Earlier this summer, I talked about the Shenmue 3 Kickstarter; I wanted to expand on my approach to questions like that.

In general, the price of a game (or of any product!) will be somewhere between how much it costs to develop, produce, and sell the product, and how much value people get from the product. This isn’t an immutable rule, especially the lower bound: there are plenty of companies that try to sell something and then discover that people aren’t willing to pay what that product costs; but if a company makes a habit of that, they’ll go out of business. And there are certainly situations where people buy a game and then decide that, in retrospect, that wasn’t a good use of their money. Still, that range is a good place to start.

Diving down: one big determinant of which end of that range a product (not necessarily a game, any sort of product) sells for is how many other options buyers have to satisfy the relevant desire. If a product is a commodity, something with lots of options that are essentially identical (a gallon of milk, a box of nails), then different producers can’t compete on value: they’ll compete on price, so the price will go down. And, in fact the natural floor isn’t the overall cost to develop and produce the product: it’s just the cost to produce a copy of the product (the marginal cost of the product), ignoring the development cost. Once a company has figured out how to produce a given commodity, the development cost is a sunk cost: at that point, if competitor A is selling for their marginal cost plus $5, competitor B can steal sales by selling for their marginal cost plus $4, and doing so will be more profitable than stopping production of the product. (That example assumes that both companies have the same marginal cost, so if you can produce at a lower marginal cost than your competitors, then you can steal sales by undercutting your competitors without hurting your relative profitability: this drive towards marginal cost reduction makes commodity businesses brutal.) In contrast, if a product can’t be substituted easily, then people will be willing to pay as much for the product as the value they get out of it.

Returning to games: one of the huge changes in the game industry over the last 5–10 years has been the rise of digital distribution. And, with modern networks, the marginal cost of producing another copy of a game that will be distributed digitally is effectively zero. This means that, to the extent that games are commodities, their natural price is $0; this is potentially devastating for game companies, but that hasn’t stopped it from happening.

Games, however, aren’t always commodities: some are a lot more popular than others. The problem there is that it’s hard to tell in advance which games will be popular enough to be able to command enough of a price premium to be able to recoup their development cost. This is why large studios have an advantage (because they can take a portfolio approach to average out the risk), and why those studios prefer sequels to popular games (to reduce the expected variance); I suspect it’s also why, if we go back a decade, new games all cost about $60 even though they weren’t commodities, because that was one way that stores could amortize the risk of pricing games of unknown value. (And then, as the value of those games become known, games that are less valued by buyers have their sales price correspondingly reduced; the used game market isn’t hampered by this uncertainty, so prices float earlier there.)

At any rate, there’s nothing predetermined about the price of a product, at least in the non-commodity case. Sellers can find buyers for a product at a range of prices, but the higher the price, the fewer the buyers. Sellers generally want to maximize their profits: so they try to pick a price at which the number of buyers times the per-unit profit is highest.


To recap:

  1. Prices for products tend to fall in a range between its cost and value.
  2. In a commodity market, prices are at the bottom end of the range, approaching the marginal cost of production.
  3. While making a game, there’s quite a bit of uncertainty as to how much potential buyers will value that game.
  4. Different potential buyers will value the same game differently.

And that second point plays out very differently now than a decade ago. Digital distribution means that the marginal cost of production is $0 in most cases; and the explosion of cheap computing devices mean that the potential target audience has increased by a billion people or more, which in turn increases the number of companies competing for the space, increasing the commoditization of the market. This is a recipe for companies being unable to recoup their production costs, unless they can do something to address these factors; as a result, we’ve seen huge changes on the business side of game production (and of many other markets). So I want to talk about those new models through the lenses of the four points above.


Free to play games are a reflection of the zero marginal cost of production: so, unsurprisingly, free-to-play games are everywhere. (I’m fairly sure that a majority of game titles are free to play these days.) “Free to play” isn’t a business model, though, since it doesn’t explain how companies will survive in the face of that pricing choice.

Ad-supported games are one way to make money: in a commodity market, players may not pay for the value of games, but advertisers see game-players as (just barely) valuable enough to pay for. (As the saying goes: if you’re not paying for a product, then you are the product.) This is embracing games as a commodity; in particular, you’ll see flash portals and ad networks promoting this, sites for whom games really are a commodity.

Microtransactions start to segment the market: separate the players who see the game as a commodity from the players who value this specific game over other games competing for their time. For example, we have energy mechanics: people who are grazing the game will play until their meters run out and then move on to another game, whereas players who value their time with this specific game are willing to pay more. (I see this as bringing in dynamics from real-world activities in physical spaces into virtual games: movie theaters charge you money for using their space to watch a movie, video games charge you money for spending time in their virtual space.) But there are other common forms of microtransactions: e.g. paying to unlock different characters or skins for characters.

Downloadable Content segments the market in a similar way to microtransactions: people who care about this game in particular will buy DLC, whereas people who see the game as a commodity don’t. There are a couple of differences that make me break the two apart: for one thing, DLC doesn’t generally have a recurring cost in the same way that (many) microtransactions do, so the analogy to renting a physical space doesn’t hold. And, for another thing, DLC starts to address the uncertainty question as well: if a game gets no interest on launch, then the publisher won’t spend money developing DLC, whereas if it does get money on launch, then that’s a sign that the game will reward an investment of further development dollars, at which point we have DLC. This is, of course, not a new idea: it’s the same reasoning that led to sequels.

Social Games are has some similarities to an ad-based approach, but with a couple of differences. One is that it sets up mechanisms where your friends are playing the same game as you, which (if successful) gets the game out of the commodity space. And the second is the how the costs of the ads play out: if people aren’t willing to pay money for a product, maybe they’re willing to pay not only in their attention (the ad model) but also in their friends’ attention. One analogue for the former is online games, or indeed multiplayer games in general; one analogue for the latter is multi-level marketing.

Minimal Viable Products are a way to address the uncertainty question: start by releasing a version of the game that is playable and that people can conceivably enjoy, and then, if the game gets popular, devote more resources to it. A lot of Facebook games took this approach, but it hasn’t been particularly common in general; given the phenomenal success that Minecraft has had with this approach, however, I don’t know why it’s not more common.

Patreon puts the value question front and center. While Patreons typically have tiers with some limited rewards, in the ones that I’ve seen, the rewards are more along the lines of thank-you acknowledgements than a serious sales attempt. (The Critical Distance Patreon is a good example: while Jason is, of course, a very photogenic cat, and while I do genuinely enjoy the cat picture e-mail every month, the cat pictures aren’t really the reason why I’m giving Critical Distance $10 a month.) So it’s approaching the question of value and market segmentation in a way that actually ends up setting the market to the side: Patreons generally explicitly ask you to give based on the value you see in the works under consideration, ignoring market considerations.

Kickstarter, however, addresses all four points that I listed above. It asks you to pay for a game that isn’t going to be released for months or years: that alone takes the game out of the commodity space, because if all you needed was a game to fill your time, you’d pick a game that actually existed! The different reward tiers further segment the market: people pay more money depending on how much they value the game. And it addresses the uncertainty question on multiple levels: if people don’t show enough interest in the game to make it worth developing at all, then you learn that upfront without you or potential buyers having to put in too many resources, while stretch goals are a way for companies to decide whether to put out a minimal version of the game or to bulk up the staff to develop something larger.

So, to my mind, Kickstarter is the most promising of all of the above business models, because it has an answer to all of these questions: the ideas there give hope at helping games avoid commodity hell.


Business models, of course, have consequences. Which is fine: constraints can be valuable in an artistic process, and that applies for business model constraints just like any other constraints. And business model constraints aren’t anything new: the business model of selling $60 disks in stores led to a focus on previews (with effects on both game development and game criticism), it led to measurements of games’ values in terms of length, and most companies found it difficult to create evergreen games in that model, turning instead to annual or biannual sequels. But with new business models come new constraints which will play out in different ways.

To take one example, I don’t think game designers, in general, yet have a good answer for how to best meld game design concerns with microtransactions. (Though there are some successes: I’m not a League of Legends player, but it looks to me from the outside like Riot is doing a very good job.) Energy mechanics in particular are frequently cited as something that hurt game design; in general, I think that’s probably true, but I also think that a lot of game developers get their pacing wrong even in the absence of energy mechanics, and part of the reason why I loved Social City so much was related to the pacing suggested by those energy mechanics, so I’m not sure that example is clear cut. Still, microtransactions definitely impose constraints on game design, and there’s no a priori reason to believe that all choices of constraints lead to equally good game design outcomes.

And I’ve seen complaints about (many of) these new business models on ethical grounds. I don’t agree with most of those complaints, but as with game design outcomes, there’s also no reason to believe that all constraints will tend equally towards good ethical outcomes. Looking at the list above, I compared the “advertise to your friends” aspect of social games to multi-level marketing, and part of that analogy is that neither side is a good way to treat your friends. Also, with microtransactions it’s easier to present cost in a misleading way than it is with macrotransactions: I’m a lot less ethically comfortable with asking people to pay $1 for a random choice of one in five items than I am at asking people to pay $5 for five fixed items. (People’s brain wiring generally treats those as similar choices, but the math is in fact very different.) I don’t think this is a tragic ethical choice, any more than crane games or random capsule toy vending machines or collectible trading cards are horrible, but it’s not a great one, either. Hopefully players will continue to get more informed about this; also, compare Android: Netrunner to Magic for a good example of game design alternatives in this space.


With the above as (lengthy!) prologue, I want to return to the Shenmue 3 Kickstarter, explaining my response to the articles by Houghton, Poprocki, and Horvath. (Though, before I go into details, I’ll repeat: I basically agree with Houghton’s claims that the Kickstarter should have been more upfront about funding sources from the start; the Kickstarter fixed that two or three days later, so yay Houghton for helping make that happen.)

Both Houghton and Poprocki were vocal against corporations using Kickstarter as risk management: for example, Houghton said “In a way, yes, Shenmue 3 did need that Kickstarter to succeed, but only because Sony made that the case. It made that the case by making the public pass a test before it offered its own support, a test that it knew would cost the public hundreds and thousands.” In other words, they argued that it is wrong for corporations to use Kickstarter as a tool to tackle the uncertainty that I listed in my third point above, and Houghton at least seemed to believe that Sony had no need to do so, that it was a charade.

I have no idea to what extent Sony was committed to Shenmue 3 even without the Kickstarter. But I am very certain that the uncertainty question in general is a real one: in all the companies that I’ve worked at (in my game company time, in my times in other industries), the question of how to spend resources is huge and always present. Lots of projects are proposed for consideration, fewer of those projects get studied seriously (with prototypes, with focused market research), still fewer of those prototypes make it into serious production, not everything that makes it into production actually gets published, and, of those products that do get published, some get large development and marketing budgets and some get small ones.

Kickstarter can help decide whether to commit to a project; and, once you’ve committed to it, it can help you decide how much money to spend on it. The fact that Sony has a thirty-three billion dollar market cap in no way lowers the importance of this sort of decision to them: yes, nothing horrible will happen to them if they make the wrong choice with Shenmue 3, but if they don’t take a disciplined approach towards evaluating projects in general, then they absolutely can go out of business. In fact, Shenmue itself provides an excellent example of how this can happen: Sega was once one of two dominant console manufacturers, then they made some choices in the Dreamcast era that led to them not recouping their costs on multiple projects (including Shenmue). The result was that the Dreamcast was their last console and their game development business is a shadow of its former self.

This isn’t to say that the articles were completely against Kickstarter as risk management: all three articles were fine with Kickstarter being used to manage uncertainty, as long as it’s the right people using it that way, where “the right people” means individuals and small businesses. And maybe I’ve spent too much time employed by large businesses and by businesses that want to become large, but this isn’t an area where I have particular populist sympathies. Don’t get me wrong, I think it’s great to see Kickstarters cropping up to fund weird projects; but when I went through my list of business model innovations above, Kickstarter was the one that seemed to me to be the most complete answer to forces that companies of all size are facing. So I don’t see why the world would be better to say that corporations shouldn’t have access to that answer: all things being equal, I like the little guy, but restricting Kickstarter just sounds to me like it takes a way a tool that would help corporations experiment with producing a wider range of goods. If corporations can’t manage uncertainty with something like Kickstarter, then they’ll manage uncertainty by making cookie-cutter games, and how is that better?


That’s the uncertainty side of the business model question, but there’s also the value side of the business model question. And all three articles were extremely uncomfortable with how the Kickstarter led to Shenmue fans showing with their wallets that they value the series. Houghton characterized the Kickstarter backings as “donations”; Poprocki took that train of thought still further and said that Sony should actually return the money to backers now that the game has been funded; and Horvath said that Sony is “strip mining” fans.

Recapping my analysis above: corporations prefer for their products not to be commodities; if a product isn’t a commodity, then its price will be in part be determined by its value; and different people will value the same product differently, which raises the possibility of corporations increasing profits via market segmentation. So: which part of this did the authors find so distasteful? Both Poprocki and Horvath focused on $60 as an expected price for games, so at the very least they seemed to not like market segmentation, though they were more silent on the commodity question. (To be sure, given that he runs Unwinnable, I can’t imagine that Horvath wants games to be commodities.)

A fondness for commodities is a reasonable enough stance: I’ve talked about why commoditization is bad for companies, but it’s great for people trying to buy stuff from companies without spending much money! Or at least it’s good in general: where it starts to go bad is where you care for value beyond what the commodity good provides, and where the non-commodity market doesn’t exist. And games feel to me like a market where that problem has the potential of happening: it is very much the case that not all games are created equal, and it’s also the case that there are lots of games that I care about that aren’t necessarily going to win overall market popularity contests. I’d like the market to provide those latter sorts of games to me, but the question is how.

So if market segmentation is the way that that the market figures out that I like specific games, and if that means that games I like turn out to be more expensive, then that could still be a good outcome for me. Also, going in a slightly different direction, some of my favorite games are evergreen; I’d like companies to have more incentive to develop more games like that, and if that means business models other than a one-time $60 purchase, then that could be a good tradeoff for me.

To sum: I don’t see anything particularly optimal about $60 for a game. Please, companies, segment me in the market! Make stuff I really like and charge me more for it! I like that more than a mass of undifferentiated games. I can think of three games that I’ve spent more than $1000 on, and a fourth where I’m pretty confident that I’ll reach that spending level. I’ve had (and continue to have!) wonderful experiences with all of those games, I have zero regrets about spending that money, I only wish there were more games that had given me as much.


I should be upfront, though: I have my own preconceptions when it comes to discussions around value in games supported by new business models. I’ve read too many articles written by people who not only dislike Farmville or Kim Kardashian: Hollywood but who treat people who value those games as either idiots or victims of a scam; I really dislike that sort of attitude, the idea that people who value a game that you personally don’t value are wrong. That attitude is harmful when talking about those games now, it was harmful in the console wars, and I’m sure anybody reading this article can think of other examples where that has hurt our culture. So I’m hypersensitive to that articles that talk about companies abusing fans of a specific game: how much are such articles criticizing the companies, and how much are they instead criticizing the fans?

When I read the three articles in question, I saw a lot of complaints about people pledging more than $60: Houghton wants to make that impossible by eliminating tiers over $50, Poprocki wants the money given back entirely. I’m more sympathetic to Horvath’s complaints about the externally-directed approach that Violet took—that feels off in the same way that the social game “spam your friends” mechanic feels off—but he also says that Violet was wrong to be happy about the Kickstarter. So the implication that I get from all three articles is: you’re wrong to value Shenmue enough to buy more than the game. You’re wrong to want Shenmue figures, you’re wrong to want a soundtrack, you’re wrong to want an art book, you’re wrong to want to play an early version of the game.

And I just don’t get it. How would our world be better if people didn’t have that sort of enthusiasm about their favorite works of art? And how is the Shenmue 3 Kickstarter bad for supporting that enthusiasm?

Post Revisions: