computer games conferences games design GDC 2008 massively multiplayer

GDC08: Social Media, Virtual Worlds, Mobile, and Other Platforms


Speaker: Peter Marx, Analog Protocol/MTV

Good to hear about virtual worlds and MMOs from the perspective of a mega content / media company. Several interesting ideas and explanations that are well worth reading if you haven’t already been tracking the way that Viacom et al have been approaching the online socializing space.

Nothing fundamentally new, but the ideas presented were clear and consistent – and I’m kicking myself for not having tried VLES sooner, it sounds fun.

My own commentary in [ square brackets ], any mistakes/misunderstandings my own fault :).

Avatars are everything

The only way to greatly change the number of active / concurrent users is to make the avatar creation process as simple, as quick, as easy as possible.

Make Me is a very fast, simple way to make an avatar and quickly send it to other people.

Second Life (SL) has a very powerful avatar-creation system, but it simply takes too long. You can (if you want) make something that looks just like you in real life.

VLES is not very realistic, but it’s simple to use and quick. It’s an aspirational model, something you’d feel comfortable with being that image, and this increased the number of people who stayed, the number of people who were active.

So: lesson to people making VWs: spend good money on the avatar-creation tool.

How avatars move, what they do, what they say is very important.

We’ve found that less is a little bit more.

Sliders and checkboxes are NOT fun, don’t use them for avatar-creation.

We have a lot of dancing avatars, a lot that anim-emote. We found people love “moving” their avatars in some sense. At the end of the day, chat and speechbubbles isn’t exciting – we already do so much plain text chat (email, forums, SMS) that plain speech doesn’t interest us much.


Having big, expansive worlds isn’t that interesting. An iconic representation of a place is far more important – the icon is enough, e.g. the hollywood sign is all you need. you don’t need to recreate any actual part of it; don’t create every sidewalk, every parking meter.

[Adam: you need the abstract representation, the stuff that people remember in their mental model – the tarmac is not part of your mental model of “what is hollywood”, even though you’re aware it’s there]

So: when choosing scale of size of world, always veer towards making it small.


We’re a content company.

If you can somehow become a part of the conversation that people are doing on mobiles, then media companies become part of the equation.

Media companies want to see avatars interacting with content. Branded content especially. Hence things like Laguna Beach, and Pimp My Ride.

They do this because they want lots of the conversation that takes place in the VW to be conversation ABOUT the branded content. If, in addition (not instead), you can do the advertising part as well, that’s a bonus.

It’s an advertisers dream if you can turn the user’s own living room to be an advertising space, by the fact that you’ve managed to get them to continue talking about it.

Media companies fundamentally aim to be “part of every conversation that takes place about every thing”.

e.g. many movies obviously would make good VW’s, without much thought. Some might not be a good medium for advertising, such as Pirates of the Caribbean (PotC), but you can do SOME advertising.

OTOH, in MTV all the VW’s themselves are the advertising. The whole thing is to expose you to the advertising, and get you thinking and talking about it.

Challenges with content:

Relevance. Getting a bottle of Coke into WoW is not easy.

Expensive. Media companies circumvent this by shifting the burden to some other point in the value chain, e.g. to the consumer (i.e. UGC). A ratings system is a form of UGC that continues to push and promote conversation.

[Adam: this makes ratings systems disproportionately important compared to their absolute social value: they preserve the conversational aspect, and grow it. They get hugely underrated by games developers, typically, IMHO]


All VW’s have a virtual currency. They do this so that the VW owner can pay the users to participate in the world. The more that people get paid, the more they are going to hang around in this world.

One of the AP’s at EA found a UO character for sale on eBay. No-one had anticipated this secondary market back then. Today, that is an instituionalized part of VW design and operation. Every time you play a game in Neopets they give you a couple of bucks.

Challenges with economies:

They get out of balance. The rich get richer, the poor get poorer.

Work is extremely boring.

[Adam: for some reason, he doesn’t mention RMT here; whether you think it’s good or bad, it’s certainly a “challenge”]


This is the key to having long-term retention of your users (at least according to all the game designers I know).

How do you create games that can bring in casual gamers? The 18-24 year olds who are just casually coming along, hanging out for a bit, consuming some advertising, then going away again, perhaps indefinitely, but eventually coming back.

The metagame in the MTV space, something that people feel they are a part of, something that gives them some progression, is really hard.

[Adam: actually, they could achieve a lot here with ARG-style shared global questing IMHO]

[here he talks a bit about a concrete example of a game they’ve got in beta – I think it was VLES – to illustrate the metagame points, but I didn’t really see how the stuff he described satisfied the problems he’d just outlined. Sounds like fun though, I’ll have to try it :)]

You can make a game out of choosing a role: artist or fan.

Only a limited number of fans can attend any band performance, where artists instead fight the challenge of getting more attention than other bands.

Currently in early beta.

They have created a series of layered metagames on top of each other, to try and increase long-term retention.


User experience is determined by distribution network, and user’s platform (client platform).

You can spend an awful lot of money creating your world, and yet still be limited to the lowest common denominator which is the [last-mile] PC that the user is using.

There are PC’s being sold that have practically no RAM, but are very very cheap – but hence are more likely to be bought by large numbers of people. They then install Vista on it (!) that makes the machine almost incapable of running anything substantial.

It doesn’t matter whether you’re doing mega MMORPGs or just things aimed at tiny mobile screens: either way you have almost no control over the end device, you can be scuppered at the last minute.

[Adam: i.e. the stereotypical disadvantage of PC development compared to standard console development]

You shouldn’t make software that needs installing: teenagers are often using computers that they don’t have administrator rights for. Even recently at the company, I couldn’t let a high-ranking person try some software because he didn’t have admin rights on his own PC. We’ve found that 60%-70% of people who download the client never log in to the world.

[Adam: this is one of the main things Runescape got right, and I’d hoped it was common knowledge by now, but perhaps not]


Everyone I know wants to make a really powerful, immersive, rich 3D virtual world. That’s the dream of a lot of people.

But even in 2008 the technology is still very much the limiting factor. The reason Facebook UI is so simple is so that it works so well on everything.

But as a content owner, you have to put powerful rich media out. e.g. LoTR doesn’t work at all for simple 2D animations, for flat visual experience.

There are stories that people want to tell that require high-end graphics.

[Adam: this left me wondering whether people who wanted to tell stories using high-end graphics were going down the right path or whether they should be focussing instead on finding ways to tell their stories that didn’t need the graphics. In contrast to the game-design talk this morning which demonstrated that good graphics are often a fundamental part of gameplay, my feeling is that “rich” visual media are never a required part]