computer games conferences GDC 2008 web 2.0

GDC08: The BioWare Live Team: Building Community through Technology


Speaker: Derek French

Given the title, this talk came far short of my expectations. At the end of the talk I also felt extra annoyed that it felt like half the talk was just waffle, mostly towards the end with lots of repetition of the same vague opinions over and over again.

HOWEVER … when I came to clean up my notes and post them here, I realised that there were a lot of concrete good points, and it was just that it got waffly at the end.

If you don’t bother reading everything below, there’s one thing I want you to read (NB: I have cut out big chunks of the talk where the speaker waffled too much, so the reading below should be information-heavy).

“We have a live team: you need one too” – he said this right at the end, and this is for NORMAL (non-MMO) games. Commentary in-line at that part of the talk, but I think it’s a very very good thing to be saying and for games companies (publishers ESPECIALLY) to be thinking about right now.

Or … don’t. It’ll make it easier for us in the MMO space to take away your customers in the offline space if you don’t, so I’d be happy with that.

Think of this as: “Introduction to using and abusing your community tools for a game”. Maybe it’s just because in MMO development most of what he said became common knowledge a long time ago that I wasn’t impressed, but that doesn’t change the fact it’s got some good points and useful to anyone who hasn’t done this stuff before.


Multiple accounts per email address was a mistake. Screws up your metrics, stats, datamining.

[Adam: this makes no sense to me at all – this shouldn’t interfere with any of metrics, stats, etc, unless you made a bad mistake with how you designed the concept of “identity”, “authentication” etc in your rools]


One of our cardinal rules: if you ask to be a moderator, we deny. The kind of people who want that kind of power are the bad mods. So we just look for people who post good posts consistently, are polite, helpful, etc. We have no requirements for them to work at given times, etc.

Forum constraints

One thing we took further was e.g. for swear filter certain words we didn’t want just to censor, we had to add excpetions. We found the town of Calimshite, and Cockatrices, caused lots of problems.

We don’t just use swear filter for bad words, we use it to cut out gold selling spam, etc.

We used a points value for the words, and if your post scores 10 points or more, we junk it immediately. So then people would go back and re-edit to optimize it.

So, then we banned you for X hours where X = number of points in your score. See you in the morning!

e.g. when Australian release was a month later than everywhere else, someone complained with great venom, got banned for about 700 hours…

Interfering and accepting

With good, consistent, fair, safe rules and constraints and moderation of your forums, you’re going to make your players love to be there on the forums.

We allow people to disagree with us, to criticise us – so long as they do it respectfully – because that fits within the community, it doesn’t damage the overall value of the forums. We want everyone to feel comfortable there about talking about our games.

Patching and updaters

For NWN, we knew we’d need some kind of patching system. We felt we didn’t want to push patches because of all the modding. So, we built a custom updater where everything was in small chunks and people could manually trigger it.

Even with auto-patching you always have to have a manual update system for the cases where something goes wrong, where you make an accident with the patch or packaging.

Many interesting ideas people have about monetizing the updater, or requiring you be logged-in to community site etc – but the truth is if there’s something broken that MIGHT be fixed by the current patch, you want to get right out of the customer’s way and let them get the patch as quickly and effortlessly as possible, or they’re probably going to get even unhappier and quit.

NWN launch

Dev team took some time off before starting expansion patches, time for live team to take over and get going.

Started using the forums and bug-reporting to gather info on the game, triage issues, and sending bugs etc back to dev team. After doing this for a while found that our own live team programmers were able to do their own builds, their own updates, and work in parallel with the dev team.

[Adam: again, this sounds weird to me, because in the MMO world “live team” has a very specific meaning and there’s a much better handover between dev team and live team: usually the live team has so many of the original dev team members initially that they are largely indistinguishable until people gradually move to other projects. Interesting to think about whether / how soon traditional game-dev teams will adopt the MMO model of merged dev/live teams…]

99% of UGC is crap

[Adam: not the speaker’s summary for this section, mine]

In early days we found the community was making a lot of almost cookie-cutter modules: entrance, monster, treasure – and that’s it. So I took a designer on our mod team and asked them to make something crazy, do something unexpected, to demonstrate exactly how powerful Aurora was, to kickstart the community creativity.

He made a M:tG clone, and a bunch of other things, about 8 different really creative modules in a month. And then we suddenly saw some really interesting and creative modules coming out of the community.

Supporting a creative community

There were several times we thought they couldnt’ do stuff, and they did it.

creatures – too hard, but they did.
tilesets – they have no idea what the file format is, but they reverse engineered it
nothing of grand scale – someone made a 300 foot scale module, and we were floored by it, it was incredible.

[Adam: obvious. Not worth mentioning. But …]

So … we did anything we could to support them: gave them proper fileformat definitions, provided data, documentation, etc.

People complained there was no database, so we put in a simplified database layer.

[Adam: anyone seeking to do modern community / web 2.0 / social networking systems would do well to look at that list and think carefully of another dozen or so items that would hide inside that “etc” – and I’d like to see a talk where someone focussed on that entirely]

Overall, quid pro quo with community: they find needed features, we developed them. We’ve released 21 full updates across 3 OS’s, worked really well overall.

You can’t mandate tools

How do community create content? With YOUR tools? Maybe.

We had to put in more error checking than originally expected. (we expected them to use our tools, community went and wrote their own, so the content wasn’t going through our editor’s built-in error checking, so we had to add stuff to verify content that was crearted externally)

Non-game utils

We noticed a mistake: whenever we had gone to make an autorun, launcher, installer, etc, we had basically passed this off to a junior tools programmer, usually in last 3 months of project. What we ended up doing was having the exact same mistakes were made every single time.

I got frustrated because we were repeating bad mistakes, so the live team took over authority for it. So, now we can make this stuff in a few days for a new title, instead of a few months, and we get much richer AND more stable supportive tools for launching games etc.

Launcher is the FIRST THING customers see. That’s the first moment you have to excite or iritate your customer. It’s not something to leave to end of game, or have as low priority. If anyting goes wrong, they enter the game with a bad feeling.

[Adam: when a player first has contact with a game, their opinions are very fluid, they can change rapidly from impressed to horrified and back again. But once that first interaction is past, the opinion begins to cement, and it’s harder and harder to make big changes in player opinion. Only, I don’t buy that the LAUNCHER experience is going to so materially affect the opinion of the game – playing the game is a NEW experience. So, basically, I agree entirely with the sentiment the speaker expresses, but disagree entirely with the particular claim he makes here]

[Adam: I cut out lots of boring notes here, this was the waffle part of the talk, with sections like “So, we had to develop everything ourselves, and we used what we already had, and it worked fine”. Hmm. Why, exactly, would I care that you did “something”, “somehow”, and that it all went “OK”? There’s just no information content for most of this part of the talk. Nevermind – you don’thave to read it! :)]

Future live team

To date, an independent group outside the other game teams. We want to embed ourselves into the next project – probably Dragon Age – work with them and leverage our previous experiences.

COD4’s RPG elements, Wow’s armory – we’re looking forwards to developing even more cool community / website integration for dragon age. Can’t talk about specifically what, but it’s great.

[Adam: /me is annoyed because the “future live team” section was basically empty on any actual info about the future. If you can’t talk about it, then DON’T MENTION IT. Your secret projects, about which you won’t tell us anything, are less exciting to us than you think they are. Sigh]

We have a live team: you need one too!

[Adam: I would be tempted to delete the rest of these notes and just leave that one line up there. Kudos to the speaker (he stuck it up right at the top of the slide, in big letters). Any game that launches today IS an online game, even if it thinks it isn’t – your community is online (even for a single-player game), and you stand so much to gain if you just throw a tiny amount of resource at it. A live team for a non-MMO is, IMHO, worth it’s weight in gold. Especially as so many other companies *do not have one* (in the MMO space, we all have them, so having one is not so much of a differentiator over competitor products – it’s just an essential, so it’s harder to stand out)]

Organizing options:
– separate independent team
– embedded in game team

[Adam: there’s loads more options than that: strike teams, split-by-project, split-by-expertise, technology-lead, feature-lead, community-lead, etc. But I guess this is just something that MMO people are more experienced with, so we can think of more options]

– this is essential: what are the boundaries of their responsibility and work
– one area is support: if your dev team or publisher cannot provide the level of support you want for your title, you can make forums, knowledgebase systems, etc.
– patch releases: again, if game dev or publisher cannot/will not do it, you can take control of it and do it well
– non-game utilities: …again if quality isn’t what you and the community want, take it over
– post-release content: downloadable content, premium content, etc

– need a lead
– + selection of people dependent on whatever your mandate was that you chose

Role details
– community manager: you need someone who is very good at interacting not only with game team but also with the commuinity itself. Advocating what game team needs, and helping dev team understand what the community wants. Sanity checks what they are hinking of doing, saying what community will react with.
– lead: lots of good scheduling skills and personnel management. Need to truly understand the areas that you’re working on, dev team needs, and community needs.

– sometimes it’s OK without a live team, but even then you end up with a fairly disconnected community/dev team pairing
– dedicated group of individuals that allows you communicate more effectively between dev and players you get a much more vibrant community and you get much more brand faith
– bioware community is 3.8 million subscribers, been around for 7 years

Q: have you ever had to shut down any community servers (websites, forums, etc)?

No, and no plans to do so. Ongoing operating cost is minimal. Systems are automatically monitored.

Project shutdown: NWN is about to get it’s final patch ever, within the next month. Will have been running for around 5 years now, much more than our expected original run of 1-2 years. Servers will stay up, but will be no further support, no further patches or content updates.

Q: how much have you used your community as a recruitment tool?

We’ve pillaged the community quite violently, we’ve hired people from all over the world from it.

We saw people putting out content that just blew our mind.

A very valuable tool for us.

Q: what was the largest size the team got to?

About nine in total, not including the infrastructure support guys (databases, server maintenance).

4 designers, 1 artist, 2 programmers, and a lead and a community manager.

Q: was it something Atari [their publisher at the time] was willing to pay for to make original live team?

very much an internal project within the developer. Publisher didn’t really notice it existed, it was just we ahppened to be the people within the developer that took responsibility for those things.

We had our own budget, so we needed to justify our continued existence internally.

Q: any problems with console games?

We only really deal with PC versions of our games now. Not really a lot to say for the 360 version right now, we don’t have so much content with the external team that does it, either. Mainly leaving it up to the standard community guys.

We do “self-help” support: a forum for support mainly helping and expecting people will discuss the issues and help themselves. Publishers do the frontline support. Sometimes we fly our support guys to the publisher and have them brief the frontline support people.

Members of our own company pile on the forums themselves. We just can’t help ourselves, we’re so desperate to make sure people enjoy our game.

Q: UGC often includes copyrighted content. Problem?

For the most part we’ll have to contact them and tell them they can’t talk about it on our forums. Since we don’t host ANY UGC, we don’t have to worry about it too much. Neverwinter vault had to deal with those situations quite a few times, because they were actually hosting the files.

That’s a lot of why we decided not to host content ourselves. I think that’s changing now for both us and for other people in the industry. We now have more faith in reporting systems and reactive handling of such problems.