The Birth: The Launch

This became our gravest mistake, if one can be said to have control over the time of birth. The child was premature. It had to be put in an incubator . We lost more players than we thought possible. I have spent countless hours thinking so much about the reason why it went so horribly, horribly wrong. What made us launch so early, and generally be so unprepared? I don't really know how honest I can be in aiding our competitors to avoid making the same mistakes, but it might be good for mental hygiene anyway. I must confess, it is still a touchy subject for many of us. Anyway, I will run quickly through the major mistakes made at launch. (These mistakes are listed in an arbitrary manner. Who knows which was the most important one?)

Mistake: Not Enough Testing in Realistic Environments

We had many thousands of Beta players ”but not online at the same time ”fighting monsters, trading items, running nanos, or playing in tightly packed areas. Basically, the tests we had pre-launch could in no way prepare us for what happened at launch. The servers crashed. The clients crashed. People lost characters and items. (Thankfully, we managed to stop people from losing their characters really quickly!)

What We Learned from This

I don't know how we could have possibly prepared better for launch, testing-wise, other than simply to have spent a lot more time on it. It is very hard to "simulate" the rush of several hundred new players every hour . We also learned that "simulating" players by using an auto-client wasn't enough, in our case. We could not predict player behavior enough. What we did do correctly is make the entrance areas "virtual." We had the ability to make "indexes" of all play areas ”to make virtual instances, that is.

Mistake: Group Thinking, or the Bay of Pigs Revisited

What makes for interesting reading in Psych101 and history books about the US invasion of the Bay of Pigs still happens in 2001. This is the deal: We launched early because no one had the strength or will to resist the launch anymore, it seemed. Normally, there is a function within game developers/publishers where management and marketing push for a punctual release (according to plan, in other words) and product QA and development teams resist if the game is not ready to go on the shelves . This is a healthy process normally, because the development team wants to tweak and polish forever and management wants to make money yesterday . In this "conflict," you make the right decision most of the time. Not so with AO . People on the development team were too tired to resist the launch; there was fatigue felt to the bone.

In addition, almost 100% of the Funcom employees have shares/options. This is one of the greatest things about working for Funcom, and it still makes people feel strongly about their job and company ”you know, the feeling of ownership and all that. We all knew that we didn't have enough money to continue development of AO past the summer without another round of investments. We all knew our shares would be watered down ”again. We wanted so much to succeed.

It is not as if I am saying the launch day wasn't defined by management; it just made us all "on the same side." Our internal, normally extremely good and impartial Testing department became compromised as well, due to pressure and being tired to the bone. Group thinking, where any nay-sayers were seen as a collective evil, was making the company go into labor.

What We Learned from This

One needs to be aware of how fatigue is affecting group behavior. The next time there will be different, fresh people deciding if we are ready, not the people who will gain from it. Maybe we will enlist an outside consultant to aid in this process.

Mistake: Too Many Features/Not Enough Focus

We let the wishes of the players guide us. We tried to incorporate what they said they wanted (during Beta and before), and it slapped us in the face. A good piece of advice: Decide for yourself what you want to listen to/what you want to include. I will give an example.

One of the features people were really excited about before launch was player-made missions/quests. "Ooh, that is nice! Yes! I want to try it!" everyone said. We put it in. We launched it. The result? No one used it, except for the exploiters. (And man, was it great for them!) After trying to fix it for several patches, we simply disabled it. The button is still there in the GUI; it simply does nothing. No one complained because no one missed it. If only that time had been spent on something else .

Players have complained about several other features " promised " pre-launch, but not that one. I'll get back to that.

What We Learned from This

Don't try to please everyone. Although you have a community to cater to, you can never , ever please everyone; by pleasing some, you displease others. Stay true to your idea and focus on it. Cut early rather than late.

Mistake: Inadequate Customer Service

CS was completely understaffed. Also, the poor people there had terrible tools at launch. The result was several days turnaround per petition. The first thing we did post-launch was try to improve those tools and staff up.

What We Learned from This

Yes, you need CS. You need them more than you think. Also, it is wise to moderate official boards heavily. One " grief " community member does more damage than you can imagine.

Mistake: Lack of Focused Project Management

At launch, project management was done by too many people. The responsibility was divided and crumbled further at launch. Of the managers, one suffered almost a near burnout, one left for vacation for five weeks, and two left the project to do other stuff. The result was that the development team became too unfocused and stayed that way more or less until Thomas Howalt was assigned to the project in September 2001. (He worked on Midgard until that time.) In addition, the design department was reorganized to cater to people leaving or working less. This led to a loss of creative focus as well.

What We Learned from This

The launch is but the beginning. It is when you need people, fresh and eager , to focus on the onward development.

Mistake: Believing the Game Was a Normal PC Game Before Launch

Having worked on other games before, we followed our usual pattern. We worked like mad, night and day, up to launch. The result was a back-catalog of vacation to be incorporated into a busy post-launch schedule and severely tired people. This is not the way to launch an MMOG. The launch day should be seen as something happening in the "middle" of development, like the start of a marathon. It is not the end; it is the beginning. It is the birth, not the death. (This is why a post-mortem is rather misplaced, but never mind .) You need fresh, motivated people for a launch ”not dead-tired burnouts ready for the beach .

What We Learned from This

Again, the launch is the beginning. If your organization has developed other games before, it might make our mistake. If this is your first game, you are almost bound to make our mistake; you simply will work too hard toward launch. Do not let your guard down!

Mistake: Totally Wrong System Requirements on the Box

They were wrong ”horribly wrong. A lot of the players who had a terrible experience when trying out the game would have had a much better time if they were told their PCs were underpowered before buying the game. It worked more or less on the original system requirements, but only if you cranked the rendering options all the way down, and who does that?

The game still suffers from what the players refer to as "lag." Most of the time, people believe this is data from the servers being delayed or something like that. It is not. It is the client PC loading textures. On an AMD 1.4GHz with 512MB DDR RAM, 64MB Geforce3, and striped hard drives , there is no "lag" at all.

What We Learned from This

Try to delay marketing from printing the box, or make "stickers" to put on the boxes if you are unsure about how the game will perform until the very end. That is not too weird because a lot of optimization is done toward the end, when you see where the pressure of the game applies.

Mistake: Underestimating the Power of Pre-Launch Promises

Speaking with the media and fan sites covering MMORPGs is actually quite different from the ones covering non-online games. First, the player community is totally different. They track you, store your words, and try to hold you to the things you said. It feels very much like being "in office," I have realized. If two years before launch you admit to a journalist , "I am thinking about doing this or that," it is like President Bush saying, "I am thinking of attacking Iraq." You can bet the world will not forget him saying that. You have to have a different mindset to make, and communicate about, an online game. You have to weigh your words all the time. That was never easy to understand for an organization that has always had to hype what it wanted to sell!

What We Learned from This

You are a politician! Yep, you are. You are running for office, and there is no filter between you and the world. Stick to the party program! Don't enter into a "design session" with a reporter, even though you can make his/her eyes glow and make headlines.

Mistake: Overestimating the Community's Ability to Forgive and Forget

At the time of the launch we rationalized, "We have been players during the launch of other online games. It was terrible, but we still played on. So will our players!"

Of course, that was true of the previous generation of such games, but not the current generation. There were no alternatives back then. Players suffered, but forgave; the world was wonderful.

In June 2001, people who are used to the stability of a 2-year-old EQ , 1.5-year-old AC , or 3 “4-year-old UO will not easily forgive having to go through birth pains again. People forgive childhood diseases the first time a new technology comes to market, but the second or third time around ”no way!

If you, the esteemed reader, are planning to launch something similar to AO , you should have something you are 110% sure of! Otherwise, test until you are. The entry barrier in this market is becoming more and more terrible. It makes me kind of proud to think that a small company like Funcom can compete with the likes of EA, Sony, and Microsoft.

What We Learned from This

Things change, the world changes; what was true will not necessarily be so again. The consumer is not a forgiving fool; he carries spears and phazers. When there is monopoly, anything goes. When there is competition, nothing but the best will survive (or sometimes, the mediocre with the biggest marketing budget.

Mistake: Not Enough Unique Content

We should have had more designer content resources added to the game. Too much became too generic. It is like we offered people dinner at McDonalds, or simply the bread and butter of online life, every day. They quickly wanted more, and we should have had it there. Especially for the high-level players, special content is vital .

What We Learned from This

Handcraft and mass-produce. As a designer catering for a world, you need both. If you only handcraft, there will be "camping," waiting, and irritation. If you only mass-produce, there will be nothing to want, only grayness, drabness, and repetition. We luckily have both now, but it takes time!

Mistake: The Learning Curve

The learning curve is too steep. The GUI is great when you know it but bad when you start out. Likewise, it is difficult to learn the skill system. This is something that should have been fixed pre-launch, and when added to crashes, lag, and other problems, it simply became too big a hurdle to many players. You need a tutorial for these types of games. You need an easy learning curve. You are creating a world, not a game. People are not forgiving of you when they are left feeling stupid and not empowered ”and other people are watching.

What We Learned from This

You need a tutorial. You need to make the game easy to get into. We knew it long before launch; we just thought launching early was more important. It wasn't.

Mistake: The Patcher

We had a big bug in this at launch, which was unforgivable. We launched with a game many couldn't start.

What We Learned from This

We learned nothing ”we already knew this. There is no excuse .

What Went Well at Launch?

There were things that went well.

The Team

The team simply rose to the challenge. It became an even more extremely focused group of people during hardship. Most did not give up or leave. We tracked down and cracked so many bugs that the floor was green with bug juice . We also managed to do necessary adjustments to the game, although the capability to communicate these was not present at first.

The Core Technology

The technology actually held together; its ideas weathered the storm . I remember in our darkest hour, post-launch, that we felt we had to rip out the core technology, from the bottom up, and change it all. We didn't. It sustained; it still does. Looking at all the mistakes mentioned, you would think it was a miracle . It wasn't. The technology was sound.

The "Lastability" of the Role-Playing System

Once you get past the learning curve, AO has proved to be remarkably last-able. It has great depth, and it keeps growing and growing. Sometimes one can almost be tempted to believe that the width of the game is the depth of the game. Most other competitors put your character on a levelling treadmill, just like we do. In AO , you can at least control its direction as you want. It is not on boring, predictable tracks.

The Beauty of the 3D Environment

The 3D environment has stood the test of time, so far. We shall improve it to stay ahead.