Last week I published the first technical chapter (chapter 4, describing Real Money Transfers) from my previously unpublished 2009 "Sustainable Virtual Economies and Business Models" paper. In that article I explain why I waited eight years to publish, and why I published the way I did. I will use the same format here so please see the previous installment for an explanation of why I am doing this.
In order to explain the need for advanced business models for games (what I would now call 2nd Gen F2P) I had to analyze the strengths, weaknesses, efficacy, and methodology of all previous existing models. That is the content of this chapter (chapter 10). This is the second to last chapter, and the last chapter gives an example of a 2nd Gen F2P model. The intermediate chapters describe in detail how to design a passively RMT-resistant virtual economy that can be sustained indefinately. Even in 2009 (and also since then) all active (and much more expensive) methods of combating RMT activity have failed due to the high rate of core players being marked as false positives.
I have no plans to publish any of the solution chapters (5,7,8,9, and 11) at this time and will continue to hold those as trade secrets. Chapter 6 is the chapter on Immersion referenced in the previous article, which was made public in 2011. While it was assumed originally that you had read that before getting to chapter 10, it is not a necessity to understand the contents.
The following terms used in this chapter are explained in the Definitions section of the original work:
PvE/PvP/PvV: Terms describing various types of combat in virtual worlds. PvE is player combat vs. an NPC or mob. PvP, or Player vs. Player, is consensual combat between two PC’s. PvV is non-consensual combat between an aggressor PC and a victim PC.
Needle: Any discreet item/ability/buff or some such, whether temporary or permanent, that a player can purchase to give them an in-game advantage against other players. [Obsolete: Replaced with the term Supremacy Good in 2012]
As before, the chapter is in its unedited August 2009 form, for historical purposes.
Current Business Models
A few distinctions will be necessary to approach this subject analytically without all of the misleading marketing hype often attached to new products. First of all, there is no such thing as a free to play game. If the game is a retail product, then someone has to pay, even if it is an advertiser. The second critical distinction I will make is that beta products are pre-retail products and as such in no way charge the consumer. All but dead virtual worlds are under constant development so for purposes of this work attaching the term “beta” to a retail product is nonsensical and has no meaning other than to suggest that the product is a retail product that is somehow inferior in content, function and/or customer service. If even one consumer is being charged for the product, it is a retail product. Charging $10 for a “beta CD” makes that product a retail product.
The following list covers the primary existing virtual gaming business models that the author is aware of. Models for non-persistant systems like card games are not covered at all by this work. There are additional models that are combinations of these models but as trying to cover every combination would be exhaustive, just the core models are discussed in detail:
Traditional (Universal) Subscription: all of the virtual worlds created in the first five years followed this model so I describe it as “traditional”. Here the consumer pays a flat monthly fee and receives full access to all game content the full 24 hours every day as long as the servers are up (early games tended to not be up frequently). This model is still popular in the West but is highly inefficient, easily exploitable, user unfriendly, and favors the least desirable customers. When you get 24 hours access to all content, those that get the most value are those that utilize the service all 24 of those hours. This group includes only farmers and those suffering severe bouts of uninterrupted intense immersion. The latter group will soon become disabled leaving only the farmers. Those casual gamers, which are your best customers because they represent the least overhead, pay the highest rate per hour or per content and thus end up subsidizing your undesirable elements.
In essence this model involves the casual gamers subsidizing the farmers, who then turn around and charge the casual gamers again for the game time the casual gamers purchased for the farmers. This is tremendously lucrative for companies generating RMT attacks and ruinous for casual gamers. Hardcore players do alright but end up having to compete with farmers for loot and so their subsidized play time is also compromised.
As you can imagine this model is extremely popular and lucrative for farmers and only really acceptable to consumers because they are familiar with it and don’t generally have any choices whatsoever other than moving to another product. That other product will likely use the same model also if it is a Western product.
Advertising Model: a fully advertising-funded model is generally reserved for dead and dying games that have outlived their time, or extremely inexpensive projects. The reason is that advertising revenue is generally pretty small. More often this model is combined with another model in order to generate additional revenue for the producers.
Time Graduated Model: I am told that in China you can play World of Warcraft in internet cafes, paying by the hour instead of by the month. This is an example of time graduated pricing where you pay only for the time you use. Presumably this ends up being cheaper for Chinese users than for Western users so this is also an example (in this case) of discriminatory pricing. This not only has the benefit of charging more affluent users more for the same or similar product, but it also allows the producers of the product to inflate their customer counts with cheap or to some extent subsidized accounts. Any system that charges more for playing more is in this category.
Content Graduated Model: this model essentially makes only a small amount of the total game world available to new players as a form of demo. To unlock the rest of the world, or additional gear to supplement play, you pay for the additional content you want to use. While some of the following variants have aspects of this model, I will use the term content graduated to refer to models that offer additional content but do not give users an overt play advantage over other customers on the same content. This almost always means that the game is absent a PvP component. Sony’s Free Realms is about as good as an example as I can come up with and a meritous if not lucrative attempt at this innovative business model.
Complete Episodic: This model involves a complete purchase of the product with no recurring fees. The player buys the game and can play it online for free. Patches are generally free but any major content addition is sold as an expansion. Without fees the revenue stream is smaller but this is usually compensated by a smaller size of content, lower development cost, and higher volume of sales. Customer service is also typically minimized. Guild Wars is the classic example of a game of this type.
Free To Die Model: this model is a PvP or PvV design where you can log on for free but if you actually want useful equipment to participate and not get killed instantly and certainly every time you come into contact with paying customers, you have to also be a paying customer. Non-payers often use these games as pretty chat rooms as that’s about all they are good for. Payers generally enjoy killing non-payers for a bit, then perhaps briefly explore the game content which is generally much more limited than a Western traditional game. Fees for payers tend to be higher than what a traditional subscription would be as the payers subsidize the brief and fatal play of the non payers.
Deer-Hunter Online Model: This is similar to the Free To Die model but utilizes a PvV system that is continually in effect even when the player is offline! New players get to be the “Deer”. Paying players get to be the “Hunter”! Deer die horribly and quickly in droves, even if they log out! This yields a high attrition rate for deer which is countered with heavy advertising for this retail game touting it as a “free beta” when it is neither. Hunter-class players are charged sometimes over $1000 per week. Desperate players can buy so many needles that they look like porcupines.
These games are so expensive, even in their “free beta” stages, that they have special VIP designations for players that spend certain amounts. The highest level costs a minimum of $2,999 to start and another $2000 per year to maintain. It is my belief that this model is patterned to some extent around Las Vegas and other gambling consortium business models. The problem here is that when you spend an enormous amount in this game you are actively attacking other people with your money where in Las Vegas hopefully only you are losing or winning. This is like one player at a table putting down $50k and using it to stop all other customers from playing in the casino, or taking the money of all other players in the casino that cannot afford to match the ante.
The health effects of this model are devastating as players are actively threatened with asset destruction if they log out…. Ever. Other players can see immediately if you log out then attack you so a great number of players in these games do not log out. Uninterrupted intense immersion is almost mandatory for the duration of play, especially if you have money invested in the game (which you are forced to protect). This is the most consumer-harmful model I have yet to observe and due to its potential for short term profit I fear it will be copied widely.
Customer service in this model also is done in an interesting fashion. For deer it is basically non-existent. VIP members get customer service, with higher levels getting direct human contact. There is a widely held belief in the community, especially with VIP members, that game rules can be broken by VIP members at will and that these players are immune to discipline. Even if 99% of players quit without spending significant money on the game, it is still quite likely the game will turn a profit as long as 1% pay to become VIP’s and to ensure they “win”.
Analysis: If an investment group puts $5M into an MMOG project using a traditional subscription model with no discriminatory pricing involved then there will be a sizable number of customers whose maximum price point is close to the subscription rate. There will also be a substantial number who are willing to spend several times the subscription rate. In the absence of any other mechanism, these additional funds will go to the RMT industry assuming no RMT countermeasures. If the aggregate consumer base was willing to pay $10M for this product and half went to the producers of the product and half went to the RMT industry, then that is $5M gross income to both. The game company breaks even, with no profit to investors, no money for development, and probably no sequel or even much in the way of patches coming down the pipeline. The RMT industry does fabulously but the game company is wrecked. The money generated by the RMT industry is unlikely to fund further game development and investors will be increasingly unlikely to invest in these projects.
Moving to a time graduated model helps the picture tremendously. Players that could not afford the base subscription of a traditional model now can afford to play limited hours. More affluent players will play as much as they want, and the design should charge them more than the price point for the traditional model. This is an example of discriminatory pricing that works and it also charges farmers at the maximum rate since they normally never log out. This method also has less tendency to encourage players to not log out so it results in healthier players which tend to be longer lasting players.
Using a content graduated model has similar benefits and can allow more enthusiastic players to help subsidize uncertain or marginally interested players. The downside is that in an attempt to mimic what the RMT industry is offering, they tend to try to mimic the offerings of the RMT industry (“visit our marketplace and buy a new sword or puppy!”). This uncreative approach utilizing the “if you can’t beat them, join them” philosophy just does not work. The problem here is that as soon as you can just buy the best equipables in the game, those items lose the greatest part of their value, which is their intrinsic prestige. Having a very expensive store-bought weapon does not make people respect you for your hard work and skill as a player. It tends to have the opposite effect and devalues your entire product.
Having a content graduated model that unlocks services or actual additional play content (like additional quests or play areas) does not reduce the prestige of players and just allows players that want to play or experience more of the virtual world to pay for the amount they want to enjoy. This approach works and can be viewed as a game with multiple expansions that launched with the expansions already included. Instead of more enthusiastic players hitting the end and having to wait a year for the next expansion, they can just buy these products as soon as they want to utilize them. This can be done without the dev team having to actually make any additional content, they can just segregate their existing content into discrete batches.
I consider all models that charge for the ability to prey on non paying customers to be short term gimmicks for products with inferior content. The real service they are selling is not an appealing game world per se as much as they are selling a grief sanctioned environment whereby some players can pay for the opportunity to hurt other players. This environment is never sustainable as pretty soon only griefers are left. As they begin to prey on each other, the product originally offered (the ability to gank someone that cannot defend themselves) disappears and the remaining griefers become unsatisfied. Heavy marketing budgets and new server starts can only delay this process so long. Furthermore, I suspect the number of players that are willing to pay to grief is a relatively small percentage of total gamers and that this group tends to skew towards younger demographics that tend to be less affluent. You might even find out that many of those “rich” 15 year old customers were paying using credit card fraud, and getting your money from a 15 year old criminal in another country is probably not going to be a profitable venture.
As I found all existing models commercially deficient, rigorously analyzing why they worked or didn't work was a critical first step in forming solutions. In 2009 I was not paying much attention to Facebook games, but all of the models used there are discussed here. In 2011 I was asked by 2K Games executives whether they should start developing games for Facebook. I spent a few days intensely analyzing games in that space and the result was "Zynga Analysis" which I was free to later publish as I produced it initially as a skill demonstration.
Zynga Analysis explained that the claims by Zynga that their success was based on "innovations in the application of data science to games" were illusory, and that the real reasons for their success no longer existed. Thus the paper explained why the company was already in free fall despite enormous hype by industry trusted sources. While the only detailed and public prediction of the "pump and dump" prior to the catastrophic (for investors) Zynga IPO, I was not well read prior to that time and was unable to avert the disaster. 2K Games stayed out of social network games and was able to avoid the substantial losses that competitors such as Electronic Arts suffered during this period in the social network games space.
By 2011 I had identified two products that were using advanced F2P elements. Those were "League of Legends" (RIOT) and "World of Tanks" (Wargaming.net). I assumed they would out perform their competitors and the next year I wrote and published my "Supremacy Goods" paper formalizing that prediction and going into some detail as to why they would outperform their competitors. I also assumed the models were the product of intuition on the part of brilliant and gifted designers that had not done this sort of rigorous analysis and thus they would have a hard time improving upon that sucess without very expensive trial and error.
I offered to help both RIOT and Wargaming.net avoid those unnessary expenses. RIOT declined to meet with me but Wargaming.net Vice President Andrei Yarantseu stopped what he was doing for two hours in the middle of the busy 2012 E3 convention to read and discuss the paper. This led to WG hiring me to optimizing their business models on their World of Tanks: Blitz and World of Warships titles, in addition to a number of still unannounced projects. Both of these products possess superior monetization metrics despite being in relatively narrow niches.
While the 2nd Gen F2P prototype model detailed in Chapter 11 (from 2009) is still more advanced than anything I've seen deployed yet in a game (including my own works published since then, due to external design constraints), I consider it simplistic as it contains none of the game neuroeconomic systems I've developed since then. At some point I will publsh it for historical purposes only. I have not attempted to articulate in writing my game neuroeconomic methodology as I did here in 2009 with game economics, because I consider the technology dangerously effective at modifying consumer behavior. Such technologies, despite the best intentions, invariably end up in the wrong hands after being released.
I would like to stress that game monetization is just a very narrow slice of the larger field of game economics, which is itself much more simplistic than the field of game neuroeconomics. I write mostly about game monetization because that is what I'm most asked about, but I feel this narrow focus causes our industry to miss the big picture and limits our ability as developers to successfully meet consumer demand. In other words, it's a fad.
Current industry-standard development methodologies are unscientific and extremely limited in scope and performance. Rare super hits like World of Tanks, World of Warcraft, League of Legends, and Pokemon GO are essentially leaps of intuition that our industry has been unable to replicate because we do not apply the tools necessary to understand why they were successful or predict that success. Thus I keep hearing that "you never know what will be a hit" when this sort of dogma is just not true.