he believes in the collective autonomy and decision making advocated throughout the Handbook. Don’t burst Gabe’s bubble by treating him like the boss. But everyone knows he’s got the authority. After all, it was Gabe who had the power and right to hire the whole Portal team on the spot after a brief interview. Everybody’s equal at Valve, but not everybody can hire whole teams on the spot. Another difference: you earn a salary, but Gabe owns the bank account into which Valve’s coffers flow. So Gabe’s your boss, but we all pretend he isn’t, because that’s what the boss wants. We leave it to you to work out the tension.
“Manager—The kind of people we don’t have any of. So if you see one, tell somebody, because it’s probably the ghost of whoever was in this building before us…”
The message: Valve does indeed have managers. Or rather, as revealed elsewhere in the document (pp. 16, 32), people sometimes step up—temporarily—to take on a lead role. The cynic in me wonders just how “temporary” these managerial roles tend to be. But I’m willing to suspend my skepticism and give the benefit of the doubt until new evidence arrives.
So I’m optimistic about Valve as an employer. Since I teach superb game development students who seek jobs at graduation, I’m interested to see whether they are drawn to Valve and what they report when they get there.
My own take is that Valve probably really is a pretty darned cool place to work. No place is perfect—people are people, after all, and every workplace has its disappointments, its tedium, its betrayals and mistakes. But Valve sounds like they’re trying a different approach, an exciting one, and one that seems to be working out well. For once, I’m not (quite) cynical.
Update 2013-07-09: Jeri Ellsworth was recently fired from Valve, and gives an interview that reveals some of the inner workings. It’s worth a watch (see about 11:00).