In-Depth: On Company Culture, Hiring, And Retention

Consultant David Edery delivered a fascinating talk about company culture at the recent IGDA Leadership Forum in San Francisco, examining the variant approaches of successful studios -- and suggesting that changing corporate culture is harder than it looks. According to former Xbox Live Arcade portfolio manager Edery, the recent trend -- as reflected by Electronic Arts' strategy with Playfish -- "is to leave acquisitions alone. It's not because they're lazy, it's not because they don't want to go through the effort... It's in many ways because it's the best way to go through the integration." What Kind of Company Is It? Edery presented research conducted at Stanford from 1994 to 2002; the subject was corporate culture and success at 200 tech companies. Many of the observations apply to the game industry very well, particularly when filtered by Edery's observations. 
 The study broke companies down into five different cultural types -- finding that 70 percent of companies hit these norms. Moreover, those that don't, don't do well. Star. Companies with a "star" culture rely on rockstar talent; people join the company out of interest in the work, and are selected for their potential, and are paid well -- the emphasis is on performance. Engineering. People also join this company for the work, but are selected for their existing skills rather than potential. Their peers are more of a contributing factor to how they work. "The attachment is to the work itself," Edery says, and there is "not a lot of bureaucracy." Commitment. "These are companies that want to be families," says Edery. "This is one where the founder might literally might say 'I want my employees to think of this as their second home' ... The loyalty is to the company. Why are people hired? The first and foremost criterion is do they fit in." Bureaucracy. These companies are known for their adherence to strict formality and have a lot of management -- people are hired for the skills and join for the work. "These tend to be driven by their process," says Edery, and notes, "People tend to think these are large companies but that's not true, some people found bureaucracies." Autocracy. Says Edery, "People are there for the money. People who run the show couldn't care less about their employees." For people who are attached to the work, they'll jump to another company for a better opportunity -- something you frequently see in the game industry. Commitment and Autocracy companies are the only two types where people don't join for the work -- it's for the love and money, respectively. Extant company cultures "don't change easily," says Edery. If a culture tried to change, even at the behest of its founder, "you instantly see two things happen statistically. The chances of failure more than double if you try to change your company culture. Your revenue drops something like 3 percent per month, cumulatively." Thus when starting a company, it's "really, really important to think about your culture -- you only get one shot." On that note, if you join a company "and you don't like its culture -- you're stuck," warns Edery. However, "You can improve it incrementally." The Effects of Type Now, depending on what type of company you have, there are drastic effects. Bureaucracies, predictably, spend the most on managerial overhead -- 30 percent of the company is management once it reaches 500 workers; in a commitment company, that figure is 12.8 percent. The figures range from 14 percent to 20 percent in other types. "This can dramatically affect your costs," says Edery. And the management requirement is consistent with the founder's cultural map -- even if he or she is not the current CEO. 
 There's also a major variance in the likelihood of failure. An Autocratic tech company is 150 percent more likely to fail than an Engineering-based one, which is used as the baseline for these statistics. "The 'you work, you get paid' [model] doesn't work in any kind of knowledge industry," says Edery. On the other hand, Commitment companies are twice as likely to avoid failure, followed by Star and Bureaucracy. Commitment companies, in tech, have the highest chance of IPO -- 200 percent, when compared to Engineering (again used as the baseline.) Bureaucracies have a small chance; Autocratics a tiny chance, and Stars have a negative-50 percent chance. But if they do IPO, Stars have the biggest chance of growing in market capitalization. Suddenly, and for the first time, Commitment companies are at a disadvantage, with negative growth. "Why is this?" Edery asks. "If you are a family culture, and you really care about your people, it makes it hard to grow fast. When they do try to grow too fast, that's where you see a collapse." On the other hand, Star companies swing for the fences -- so they will go far, but they might strike out. The stars at star companies "have to be a good manager, or person to work with, because so much authority rests in him," says Edery. "They need a really great team, and that really great team needs to be happy being overshadowed by the star. There's a misconception that everyone who is not a star is shushed up," he continues. "It's important that the star get feedback, and it's important for the star to be able to take feedback." Two Quick Case Studies Edery interviewed developers at game companies to try and ascertain what kind of culture they have. While he emphasized that he has not worked at these companies and can't be certain of their culture -- "it's just a guess and I could be wrong" -- these are his educated guesses. Edery believes that there's a high chance that Blizzard is a Commitment company. Three quotes he got from developers cut to the core: in recruitment, said one interviewee, "we want ninjas, not rockstars." In reference to a colleague, one remarked, "that guy bleeds Blizzard blue." And when it comes to retention, one said, "employees who have been here five years still feel like newbies." Edery spoke to a Blizzard employee who said that if Valve offered him double his current salary, he would not leave Blizzard. Edery, conversely, thinks Valve is more likely to be an Engineering culture. Some quotes: "No one has a manager and there are no producers." The famous development philosophy: "absolutely everything is driven by playtests." And when it comes to new hires, an employee asks himself, "would you split your bonus with this person?" That's not a sign of a Commitment culture, most probably, suggests Edery. Though it's probably not a Commitment culture, Edery was quick to point out that's not a negative. He says, "I know the people who work for Valve really like it. It's clearly a great place to work that does great things. I don't want people to see these slides and think I knock companies that aren't Commitment companies." The important thing to take away, though, is that Commitment cultures really do work. Says Edery, "I've talked to several people who have this desire to start a company that's like a family, but they've talked to people and been told it's not a good idea." The primary reason is that, in the game industry, layoffs are a big issue. "That's a bunch of BS," says Edery. "There are lots of other good examples out there, outside the game industry." He also says that many people he talked to referred to the perks that their companies offer -- like sodas, vacations, and more -- as evidence that they're Commitment companies. However, says Edery, "perks are not the sole key to a Commitment culture. Perks are a mask for what the true culture is. The true way to know if it's a Commitment culture is if, you took away the perks, would people stay?"
 Again, says Edery, change is unwise. "That's really going to hurt. Improve incrementally. Identify what your culture is and maximize its strengths." Recruiting For Cultural Fit So how do you recruit for a cultural fit? That's a problem -- particularly as the best way to do that is to observe them in action. "We put people in interviews -- we test their ability to interview well. We all know this." One great way to observe someone in action is to not hire them initially -- if you can, take them on as a contractor "and don't do it abusively" -- pay well. That's not going to work due to job security, in many cases, but it's a possibility. Another option is to observe through eyes of others -- in the form of a reference check, something Edery says "people tend to dismiss." Reference checks can obviously be gamed; many people put down friends, rather than their bosses, or point you at dead ends. Some companies don't like to give honest references, due to liability issues. But that's "not an excuse not to try. If you make just a bit of effort, you might reach someone who will give you an honest answer." One tactic Edery has seen to circumvent that involves "intentionally call[ing] the reference when he thought he'd be away and get voicemail." At that point, leave a simple voicemail -- "If candidate XYZ was an excellent employee, call me back." No call equals not an excellent candidate. Google also tried a recruiting game -- it put up a billboard with math riddles near engineering schools; it tested high level mathematical engineering skills. When the answer was fed into Google, it led to another puzzle, on and on -- with the ultimate goal being an application. In the game industry, a good example is BioWare, says Edery, which "has a history of hiring people who use their tools to make content for their games." And Edery also points to Valve, which recruited the entire Narbacular Drop team from DigiPen to make Portal, knowing they could effectively work together. A mission statement can help clarify things, too. Harmonix's is "we invent new ways for non-musicians to experience the joy of making music." If you do not want to do this, says Edery, "you have no business working at Harmonix." Employee Retention Now that you've found your employee, how do you keep him or her?
 When it comes to compensation, bonuses are "not such a great way," it turns out, says Edery. The best thing to offer, rather, is fair compensation that people don't ask for more of. You should be "surprising the employee pleasantly with raises. They shouldn't have to ask for it." Netflix keeps employees at the maximum the company would be willing to pay for an employee. They're less likely to leave, and if they are, HR knows they're already at max -- so there's no real way to counteroffer. And if you do want to give bonuses? A smaller bonus, given unexpectedly, "is much more effective than giving at the end of the year that that they're expecting." It should be tied to performance, still -- but maybe for success on a specific project or task. One of the most "underleveraged" ways to keep employees interested, says Edery, is challenge -- important particularly for the three types of companies that rely on interest in the work. When it comes to employee feedback, says Edery, "a lot of companies are really bad about this." Those that give it one or two times a year are "a disaster", he says, particularly new employees who need to find out how they're doing quickly and frequently. "The best way to drive them away is to find out you're not happy with them." It's important to get rid of the bad apples, however, no matter your culture. Is it counter to a commitment company? "Absolutely not," says Edery. "It doesn't help a commitment culture to keep bad apples around." Bear in mind a bad apple is not someone who is having a hard time -- but someone who is a legitimate negative performer or influence. Dealing With Resignations When it comes to resignations, says Edery, "at most companies, they tend to assume they've lost them at that point," and only make "some token effort" at retention. But Edery thinks this is foolish: resignations merit more attention than they usually get. "You should be more thoughtful," he says. 
 Edery sees resignations as very costly -- you've lost your investment into that employee, you've got to hire a new one and spend on their behalf, and you've got to wait for that employee to ramp up to full productivity, which is a drag on resources. One tech company, Cypress Semiconductor, has a lower than 50 percent successful resignation rate. "When their valued employees resign, they claim they convince more than half of them to stay," says Edery. The manager, once he or she hears of the resignation, will cancel whatever appointment he or she may have -- including a meeting with the CEO. The resignation is kept quiet, because "public commitments we make are dramatically more important than private commitments," says Edery. After the meeting, the manager tells everyone in the management chain -- all the way up to the CEO. During these meetings, the goal "is not to convince them he's an idiot for leaving, or why they should stick around. You're literally there to hear what they say and accept it." Then you create a retention plan and "do whatever it takes to win." If it doesn't work, the employee goes up the chain, one by one; 50 percent of resigning employees end up in a 30 minute meeting with the CEO. Getting them to say they'll stay "is so incredibly powerful," says Edery, because people do not like to renege. Don't worry about justifying the move to them; they will justify it to themselves once they've made the decision, he says. The last step: wipe out the competitor. Get the employee to call and reject the offer.