Behind the Screens: Q&A with the Technical Lead of Zynga’s Mafia Wars

CEOs tend to get all the credit, but as many of them would likely admit, it’s countless invisible yet talented individuals who make up the vast amount of value in Silicon Valley.

Beginning this week, we’re hoping to shine a light on some of these unexposed stars through a regular new feature called Behind the Screens.
Kicking off the series is 28-year-old Erik “Dutch” Vanbragt, the franchise lead of technology on Mafia Wars, an enormously popular multiplayer browser game from the social networking game company Zynga.

The son of a professional soccer-playing father and an athletic mother, Vanbragt says that his parents were “a little concerned” when instead of playing outside, their 10-year-old son attached himself to their IBM 286 computer and began teaching himself different programming languages. His obsession with coding never faded, though, and that fixation is paying off. After nabbing his degree at Carnegie Mellon and trying life at three other startups, Vanbragt now helps head a game studio that employs 65 people — all of whom are riding a rocket ship. With roughly $220 million in venture backing and annual revenues of nearly $300 million, two-year-old Zynga has become one of the hottest startups on the planet.

Today, Vanbragt and I talked about Zynga’s explosive growth, how its culture has evolved, and what most surprises him about Mafia Wars’ users.

What’s your day like? I have this idea that most developers wake up at noon and work until midnight. Is that anywhere close to reality?

On the Mafia Wars’ team, that’s definitely not true. I usually get in around 9:30 a.m. and most of the engineers are in by 10 or 10:30. We have a morning meeting so people sync up, then we get progress reports on what’s going to happen, we work throughout the day, then most people go home after 8 o’clock. Zynga has free lunch and dinner, so people try to stay through that. [Laughs.] But the days are long — 10 hours on average. A lot of people will go home and do a bunch of work after putting their kids to bed.

What’s the average age of a developer in your group?

Our team is 65 people in total, including product managers, developers, everyone. There are just 24 people on the engineering side, and they’re spread across two development teams, a feature development team and a systems team, which includes our server guys who run the back end and handle the performance optimization that no one can really see but that makes everything run. I’d say the people coding on the front end are around 25 years old on average; on the systems team, it’s probably closer to 30.

And what’s the culture like? Do you feel like an employee of Zynga or like an employee of Mafia Wars, given its size?

Like other game developers, we’re broken into different studios. So while we’re all part of the parent, Zynga, we’re all our own entity and we’ve built our own culture. We’re all very close friends. We do lots of team outings. We have a weekly soccer match and go out after work. It doesn’t feel like we’re going to work but coming to hang out with our buddies.

You were at Mafia Wars from pretty much the outset in 2008. How far along was it when it was released? Sixty percent developed? More? These games undergo lots of iterations, right?

No, when Mafia Wars was released, it was as finished product, not as a needs-to-be-iterated-upon thing. After the release, it was: now that we have this game, how do we expand on it? Since then, it’s been about getting user feedback, seeing how people in forums react, and making a better genre of a game that people have been making since the ‘80s.

What was one of the assumptions you made when you created the game that you had to change because of user behavior?

Well, for one thing, you want people to fight each other and to have healthy competition. We weren’t at all anticipating people griefing each other.

Griefing? I’m an old lady. I don’t know what that means.

It means attacking someone else nonstop. I’m having fun but I’m ruining this other person’s game play experience. We didn’t see that coming into this.

We also didn’t expect such explosive growth. We were in no way ready for the growth we experienced. We just didn’t have the systems in place for it, so we had to stop a lot of the development and reengineer our backend systems so we could grow. Otherwise, we’d still have less than a million users.

And instead you have…

Seven million daily users.

Wow, and is that success for you? Do you judge the quality of your games purely through user growth or do you have some other ideal that you’re shooting for?

We use a few different metrics around users, but also retention and how many installs come back every day. We’re always measuring: will a user come back in a week? Will he come back in two weeks? We also have metrics based on community involvement. We don’t want our customer service hotline blowing up with 5,000 tickets a day.

How do you retain people’s interest? How many features do you need to release — and how often — to keep users coming back?

That’s the battle. We try to introduce one new feature a week, but it’s usually more like one every other week. But every week we have a lot of code going out and we try to do a lot of iteration on what users like. We have a very robust backtracking platform, so when we see users gravitate toward certain features and use the game in ways we didn’t expect, we’ll try to enrich the experience for them.

We also have a community team that hosts forums and communicates closely with users and they really let us know what the gripes are so that, by the next day’s meeting, we can brainstorm out ideas and within a week actually get what users want out in front of them to see if they actually do like it.

Do you look at your competitors’ games?

Yeah, I play all my competitors’ games. I enjoy games in general, so I actually have fun playing their games. I don’t look to them for inspiration, but I’ll look at them for their mechanics.

I feel like we’re trailblazing a lot of the mechanics, but I like to look at the widest array of games I can and when there’s a game mechanic that we like, we’ll incorporate it.

To what extent do you guys rely on Zynga’s network to grow Mafia Wars? Is it possible to isolate the quality of the game versus the network effect?

Mafia Wars started out before we had a lot of the cross-promotion. So we’ve actually been one of the primary teams that has grown FarmVille and Café World. But now, yeah, we do get a lot of reverse traffic from them.

We do rely on the Zynga effect, the ability to drive traffic between applications, as well as enjoy decent organic growth.

Let’s talk about virtual currency, which has turned Zynga into a cash cow. When you launched Mafia Wars, did you expect that people would be so interested in buying virtual items?

Not at all. And we didn’t actual start selling virtual items until later, starting with a bloody chain saw and, I think, a flame thrower. Others had done it before, of course. The Korean gaming market has [been selling virtual goods] for a long while. We just sort of viewed it as something with low risk and a high reward potential. Still, [the success of the goods] was just shocking. We pushed them out late one night and we all got in the next morning, saw the revenue numbers and it was unreal. You just sort of had a grin across your face from ear to ear.

What’s Zynga like today? How does it compare with your past gigs?

My other jobs have been very different, very

based where there wasn’t a lot of love for the engineers. At Zynga, instead of being monkeys who are made to adhere exactly to a code that a PM writes, we can express ourselves. Engineers have the ability to push back on specs, to say, “Maybe this feature would be a little better if we put in this little twist.” It’s pretty great.