MMORPGs, and their need for a live team, are dragging the game world kicking and screaming into a new publishing paradigm -- one that it doesn't really know what to do about. What's more, we are being dragged without preamble and without preparation. Before we had live MMORPG products, the entire game industry was set up top to bottom to deliver a boxed product to the stores. Once the game got to the consumers, publishers then sat back and watched the retail sales of that product. The way games' production cycle is geared, the way games' advertising money is spent, the way games' development mentality focuses its efforts -- it's all designed to support an "arrowhead" approach. That is, all team members aim towards one specific delivery point so as to gain the deepest penetration for the product in the marketplace. We do this because we all know that if a game isn't a hit in the first month or two, it isn't going to be a hit at all. Rarely does a game rebound or have its sales curve accelerate and have an extended life in the months following its release. It happens, but I can only think of a few games that have done so.
But MMORPGs have a new and different issue to contend with: they exist live online for long periods of time, and thus have an extended product "shelf life" by their very nature. There's the launch, then the public reaction to that launch, and how the stability of that launch affects the sales (recall Anarchy Online?) and then after the dust settles, the product begins to exist on its own in and the audience starts to grow (or not). I believe we have to re-think how we deliver these kinds of products to the public. Perhaps we need to learn some lessons from those people who live and die by delivering live product to their audience on a scheduled regular basis: mainstream entertainment like broadcast TV.
I have some experience in both areas: my original training was in the dramatic mediums (theater, film and TV). I have worked in theater and TV, and have many friends who make their living there. For the last 22 years, I have been a creative director / designer in both the pencil & paper and electronic game arenas. I have over two-dozen shipped titles to my credit, and have for the last 2 1/2 years worked on Earth & Beyond, EA Online's science fiction massive multiplayer online game. I can speak to the comparison in process and focus between the two publishing models.
If there is one thing the MMO space has proven so far it is that no product can survive once launched without giving the players more of what they want: some have delivered expansion packs that contain new lands to explore, some have updated the game's systems and content through patches, and some have started to try and tell stories on line that involve the players as active participants. But everyone realizes that only shipping the box isn't enough. At the very least, a team of developers has to be in place to fix the bugs that crop up as people play the game. My question is this: are the skills needed to get a game out the door in the first place the same set of skills that are needed to maintain the game once it goes live and also change the game experience in a way that helps to grow the audience?
These games are very expensive to build and publishers want, understandably, to begin recouping their investment as soon as possible. The economic model of these games is designed not so much around the sales of the box in stores but around the continuing income from monthly subscriptions. Frequent updates and expansions of the product support and drive up subscriptions, which in turn drives long-term revenue. So, it is no exaggeration to say that equal in importance to the eventual economic success of a game are BOTH the quality produced by the original development team as well as the quality produced by the ongoing live team. We all know what a development team does, and a little later in this article we'll discuss what the responsibilities of the live team are and how the priorities of live team development might dictate a different set of talents from the development team.
In the game business, because the retail sales are (historically) the only way to recoup investment, publishers have developed predictive methods to determine very quickly after a product hits the shelves whether they are going to recoup or not. But in TV, the profit comes from the long term, not the short term. Unless the show is sold into syndication, it is hard to call it a hit. This payday comes years after the successful (if it is successful) first season. TV series have to acquire a backlog of episodes they can sell as a package to these local channels, and it is also much easier to sell packages of half-hour sitcoms than hour-long dramas. Syndication is where the real money is really made. Any investment is truly a strategic one for the producers.
So, in TV land, the development mentality is geared towards trying to generate a series that is reasonably successful from the moment it launches, in order to build economic success over the long term. When you think about it, this economic model used by the TV industry is much more akin the MMORPG model than to the boxed game model. The television industry is used to that approach, and today it is the only profitable model for that industry.
In the game business, we're totally focused on success at launch. TV is focused on a successful "run" of a show. These economic forces affect everything about development in both mediums. Let's examine the differences in the two development environments.
Comparison Between Broadcast TV and MMORPGs
TV: TV delivers new episodes in a series, which everyone assumes have new content. In fact, TV has shown that recycled content (re-runs) don't get as high a rating as new content.
Games: Games deliver new patches. What is a "patch" except a fix to something that's broken? We're used to patches because we live in an environment where a games ships and then, because we never have enough time to really polish it before it shipped, we issue a "patch" over the internet which fixes some of the things that are broken. Part of our mental adjustment to the new paradigm needs to be to see these "updates" as opportunities to continuously deliver to the players new and interesting systems and content to make their experience fresh.
Earth & Beyond delivers an update each month, but going to a weekly, or even bi-weekly, update schedule would be impossible for the team right now.
Conclusion: The question we have to ask is, "How much does this lesson apply to MMORPG games?" Earth & Beyond tries to deliver one update a month. Is this enough? Is this too much? Does frequency matter in this space? It is hard to prove one way or the other at this point. I can say that on Earth & Beyond, the players look forward to content updates and some even take days off from work on days we've announced updates in order to play all day and see what we've added. One could extrapolate from that the idea of frequent updates is desirable to the audience. But how do we reach a goal of weekly or even bi-weekly updates? Right now that is impossible for us. I believe that marketing or design should determine the best timing to deliver these updates in order to increase or maintain the subscriber base. The evidence is there that when done with this in mind, we create excitement in the audience. Just like new episodes of a TV series do.
TV's Tight Schedule: Every minute in the development of a new episode is accounted and budgeted for. They cannot keep up with the production cycle otherwise. They have learned what it takes to create and produce a weekly show. Budgets are tight but the machine is well oiled.
Games' Flexible Schedule: Audience doesn't know when an update will come or on what day, so everything can be in flux. This is good for the development team, bad for the audience. Well, it is ultimately bad for the development team as well because it encourages the development team to be sloppy in its processes. "It's okay if we slip, because we'll just delay the content update, or we'll delay the system patch" or whatever. This implies that there is no sense of timeliness to the update, no sense of timeliness to the content, no sense of urgency resulting in no momentum building up in the audience. Think of the problems that Babylon 5 had with new shows, followed by reruns, followed by new shows again. This is very tough to fight against.
Conclusion: We should try to build a team and a technology base designed from the ground up to deliver regular updates on schedule.
TV staffing: Every job has a specific task. There are costume people, prop people, wardrobe people, and many other dedicated staff. While the unions regulate much of this, the production schedule dictates it even more. There isn't the time to be inefficient. There isn't the time to train people as you go. Unions also control the economics to some extent; producers cannot go overtime without serious financial repercussions. How many hours before we have to break for lunch, how many hours in a day? How many in a week? Working conditions are intense, but there is some protection for the workers.
Games' staffing: One of the dangers on any game development project is for the morale and health of the workers involved. There are no financial repercussions of going overtime. And since we don't know what in the heck we're doing to a large extent, sometimes it is very hard to correctly estimate timeframes. But because this isn't a sprint to a specific finish line (as is the case with a boxed product), the danger is greater.
On Earth & Beyond, when we were planning our monthly
updates, one of the writing staff asked me when during the year
we expected to give the writing staff a vacation. In TV, hiatus
comes and people can take time off then if they wish. The often-excruciating
hours are somewhat ameliorated by the knowledge the hiatus is coming
and everyone can collapse if they wish. Just as the rush to ship
crunch is somewhat eased by the knowledge that after the game ships,
people can take off and rest. MMORP staffing and development needs
to adjust to this.
TV delivery commitments are serious business. The client (in reality an advertising agency) has paid for a specific delivery time slot, and if they don't deliver that time, then the client gets a refund. They have paid for ads on a specific show that they expect to be seen by millions of people during the commercial breaks. If the show doesn't deliver the audience, then the advertisers can expect paybacks. Thus there is additional pressure not only to deliver the product when promised but also to deliver the audience when promised.
Games delivery commitments are treated loosely. The audience has no firm commitment from the publisher. The game gets updated when it gets updated. The client, who is the consumer in this case, only has a commitment from the publisher to keep the product on-line. Product development has no plan except a vague plan for the next six months, if that. We're all trying to adjust to the idea of what updating a product means. We don't know what it is we're supposed to deliver, or what things we can deliver to grow the audience. Everquest delivered only expansion packs and fixed systemic issues, and that game's audience grew a great deal. So what really works? E&B believes that frequent regular content updates work, but we're not really set up to do that. On the E&B team, we're learning to do that now.
Conclusion: We need to get real and deliver what we promise when we promise it. There's too much money at stake otherwise. We also need to get better data about how much impact any form of update matters. One lesson that TV knows very well: once you become part of a viewer's life, you've really made progress. It is one of the reasons why syndicated sitcoms really work. Everyone knows that at, say, 5:30 PM every day they can sit down with their dinner and watch Seinfeld. This is critical to the show's success. I knew a fine artist when I lived in New York who, every night, rain or shine, watched Mary Tyler Moore at 11pm on WPIX. This guy was a very influential illustrator in the NYC advertising scene, way too cool to admit to anyone but his closest friends what he liked to watch on TV, and Mary had him in her grip. These things become part of the fabric of our lives. MMORPG games do that as well, and we can do it even more so by understanding these techniques.
TV professionals know their production issues. The TV business has been doing this for years, and as a a result it knows how things work, how many people it needs, what technological processes work and which don't, how much lead time is needed, and how to skirt the edge of union rules and still keep everyone happy. The technology is stable. The workflow (write a script, cast the script, prep the script, shoot the script, post production, upload it via satellite, broadcast it) is proven. The technology gets incrementally better each year, costs rise, but these are all predictable concerns. And once the Fall season starts, it is all about keeping on schedule, keeping people sane, getting things done. When anyone, even a star actor, starts to get in the way, efforts are made to smooth over that bump, because bumps have a very detrimental effect on everyone's morale. Workflow is everything in TV.
Game developers don't have any idea what their production issues are. We've been doing this only a short time and only by a few groups. Let's examine what the existing worlds do to update their product:
- Everquest: One of the oldest and most successful. Fantasy role-playing on line. EQ's world expands through frequent unscheduled systemic updates and also expansion packs delivered every nine months to a year. By "unscheduled", I mean that the developers don't leverage the update or advertise it. They certainly do leverage and advertise the expansion packs. Their live team is built to deliver these expansion packs and to maintain the existing world. The skills needed for that are very similar to developing a new product, because an expansion pack is really a lot like a sequel.
- Dark Age of Camelot: An online fantasy RPG set in Arthurian England, and in the world of Celtic and Norse myth. Their update model is identical to EQ. Frequent systemic fixes/enhancements coupled with occasional expansion packs.
- Ultima Online: An online fantasy RPG in an original campaign setting. They have tried to update frequently. At first this was to fix the problems they encountered, and later to make the product better. They have tried to update content with mixed success, like the rest of us.
- Star Wars Galaxies: Too new to really see what their plan is. They are trying some story updates along with some bug fixes.
- Earth & Beyond: Working to find its audience, but updating systems, content and story on a monthly basis. The team believes that our retention rate is mainly due to our commitment to monthly updates.
So, you can see that we're all really doing our own thing for the first time. We don't know what to do, we don't know how to do it, and we really don't know what makes a difference and what doesn't.
Conclusion: Until we know what works, it will be hard to actually determine how to configure the live team.
Who is TV's audience? The TV audience has segments that are very well established. Advertisers sell to that audience. Content is created to fill and please certain demographics. While superficially this sounds like a bad thing, it is actually a good thing because then you get variety: from the Nature Channel to Bravo to MTV, and from West Wing to Golden Girls to ER to Married with Children to Will and Grace. The audience changes very gradually, but it really does change. TV is right now trying to adapt to the perceived "new audience" -- the one used to channel surfing and Internet content. The audience demographic changes based on the reaction to individual shows. The content dictates the direction of the audience. Reality shows added an audience category, essentially. "Reality" is a segment of the audience that no one knew existed before the first reality show hit the airwaves.
Who is the MMORPG Audience? The games audience is not passive. When the original Star Trek audience reacted loudly when the show was first cancelled, TV executives couldn't believe how vocal the show's audience was. That encounter changed the way audiences and TV interacted from then on. MMO audiences believe they own the product. E&B's audience has asked me specifically whether they could "sit on a board" and help us make decisions about the way the game should go. The audience is unknown and perceived as "geeky".
Conclusion: We need to spend more dollars and determine who this audience really is and also predict how big it will grow. If we're going to move towards a different production model, that model will probably be more expensive to run, and we need to determine whether the audience will show up to justify that spending.
TV Technology: Technology was very new and at first, very limited. Now it is ubiquitous. TV delivery started out much like broadband. Few homes, TV sets were expensive. Few families had them. Then it spread. There was a history of families gathering together to listen to the radio. In that same way, families gathered to watch TV. Families do not gather to log onto the Internet.
Game Technology: For online games, the technology is broadband. The difference between TV and broadband is mainly the age of the mediums. No so much the age of games but the age of broadband Internet.
Conclusion: The computer-human experience is solitary while the TV-human experience is communal. This is a big deal in terms of how we get people to watch. The technology is still very new. The delivery mechanism means that a large portion of homes doesn't have it yet. The percentage that matters isn't what percentage of computer users have broadband but, rather, how many homes have a computer AND broadband access. Broadband isn't needed so much for playing the game as it is for receiving the updates, which are huge in size
"We're Experiencing Technical Difficulties"
Correcting errors: Anything made by man has errors. Every TV show ever produced has some acceptable degree of error in it: bad continuity, incorrect props, lines delivered without the right inflection, and so on. The TV production system has people in the loop whose job it is to minimize those errors. But the errors are generally of quality, not of functionality. And so, because of the overwhelming pressure of the deadline, the TV product "ships" with some of those quality errors intact. The system prevents the gross errors from getting through. An example of a functional error in TV broadcast would be equipment failure: we've all been watching the news and become aware of the videotape machine breaking down. But that's about it. TV doesn't really have major functional errors.
Making the product bug-free: The errors on the game development side are both kinds: quality errors and functionality errors. I differentiate them as follows: any bug that does not prevent the playing of the game is a quality error, even if the bug is systemic and causes performance problems (example: lag, a common problem for online games). While most developers would classify lag as not an error of quality, it is akin to a quality error in that it does not prevent the game from being played but prevents the game from being played enjoyably. Since there is a greater possibility for games to contain a functionality bug, one that prevents the game from being played (a crash-to-desktop bug) because of the nature of our business, there is a greater possibility that we'll end up arriving at the go-live date and have to pull back because an error of this type is discovered. I would say that 10% of the time Earth & Beyond arrived at a go-no-go point, we had to delay it because the update was unstable or contained a serious quality bug that we deemed unacceptable for shipment. This doesn't illustrate anything but how new this all is to us.
Conclusion: The quantity of these errors can be lowered through various techniques -- layers of quality checkers, better scheduling, more mature technology, and so on. Many of these techniques will come on line as the audience grows and we can afford them. The real mistake here is to lower our target, to accept as "normal" an environment where we cannot deliver what we say we'll deliver when we say we'll deliver it. That has to be viewed as unacceptable instead of business as usual.
How Visible are the Errors
Visibility of TV's Errors: Most of the errors that do get through are unnoticeable by the viewers of the product. Many times, for instance, an actor's performance that doesn't deliver is itself not perceived as the problem it is but attributed to writing or just to as boring basic plotline, and vice versa. What's the worst that can happen in the event of a catastrophic error? The show isn't broadcast and people have nothing to do in that hour.
When television shows "ship with bugs", they're called "bloopers" and we laugh. There's even an ancillary market of shows dedicated to airing TV bloopers. Games don't have that luxury, unfortunately.
Visibility of MMORPG Errors: Gamers are used to products having bugs. In fact, they often attribute bugs where one doesn't exist. However, it is an accepted part of the gaming landscape. On Earth & Beyond we operated with a bug threshold. The game shipped with a number of known bugs and we strive to keep the number of known bug below that number while the game is live in order to keep the quality of game play acceptable. We believed that once the total number of bugs exceeded a certain amount, that even if any single user never encountered the major bugs, that the cumulative effect of many small bugs would leave such a bad taste in the mouth of that user it would drive them away from the product. This policy, coupled with the possibility that bugs can cause crashes and potentially damage the user's computer raised the impact of bug status as a variable of shipment to a very high level on E&B. It is not an exaggeration to say that we never rated timely shipment of an update above the need to keep the number and seriousness of existing bugs below a certain threshold. In other words, if the product was too buggy, we would not ship an update even if our players were screaming for it -- even if it was otherwise ready to go. While this makes sense from the standpoint of quality assurance, it does not make sense from a "keeping the customer's excitement level peaked" point of view.
Conclusion: Our process has to change to gain a higher level of reliability while we learn to deliver more frequent updates. While our current customers might put up with this because to them it is just business as usual, our growth as an industry is dependent on our ability to deliver frequent content updates that are bug-free. That bar has been set high by our competition.
The broadcast mediums have shown us that audiences crave new content delivered into their homes on predictable schedules. Online games are going to want to emulate that paradigm as they begin to attract a largerand larger audience because the methodology has been proven effective. Our game's own audience has shown that it will consume new content as rapidly as we can deliver it. We have to figure out how to do that, as the products that do will gain market share.