Development without a dev team
Thursday, 5th May 2011 at 7:59 am
Creators of the MMO Illyriad discuss the pleasures and pains of external collaboration
Illyriad, a free to play strategy MMO, is the brain-child of London-based James Niesewand and his team at Illyriad Games.
Recently praised by Massively.com as a “fantastic browser-based city-builder” with “perfect” gameplay, Illyriad went live as a Minimum Viable Product a year ago, and now, after two years in development, is approaching its official launch.
With the first iteration of the game nearly complete, Develop caught up with a selection of the developers involved in the project, to discuss the team’s approach to developing Illyriad.
Rather than developing the game with a single internal team, Illyriad Games used a highly distributed model.
Games are conventionally developed by a single team - coders, artists, designers, and so on - but Illyriad Games deliberately chose a very different approach. How?
Illyriad Games founder James Niesewand: Well, the core Illyriad group have built their careers on web and enterprise software, not games. So we didn’t have any preconceptions about how a game should be made.
We started by defining what the game should be like, what the experience should be for the player, and we simply broke that down into discrete functionality for the coders to implement.
From start to finish it took us about a year to get something playable. You might call it a Minimum Viable Product, or an Alpha, but for us that was just the point at which we had something fun that people could enjoy. At that stage we put the game live, invited people to play, and worked on improving the game in line with the players’ feedback.
Content, in the early days, was created by volunteers. Our volunteers are a great bunch of people; however, we need to take very strict approach to editing and converging the content being supplied. It was as we started to reconsider the content process that Kevin [Hassall, who runs the consultancy Beriah, and worked as production manager] began to get involved, bringing a great wealth of hard-won production expertise from the game industry and helping us shortcut around many of our early development mistakes.
Kevin Hassall: James has never laid out a formal methodology, but when you listen to him talk, it's very SCRUM-like, and centred around the coding.
His focus is on user-stories - functionality that gives a benefit to the players - implemented by the coders. He isn't especially focused on the assets that those features might require, though each might require hundreds of art assets or thousands of words of text.
Now, for a game like Illyriad, James's approach is perfect. So what I had to do for the production, for the art and text, was to deliver in a way that allowed James to keep working as he wanted. We needed to be able to turn people on and off as needed, cost-effectively and reliably. And the way to do that, on this project, was to use a really solid art studio and a bank of freelance writers.
So you decided that rather than bring in more staff, you would outsource?
Niesewand: Exactly. We needed to accomplish more than we could with volunteers, so we sat down with Kevin and looked at all the options.
Hassall: And in this case it made little sense to expand the core, internal team. Illyriad had a broad range of requirements, and they were going to need them work in uneven lumps – a big chunk of art at one point, thousands of words of copy at another... We needed people we could pull onto the project for short bursts.
And so you decided against expanding the core team. But it sounds like you also made few long-term commitments to your contractors? This all sounds quite chaotic. How did everyone feel about that?
Illyriad Games CTO Jeremy Hopkin: I will admit to a certain reticence. From my previous experience working for two major FTSE100 companies, whenever core development activities are outsourced it seems to be a bit of a hit and miss affair. Especially when it comes to integration and hand-over, there are generally always teething issues and code style conflicts.
However, continuing in the way we were was stretching us too thin across too many disciplines. And as it turned out, working with [outsource group] Quantic Labs, Kevin, and the writers dovetailed almost perfectly with our internal Agile working practice.
Quantic Labs exec Stefan Seicarescu: As an outsourcing company we are familiar with being thrown in hot situations and have to rise to the expectation in an extremely short time.
Quantic Labs exec Marius Popa: Yeah, when Kevin came to us with this project, we were very enthusiastic. After all, we all joined this business from passion.
Nicole Le Strange: As a freelancer, well. Freelancing means lots of variety and no security. It means working seventy hours one week, then having nothing to do for the next week. That’s just how it works.
And what was the impact on the actual work of the approach that you took? Was the work better or worse? Was it completed faster, or was the project delayed?
Niesewand: As soon as we switched over to using professional game-industry content providers, not only did the work quality improve beyond recognition but it was also completed at a phenomenal speed - compared to what we were used to getting.
Hopkin: This proved to be a double benefit; not only was Quantic et al’s work a massive increase in quality on what we had previously received, it also allowed us to concentrate on our core competencies - which then caused our work and timescales to improve tremendously.
Niesewand: We've also found the costs to be very reasonable, and we've certainly appreciated the flexibility we get in occasionally changing our mind, adding or subtracting items from the worksheet - or sending concept work we're not entirely happy with back to be rejigged. Whilst working solely with our volunteer army was an early exigency (and one for which we're eternally grateful to them), it was always difficult asking an unpaid volunteer to go and rewrite his or her lovingly handcrafted NPC description in a different way; whereas with paid professionals we have no such qualms nowadays!
It sounds like things went pretty smoothly? What were the most important factors in keeping the work on track?
Seicarescu: We tried to be very clear in communicating the progress of our work, and when possible we sent daily overviews of the work finished so far. The Illyriad team sent us clearly documented requests, so we could start working fast on the assets.
Popa: After that it all went smoothly, and we were always on schedule. We always managed to assess correctly the time frame needed to create an asset, and the very comprehensive explanations we got from the Illyriad team helped us stay on schedule.
Seicarescu: We could deliver fast, because we had good information and they knew what they wanted from us.
Popa: Furthermore, there is an advantage that Quantic Lab has a diversified business, from Quality Assurance to Development, Mastering and many more….
Seicarescu: … so we also see the graphic projects from the end user’s viewpoint and as a qualified testing lab; so all-in-all we take a critical view of our work.
Le Strange: With the writing, the best thing was the freedom. Illyriad defined the goals and gave some pointers, then just left me to get on with it. If I’d been trying to second-guess what someone a hundred miles away had in mind, I couldn’t have done it. Illyriad were great at explaining what I had to achieve, not telling me what to do, if you see what I mean!
Hassall: The key is to be very clear, but also very flexible. It's a question of allowing chaos, but within very well defined boundaries!
So, for the art, we defined types of output - buildings, terrain tiles, icons, etc. - and for each of these we had a cost and time estimate; therefore, it was easy for James or myself to say "OK, here's a list of another 50" without having to renegotiate schedule or price - we knew immediately what the financial and time impacts of our requests were. So, we could keep moving forward really quickly, and focus on the important issues - creative and technical.
A more formal process, where each new request lead to a formal re-quote and an bureaucratic sign-off process, would have slowed us right down and driven everyone mad into the bargain.
At the same time, we gave the artists and writers a lot of creative freedom, within very clear parameters, and we got back to them very swiftly when they had questions. Managing people remotely is absolutely not the same as managing people in the same office, and that fact has to be respected.
Hopkin: With tens of thousands of players online, every release we make receives almost instant feedback and we can see how the players react. These factors mean we have to be in constant communication and rapidly adapt our schedule and priorities. I see this as becoming common practice in the online and social networking spaces. You have to adapt and respond at the speed of your users.
But there must have been problems?
Hopkin: [laughs] After the initial samples and discussions to make sure everyone was working to the correct style and concepts; it turned out we weren't quite prepared for just how quickly the content teams produced their work. It sure kept us on our toes!
Hassall: [laughs] If that's the only problem Jeremy saw, then I'm delighted! Certainly the art seemed very straightforward, and that's down to the guys at Quantic. A good, professional art studio will communicate clearly and keep their promises, which is exactly what Quantic did.
If they did have any issues, they handled them themselves - their problems never became our problems - which is exactly what you want. I certainly saw more issues with the freelancers, but these were just things that I dealt with on a day to day basis - just as Quantic managed their internal artists to ensure that James and Jeremy got what they needed when they needed it, I did the same with the writers. Probably the big issue for me was that having writers in different timezones dictated different styles of working.
So, I had Nicole on UK time, another writer on Australian time, and a proof-reader on US time. In many ways this is wonderful, as one can pick up where the other leaves off - so our US-based editor could check text created earlier the same day further east.
But it can also lead to an inconsistency in the style of work. When I dealt with Nicole in the UK we had a lot of informal conversations, batting ideas around, while each piece of text was being written, so the UK-based text is very lush, densely packed with ideas.
On the other hand, with Tim, on the other side of the planet, I didn't have the same frequency of contact, so he would work while I slept and deliver pieces that were fully formed but generally more direct in style.
Because of the nature of the project we could turn this to our advantage, so different writers adopted different voices. But for the next body of text, we are going to have a more unified approach, having the teamwork much more closely together, and James has built an on-line collaboration system to handle this.
Le Strange: The biggest challenge was just the type of writing that Illyriad needed. If I’m writing game manuals, or editing a wiki, then that’s quite straightforward. But some of the things Illyriad wanted were really unusual.
I’ve still got a work-order email from Kevin that begins “1,000 words, first person narration, really nasty and violent, based on Aztec culture, but not constituting Mature content…” It’s a challenge dropping into something this complicated, for a short period. You have to get your bearings fast.
Seicarescu: The most problematic part, as it is constrained by the work itself, is to put on paper the image the client has in his mind by understanding his imagination and expectation. In this case we had some general rules: it should be all hand drawn, small icons should be still recognisable and most important, everything should have a Tolkien touch.
Now, the best known Tolkien drawings are the Middle Earth maps, which have clean, black lines. We had to transfer this feel into coloured pictures and invent by ourselves various assets. Tolkien never drew pictures of how a coloured tundra map would look for example….
A special problem when hand draws are requested is that every artist has his own style. When splitting two hundred assets over several artists, the look of every asset has to have the same style. To keep a similar look for all assets, the dedicated lead artist had to make the final changes and retouches for the assets delivered by the team, if necessary.
In the whole process, the final approval of some assets took longer then the delivery of them, because the teams had to wait until the assets could be tested in the engine. And as Jeremy said, they didn't expect to get the assets so fast!.
Is this style of working something you'd recommend to other studios?
James: That's really hard to say, as I have no experience of other studios and their working practice!
Hassall: Personally, my objective with Illyriad was to find a way of working that suited Illyriad. It isn't a one-size-fits-all solution. But it's always worth considering - intelligently and with an open mind - what functions should be handled by a core team, by an outsourcer or partner company, or by freelancers.
People find it easy at the start of a project to convince themselves that everything should be done internally. And often when they hit a problem or a bottle-neck they tell themselves that they can buy in a "solution", for example by using an outsourcing company. But the truth is more complex. The key to getting a really efficient development is to consider each project on its merits, and as early as possible.
Hopkin: With the benefit of hindsight, if anything I think we should have started working this way earlier. Having said that, it’s very important to find the right partners and we certainly didn’t have enough experience in the beginning in what was to us a new industry. Kevin’s help in finding the right partners was simply invaluable.
And does Illyriad Games intend to keep using this approach to development, as Illyriad grows? Or as the game gets bigger will you have to move to a move concentrated approach, with a single team?
Niesewand: For the meantime, yes absolutely, we’ll keep the same approach.
The flexibility we get on a project-by-project basis and the (so far anyway!) lower cost combined with the number of different styles and skills we can get by outsourcing works very much to our advantage.
I suspect that as Illyriad gets bigger we may have to introduce layers of internal staff to filter and to insulate the core in-house development team from other roles such as moderation, bug-fixing etc.
At the moment we're very happy to spend a lot of time at the coalface and we speak daily with the player-base - but I know, sadly, that this can't continue in quite the same way forever. And, there are definitely going to be certain other functions (such as marketing) that we're going to bring in-house as we grow.
Hopkin: We do have a frighteningly ambitious development plan for the future and it makes sense for us to concentrate on our core IP and nurture the talent which makes us unique. We have a clear vision, the ideas and technical ability to see it through – however until we brought on board the content teams, we’ve had to temper some of our ambitions as they were simply too big to be able to complete in any realistic timescale.
With this approach it allows us to perform as a much larger development house with experts in a diverse range of skills, yet it does not dilute our focus. We will still have to grow the core team, but I can only see us using our outsourced and freelance partners even more in the future as their contributions have been so good; it would be insanity not to.
© Develop 2013. All rights reserved.