Here is the Gabe interview itself; and here is their awesome tour of the Valve offices. But we can’t talk about this interview, or the tour, without first bringing up one of its most important revelations: Gabe Newell now has a beard. Take a look, in case you didn’t see it up there:
I have to admit – he looks quite badass! It’s as if Future Gabe stepped out of a time machine from the year 2022, bringing tidings of… I don’t know, Dear Esther 2. Maybe it’s so amazing it grew Gabe’s beard for him. It does look a bit unkempt, but he’s looking pretty cool.
But, enough with the beard-talk – this isn’t a Bethesda game. This interview is absolutely terrific, so we’re just gonna blaze through it with some of the big highlights. Not all of it, so if you wish, you can simply check the actual interview out here. Anyways, let’s get started:
Ben Kuchera: It’s been said you’re not much of an early riser. Can you walk us through a day of your life at Valve?
Gabe Newell: It used to be that I was the best multiplayer gamer here, and then we hired someone. It’s been downhill ever since. Just about anything I do, there’s someone here who does it better. Whether it’s game design, or business issues, or writing, or programming, or anything. I do a lot of things, but I tend to do them not as well as other people, so I tend to fill in holes as much as anything. If there’s something that needs to get done that’s not getting done, that’s what I end up doing. I don’t really have the same day, my days tend to be more reactive than many people’s.
Ben Kuchera: That seems tricky for someone in your position. Is it hard to keep track of what’s going on in the company?
Gabe Newell: We’re a very flat organization, so we expect everybody to manage themselves. One of the things that people have to do when they’re here, they need to know when to broadcast to me when something’s important. [...] Most of the time when information is being distributed it’s because someone decided it was important and people needed to know about it. So the issues of keeping me on top of everything is more of a question of people deciding that getting me involved would be helpful for something they’re working on. It’s not my job to track down or supervise everyone and make sure they’re doing the right thing. It’s more like I’m a resource here and I can be helpful at addressing a problem. [...] We’re optimized for people who are very experienced and have been working for a long time and don’t really need someone looking over their shoulder or second-guessing their decisions.
That sounds pretty damn neat. Valve really sounds like an amazing place to work at! Imagine – it’s up to you what you work on. All of the work you do is stuff you really enjoy doing, and feel bonded with. It isn’t just “work assignment #1″ and so on, ad infinitum until your employment is terminated.
Ben Kuchera: Is there anything you’re seeing in terms of the inputs or outputs you’re experimenting with that you’re personally jazzed about, or you think that is really bearing fruit, or will in the near future?
Gabe Newell: Yeah, I mean there’s a surprising amount going on with new – they used to be called wearable computing before those all got kind of set on fire by losing investment firms hundreds of millions of dollars, so nobody wants to call them wearable computing, but they sort of look like the old wearable computing solutions, the difference being that they’re way higher resolution, way lighter weight, much better battery life, and things like that. It seems like just about the time that everybody gave up on them they actually started to become interesting, so we’ve been seeing a lot of stuff go on in that space that gets us excited. We’re trying to get our–the experiments we’ve been doing in–you know we did a ton of work on biofeedback, on biometrics, and that’ll, you know, from our point of view we were like “okay, this is all sort of proven out” and we’re just sort of scratching our heads trying to figure out the best way to get that hardware out to customers without something where we’d just say “okay, this works.” it’s not a question of whether or not this is going to be useful for customers, whether or not it’s going to be useful for content developers, you know, it’s figuring out the best way we can get these into people’s hands.
So we’re thinking of trying to figure out how to do the equivalent of the [Team Fortress] incremental approach in software design and try to figure out how would you get something similar to that in the hardware space as well. The sort of old method of, you know, let’s go make a giant pile of inventory and hope that some set of applications emerge to justify this giant hardware investment doesn’t seem to be the – very consistent with what we’ve seen to be the fastest ways to move stuff forward, so we’re trying to come up with an alternative to that that gives us the ability to iterate more rapidly. That stuff we’re like “this is good,” now we just need to figure out how we can start giving these to customers and iterating on the design quickly enough without having to go off and buy ten million of them and then find out we did something mildly stupid and then having to throw them all away and start over.
Valve hardware? Interesting – especially for what they’re doing in the bio-feedback department. If they succeed in making that kind of hardware available and customer-friendly, then bio-feedback could become far more than a playtesting tool. We could see a new way to make games: “bio-gameplay”, if you will. And it’d finally be a great opportunity to play those prototypes Valve made for their internal bio-feedback testing.
Perhaps a Valve console of sorts is no longer looking so unlikely!
Ben Kuchera: Now do you see a future where Valve is actually selling hardware or do you just want to have things that could take advantage of that technology should it be popular?
Gabe Newell: Well, if we have to sell hardware we will. We have no reason to believe we’re any good at it, it’s more we think that we need to continue to have innovation and if the only way to get these kind of projects started is by us going and developing and selling the hardware directly then that’s what we’ll do. It’s definitely not the first thought that crosses our mind; we’d rather hardware people that are good at manufacturing and distributing hardware do that. We think it’s important enough that if that’s what we end up having to do then that’s what we end up having to do.
Interesting. Quite interesting. Did I mention how interesting it is?
Ben Kuchera: Now we’re talking about things like wearable computers and biofeedback and we’ve touched on Steam a little bit but there’s millions of people out there who are gnashing their teeth when they read this who just want to play Episode 3 or Left 4 Dead 3. Is there ever tension between all the different things that Valve is interested in doing?
Gabe Newell: Oh absolutely. We’re acutely aware of how much we annoy our fans and it’s pretty frustrating to us when we put them into that situation. We try to go as fast as we can and we try to pick the things that we think are going to be most valuable to our customers and if there’s some magic way we can get more work done in a day then we’d love to hear about it, but we recognize that it’s been a long time whereas we have so many games that people really love–Counterstrike, Half Life, Portal, Left 4 Dead, not a whole lot of Ricochet enthusiasts out there, and at the same time we want to be making sure that those games and those stories and those characters are moving forward while also making sure that we don’t just get into terminal sequelitis.
But we’ve always somehow, you know, part of the reason that we backed off talking so much about what was happening in the future is that when we’ve done that in the past, you know, with Half Life 1 it was a year after we originally said it would be, Half Life 2 basically if you go and read the forum posts apparently took us fifty or sixty years to get done so we’re trying to be careful not to get people too excited and then have to go and disappoint them. So we’re sort of reacting in the other direction and saying “okay, well let’s have things a little more baked before we start getting people all excited about it.”
Gabe makes a good point, but no one’s going to be disappointed by learning the thing’s actually in development. If anything, the silence only increases excitement and hype to levels Valve themselves might not manage to exceed. We don’t really know it’s on their agenda to begin with. A simple “we are working on it, we’ll give more details soon” would suffice, and that’s all that the fans are asking for at this point. No one wants the game to be released 3 months early, and no one wants a reveal trailer. Just a simple heads-up so we can all relax knowing that Valve hasn’t given up.
Moving on to page 2…:
Ben: How do you sit down and try to extract the data about which players are going to bring value into the game without almost embedding people into vent servers to listen to how they interact?
Gabe: Well, we all play games all the time. It’s not Valve who will be saying that somebody’s more or less valuable, right? You have to create and design systems where the fact that somebody is valuable to some other group of people is discoverable by somebody else. The fact that you’re fun to play with is just one of many, many different ways that people can potentially create value in this kind of environment. The fact that I found that somebody was fun to play with needs to be efficiently and transparently communicated to other people who like to play with people who are like people I like to play with, which is totally different. You might find that person incredibly annoying and exasperating and not at all interesting to play with, so if you can create a system in which both of us end up getting the greatest amount enjoyment out of our experiences then you created a lot more value collectively, right? Each person is going to have a very different weighting of a bunch of different values, choices, preferences, and getting you connected to the right experiences with the right people with the right content is going to be a characteristic of a more successful systems, and just sort of treating everybody as if they’re exactly the same and only viewing them as opportunities to extract a retail entrance fee is gonna seem very archaic, I think, fairly soon.
Very well said. Let’s see what else he’s thinking on that front:
Ben: Now this all sounds wonderful and it makes sense, but what kind of concrete mechanisms can you put in place that allow people to express who they like to play with and how they’re either giving or gaining value from the system?
Gabe: Well, one of the things you have to do is start to come up with metrics, so we look at huge amounts of data all the time and then we try to figure which one’s predictive and which ones aren’t. So if you look at – here would be a simple example– so one of the decisions we made with Portal 2 was that we were gonna care a lot about whether or not people actually completed the game. And once you start paying attention to it and start changing stuff, you know, we made a bunch of changes after Portal 2 shipped that significantly increased the number of people who actually were completing the game with, I think it was – we had some window like 30 days, so as soon as you just simply start paying attention to some metric and start making changes based on that metric and then measuring those results, you tend to get somebody somewhere fairly quickly. So if you were looking at people and saying: “let’s not really care about who, let’s just take a trait-agnostic approach to this and just simply view people as clumping together,” we’re not really sure why, some people would say: “oh, I can assign various traits and characteristics to these people and I’ll just start measuring those things directly.”
The problem is that if your framework is wrong or if they don’t actually matter, you end up being screwed. So instead you just sort of say “let’s just see – let’s pick some behavior that we can measure and then start to look for correlations that we’ll then try to test for causalities.” So if somebody comes into a server who are the set of people that tend to stay on the server and who are the set of people that tend to leave. And then let’s look and see if that’s random or if it’s correlated, in other words if there are groups of people who tend to stay together and other groups of people who tend to avoid those people or leave or be mad. So then let’s use matchmaking and then say “so we’re going to put these people together,” and in fact what they should do is that they should all be more likely to stay through the entire match, and they also should be more likely to play again within the next week. And then you make those changes and you group those people together and you find out no I was totally wrong, right, none of those people actually did it.
Then you go, okay, there’s something about our model here that’s not working and we need to figure it out, and then you tinker with the model until you say “good, now we can actually with a lot of confidence say that there are groups of people and when we cluster them together they tend to play longer and they tend to play more often.” That way we don’t have to interview them, we don’t have to read their chat logs, but we probably made a set of changes that those people would perceive as being fairly valuable, so that’s just like a really simple example of the kind of thing that you would do without being particularly intrusive into the relationships yet would still probably have a pretty positive outcome for people in the community.
I guess now we might know why Valve made the move to lobby matchmaking!
Ben: There’s also this huge conversation going, especially now with rumors of the new consoles, about used games and piracy. Do you feel like you’ve kind of successfully sidestepped those issues with Steam as a service provider?
Gabe: You know, I get fairly frustrated when I hear how the issue is framed in a lot of cases. To us it seems pretty obvious that people always want to treat it as a pricing issue, that people are doing this because they can get it for free and so we just need to create these draconian DRM systems or anti-piracy systems, and that just really doesn’t match up with the data. If you do a good job of providing a great service giving people… as a customer I want to be able to access my stuff wherever I am, and if you put in place a system that makes me wonder if I’ll be able to get it then you’ve significantly decreased the value of it. So, you know, people were worried when we started using Steam initially because, oh my gosh, if I don’t have my discs what happens when I get a new machine? And after they’ve done this a couple times they’re like “oh my god, this is so much better, I’m so much more likely” – you know, this isn’t a legal argument, this is a real world argument – “I’m so much more likely to lose my discs than I am to have any problem with my Steam account, that seems way better than having a physical token that I use to access my content.”
A lot of times the systems that are put in place when you’re just trying to punish your evil customers for maybe doing something that’s not in their terms of service end up driving people towards service providers who don’t, right? So, you know, if I have to wait six months to get my Russian language translation and where I can get at this other guy on the street who will give me my Russian translation right away, it seems pretty obvious when you talk about it in those terms how the pirate selling pirated DVDs has a higher product than some of the people who try to DRM their way out of not giving customers what they really want.
Even more on Valve’s stance on DRM:
Ben: If you’re going to sell a game on Steam, has there ever been a temptation by you to kind of create a standardized set of DRM and holding publishers to it, or saying this kind of thing is inadmissible but will allow these certain solutions? Have you ever been tempted to get more hands-on on what kind of DRM is offered through what amounts to your storefront?
Gabe: We tend to try to avoid being super dictatorial to either customers or partners. Recently I was in a meeting and there’s a company that had a third party DRM solution and we showed them: “look, this is what happens, at this point in your life cycle your DRM got hacked, right? Now let’s look at the data, did your sales change at all? No, your sales didn’t change one bit. Right? So here’s before and after, here’s where you have DRM that annoys your customers and causing huge numbers of support calls and in theory you would think that you would see a huge drop off in sales after that got hacked, and instead there was absolutely no difference in sales before or after”. You know, and then we tell them: “you actually probably lost a whole bunch of sales as near as we can tell, here’s how much money you lost by bundling that with your product”. [...] I also tend to think that customers don’t really like it when you try to impose rigid rules on them as well, so we tend to think and hope that over time people will move towards doing the things that are in the best interests of both the customers and the content developers.
You know, it’s a really bad idea to start off on the assumption that your customers are on the other side of some sort of battle with you. I really don’t think that is either accurate or a really good business strategy, and so we just sort of keep trying to show – you know, I think that we have a lot more credibility now with developers on issues like this simply because there’s so much data that we can show them where we say look, we’ve run all of these experiments, you know, this has been going on for many years now and we all can look at what the outcomes are and there really isn’t – there are lots of compelling instances where making customers – you know, giving customers a great experience and thinking of ways to create value for them is way more important than making it incredibly hard for the customers to move their products from one machine to another.
Terrific stuff right there.
As for page 3 of the interview… well, not much else on that one, and I’ll leave you to discover what’s left. For we have greater business to attend to – the PA Report’s visual tour of Valve HQ! Including Valve’s great wall of magazine covers; Valve’s great wall of games; Valve’s coathooks; Valve’s hats; Valve’s awards, and quite a few more that’ll make you wish you ever set foot in that marvelous place.
In concluson, cheers to Ben Kuchera and all the folks at Penny Arcade Report; and of course, cheers to Valve, for having a workplace that’s so awesome we get drooling just looking at a few pictures of it!