Balancing game economies is an art and a science

Game economies are a delicate thing. It doesn’t take much to break them. If players perceive a greedy cash grab by developers, as has happened with controversies around loot boxes, they may boycott the game. But if developers don’t price the items in a free-to-play game correctly, they may be stuck with escalating costs and insufficient revenue.

From Electronic Arts’ Star Wars Battlefront II to Blizzard’s Diablo III auction house to Bethesda’s Fallout Shelter, we’ve seen the consequences of failing to get this right.

How can developers spread the monetization of game economies across larger numbers of players? Are subscriptions the answer? Or can blockchain help? And what role does account and credential fraud play in further disrupting game economies? Akamai recently sponsored a GamesBeat breakfast that we held in Anaheim, California, to address these questions.

I moderated the session, and our panel of experts included Nelson Rodriguez, Akamai; Spencer Tucker, senior vice president of Scopely; and John Linden, CEO of Mythical. We had a couple dozen game industry CEOs and executives in the audience.

Here’s an edited transcript of our panel.

[Update 9/16/19 — Author’s note: I originally published a version of this story on 9/14/19 that contained material that I had previously decided to delete for both clarity and brevity. A party familiar with the conversation contacted me and noted that the material in the actual post contained information that appeared to be in there by mistake. I realized this mistake and updated the story on Saturday night with replacement text that was shorter and briefer.

The shorter version had material that I did not review thoroughly though. This confused people who saw both versions.

Given the conversation in the player community regarding this topic and after consulting with my editor, we’ve decided to restore the original article with the longer length, as it was spoken at the breakfast. This should clarify what was said. I apologize for any confusion this may have caused. I would note that we always edit our transcripts, often for brevity and clarity. In this case, my editing and communication were flawed. I apologize for this].

GamesBeat: I think everyone knows that building game economies is important. The times when you notice it are usually when something goes horribly wrong. There are a lot of lessons that come out of that, and that’s what we’re going to try to convey today. [Former Blizzard developer], I guess you know about Diablo III’s auction house [laughter]. The original disaster of game economies. But there have been plenty of others, like Battlefront II. All kinds of things related to loot boxes these days. We’re going to talk about the future as well, things like blockchain games.

When we proposed this topic, did you guys have some things in mind that you wanted to bring up about what balanced game economies are?

Spencer Tucker: One that we’re focused on that I think is going to become more important over time is the concept of personalized game economies and that value is a relative thing. Looking at various behaviors on a segmented basis — how much, for instance, on a 30-day rolling period, does somebody spend? And then determining value relative to the amount of money that they’ve spent over the past 30 days on a particular piece of content.

That’s something we’re looking into. It’s a good jumping-off point for conversation. The idea there being that historically, if you’re familiar with mystery boxes, things like that, typically it’s an expected value system. This thing is worth whatever amount of money, and that’s some sort of calculus based on either historical aggregate-level spend, or some sort of calculus between the power of an item relative to the amount that we want to charge for that item, and then we apply that against an entire population. But there aren’t a lot of companies with semi-durable and durable goods that look at the actual player behavior on an ongoing basis to stretch the behavior beyond its natural threshold on a relative period of time.

For instance, if we looked at the distribution of spend over the last month and said, “Hey, someone spent between $1 and $50,” that’s a bucket. We want to determine content to push them above that $50 threshold over a certain period of time. Maybe we price it as a percentage of their last 30-day spend, instead of a fixed actual cost. We do that differently against that entire distribution of the population. The goal there is to move people up in terms of spend velocity between buckets over time and get them more engaged in the purchasing experience, and ultimately treating purchase activity as a retention funnel.

GamesBeat: Some of that sounds like pricing airline seats, different prices for each passenger.

Tucker: Price discrimination is one thing. You could offer the same good at different prices. There are other ways to do it too. You treat value as the relative thing, aside from price. This might come into play in an MMO. You have a lot of people having conversations about how much something costs. If you do price discrimination, you could say, “Hey, I’m going to sell you this sword for $100, and sell this other guy the same sword for $20.” But then those two people meet online and ask, “Why am I paying this much while you’re paying that much?” They don’t have the context to understand it, so you have a negative reaction from that level of transparency.

If you treat the value as the relative thing, the distribution for that value is more probabilistic, like mystery boxes are obfuscated in some fashion. Then you can keep that value exchange behind the scenes. They’re paying the same price, so you attack a price point on a relative basis that moves a lot of volume, but the number of times they have to make the purchase to ultimately acquire that good is variable based on how much the spend over that 30-day rolling period.

Audience: So you’re talking about shifting the odds.

Tucker: Shifting odds, or shifting the–in a practical sort of example, one thing we do is we say–we’re doing this now. We’ll say, “Hey, we’re going to create a mystery bag. We’ll call that bag the same thing regardless of what bucket you fall into.” From a player-facing standpoint, the bag is called A. It’s A for everyone. The price of A is 99 cents. But the thing that varies is the quantity or the range and probability of the range of items within that bag.

The idea there is to optimize around pushing people beyond their natural threshold. If you look at the median value for the spend buckets and divvy up your population in a way that makes sense, you can push people to stretch beyond their natural limit. The goal is to move them over time to become more inelastic, versus being out-priced by attacking just the top spectrum of spend, which is more typically what we do. We push very hard on the very top segment.

John Linden: Are you guys able to publish odds, then? Because if the odds change per player, what’s the legality in Belgium and other countries with new regulations around that?

Tucker: We do publish odds. It depends. Right now we have a number of approaches there. We can class content. We’re publishing odds based on a class. Rarity and things like that.

Linden: But that would still shift per player too, right?

Tucker: It does, but the class is a bucket. Think of it as rarity tiers. Orange rarity has a bunch of stuff in it, but there’s a range within that tier, and you’re publishing the percentage of that range, of the total range. That’s one example. Another example is, if we talked about mystery boxes–this is something we haven’t done yet, but I think in the future there’s alternatives to mystery boxes that are effectively very similar, but function in a very different way from a player-facing standpoint.

If you’re familiar with PvP stores, where you get currency and go and–Galaxy of Heroes has one. You refresh the store to refresh the content in that store. If you were to take the concept of a mystery box, right now they function as a probabilistic system where you know what the range is, but you don’t know the outcome until you make a purchase. If what you did was you took that sort of random probabilistic piece of it and put it behind what populates in the store–imagine a PvP store where I know exactly what I’m going to buy every time, because I pay a fixed cost for any particular item that shows up in the store. But when I refresh, what shows up in the store is determined by a table.

That becomes the mystery box. The table itself generates a range of possibilities, and you know if you get that chase item, it’s going to be 99 cents, the same price as everything else. But you don’t know when it’s going to show up. Then you have an incentive to refresh. If you build an added value on the refresh, like progress toward some ultimate goal–it could be a jackpot mechanic. Then you’ve incentivized people to refresh so they’re not losing value. They’re only buying things that matter.

The good thing from a game economy standpoint is, rather than people getting stuff that has “no value,” but does have value earlier in the experience, or for a different population, you’re basically preventing people from getting that content as junk payouts. What they’re doing is only buying the stuff that they actually care about. I’m only paying 99 cents for this consumable item, because I actually need that item, rather than being given thousands of those items over many, many purchases.

Audience: You’re effectively training them to keep going back.

Tucker: Refresh until you see what you want and buy the thing you want. Chasing the thing you want becomes chasing the refresh, and refreshing that box also gives you progress against some sort of milestone.

Audience: Is this all theoretical still?

Tucker: We’ll be building that. We haven’t built it yet.

Audience: Does the reset cost money?

Tucker: The reset will cost money, yes. That’s the key. We’ve done other things. There are other plays on–things that we do have, that we have built. For instance, rather than a mystery box–one thing to understand, in a lot of mobile games they’re not really true MMOs. Economies are individual things. Saturation hits individual demand, because there’s no concept of an aftermarket. There’s no trade, no real inflation in that sense unless you have a true aftermarket.

Part of the problem there is, how do you account for demand and decaying demand over time in a system like that? These mystery boxes, we adjust odds and so on based on aggregated point view, but not on an individual basis, because we don’t really have demand on an aggregated basis. We have it on an individual basis. If we built something that was less of a mystery box and more of a finite pool, so you control the quantity and effectively create scarcity–for example, if you’re familiar with capsule machines, you put in a quarter and there’s a fixed number of items in there. Maybe you want one thing. Imagine doing that and then pooling the finance segment of the population. We’re going to say, “These two factions, or these 300 players,” and you pool them all against that fixed pool, now they’re all competing effectively for the same content pool. Then you adjust relative price based on the expected value of that pool. There’s a fixed number of people who get that chase item. You’ve created a social competitive purchasing mechanism.

Audience: Do you go as far as something like an auction for that item?

Tucker: It’s not an auction. It’s a system where you buy content, and as that content is purchased, it’s removed from the pool. The idea there being that because you have a whole bunch of people competing for that finite pool, that demand is going to be higher and you’re going to create more spend activity.

Breaking the ice is the challenge there. We have what I call a social network and externality component. Let’s say that this table is faction. You’re all individuals playing the same finite pool. We’re all competing for the same content, but we’re in a faction. For everyone in that faction, to incentivize both faction participation and encourage the behavior of participating in this particular monetization mechanic, what I might do is say, “Any one of us who makes a purchase, everyone at this table gets a fraction of a free pull, regardless how much they’ve spent.” For every purchase there’s a fraction of a free purchase being built up.

If you have a bunch of free purchases built up, what you want to do as table–we all want us as a group to have a competitive advantage for that piece of content. We want to coordinate when we spend our money. We’re going to pay attention to the pool. We’re going to pay attention to who within our group is making the purchases, how many free pulls we have, and then try to coordinate a burn to acquire that content before the box resets.

Those sorts of mechanics are things that we’re going to see more of — the social component, the proxy spending pressure. You’re pushing people to stretch beyond their normal capacity to spend by spending more on behalf of other people in the group. Then the dynamic that exists there is, people who don’t spend will apply pressure to people that do spend, because everybody benefits from that activity.

Then you’re controlling the amount of content you give out, so there’s true scarcity in that sense. We know that once that box has been drained, somebody will have everything that was in that box, whereas with a mystery box currently, there’s no guarantee that anyone will ever get something in that box unless they happen to roll the right way. We base that on the probabilities of that box, but there’s no guarantee.

Audience: Is there any grinding in this model you describe?

Tucker: There’s definitely a grind. These are distribution methods of content. I’m not saying that we would remove grind or anything like that. I actually think–we were just talking about this earlier. In the future we’ll be more into this. When you think about the distribution of payers and general population of players, it’s like 95 percent, possibly even more, players never spend. The payers are a very small group.

One of the examples that we were talking about earlier, back when I was working on PC games, we had a game — this is a story that illustrates the point — where people were packet sniffing and injecting gold and selling gold. That’s very common. The economy truly did inflate out of control. The secondary market, because this did have an aftermarket, was basically destroyed by the devaluation of that currency. The players organically determined that gold wasn’t worth anything anymore.

They moved to something like a gold standard, something with intrinsic value. They moved to a consumable item, where the only way to get it was to purchase it. The intrinsic value of this item was that it improved your probability of linking a piece of gear with a gem. In Diablo it would be rune socketing, except this one had a failure system. You’d break the gem if you failed. This improved your odds. There’s a true intrinsic value of this item.

The downside of this–great, we sell it. We saw a spike in sales of this item. But not necessarily a good result, because we’d inflated the economy to hell and this was the reaction. We said, “Well, there’s a lot of fraud. There’s no management of that aftermarket.” Plus, from a business standpoint, every time that item changed hands, we saw no value. If somebody quits and they sell their gear, they’re selling it offline. They’re trading it to someone else. That person has no demand for gear in the future and therefore they’re undercutting us on the value of those items over time.

What we ended up doing was building a secondary auction house system with a tertiary currency, and that currency had intrinsic value in that it could be used to purchase items from a store that was exclusive to that content. That was also the only place where you could buy the aftermarket currency, these consumable hammers that helped you make gear. Then we taxed the transactions. We charged a listing fee and a percentage of the rate of exchange. That’s how we started making money.

GamesBeat: Some of the things you’ve talked about–you covered a lot of ground already. But it reminds me of the past. You’re focused on Star Trek and other mobile games. [Former Blizzard developer] used to work at Blizzard and was there during the Diablo III auction house days. I don’t know if there are some things that Spencer just said that brought back memories for you. Generally people know that something went wrong with Diablo III and Blizzard had to shut it down.

Audience: I can talk about that story. A lot of people get fixated on the fact that we did a real money auction house for Diablo III. But I don’t think that was the real issue, adding any sort of auction house to the Diablo system of rewards. The real problem was we didn’t beta test the game in the way we should have to figure out how the economy was going to shake out.

In Diablo II, we only did a stress test beta test. We were really worried about giving away all the single-player content. So since that worked for Diablo II, that was what we were going to do for Diablo III. That was our mindset. Our mindset should have been that we were living in a post-World of Warcraft world, and people were going to play Diablo III like an MMO, even though it didn’t have an MMO model to it. We ended up putting in the auction house. We knew it was going to have a big effect on the game. We didn’t know entirely how. But we only stress-tested the game as a beta test. We didn’t play long enough for the game to turn into a true MMO-style economy.

The problem with the loot-reward-incentive loop with Diablo is that so much of what makes that game fun is finding upgrades to your gear through killing monsters in the game. What happened with the auction house is that people would go there, even if they weren’t spending real money — because you could also use virtual currency in the game — and use the virtual currency to upgrade all their gear. Then the loot-reward loop in the game–people would never find upgrades. The game just felt terrible. It felt broken. There was a lot of rage. From a community standpoint they fixated on the real money part of the auction house, but it was really just the auction house itself that was a problem.

I think we could have fixed it and made it interesting in the same way we did in World of Warcraft. In World of Warcraft we had a lot of the loot in the game cordoned off. We called it prestige loot. You could only get yourself if you killed a certain monster. It made it so people wanted to do the raids and high-end content. Then there was a subset of loot that we made freely tradeable, and that turned into the auction house economy.

We could have figured something like that out for Diablo, but the problem was, since we didn’t have a really long beta test like we did with World of Warcraft, we didn’t figure out all those designs. By the time we figured it out, we already had a horrible launch. We had a lot of player rage and the community pointing out a variety of different things that happened with our launch, one of which was the real money auction house. We didn’t feel like it was worth keeping in the game.

We’d rather get rid of it and get some community trust back, rather than mess around with the live version and try to correct it.

Linden: I was trying to figure out how to transition from what you were saying to what we’re doing, because it’s kind of the opposite. At Mythical we’re focused around some of these issues. The last game I did was Marvel Strike Force, a free-to-play game, and as he said, it was 95 percent unpaid. The other five percent spent hundreds of millions of dollars last year, which is great.

I started finding it really fascinating, looking at that 95 percent. How does that group of users grind and make money? They can sell something. It’s not that we’re making gamer millionaires out of this. That’s not the goal. But what happens if they put their effort into it, sell something, and make $10 or $20 or $50? What happens to that segment? These are the challenges I think are fascinating. Do they take the money and run? Probably not. Some of them will. But the others will put it back in, again and again. Almost creating gamer entrepreneurs. That’s fascinating.

What we’re looking at now is primary and secondary markets. We are using blockchain. We get called a blockchain company quite a bit, but we’re really not. We’re using blockchain tech behind the scenes to do real-world currency transactions. What we love is, if you give game developers tools that have primary and secondary combinations, there’s some really cool stuff that can come out of that.

One thing we’ve been looking at a lot–we’re not looking at loot boxes at all. We’re almost adopting the sneakerhead model, the Supreme, Bathing Ape, all these brands coming up with the new generation of consumers. They look at everything in terms of rarity, in terms of time-based. We have our first game coming out on our engine, and we’re going to sell everything time-based. We’ll tell you the rarity once it’s gone. That becomes fascinating if there’s a secondary component.

If you sell a Marvel branded item and have a million sales, your secondary transaction volume is going to be low. The value is not going to go up a lot. But if you sell a small item from Supreme and it only sells on a Saturday for two hours, and you only sell 5,000, the secondary market could skyrocket. That’s what I think is really fascinating about where we can start slowly shifting. That’s one combination we’re doing.

We’re also kind of using this blockchain thing out there–we’re not doing cryptocurrencies, but we use blockchain tech to enable that secondary market. You mentioned, how do you have secondary transactions and have liquidity for players anywhere? Whether it’s in the game economy itself, in an auction house in the game, or if it’s on eBay or a crypto exchange, how do you do that and still get your transaction fee? How do you represent the game asset outside your game?

That’s the other concept we’ve been really focused on, a concept called “D-goods,” which stands for digital goods. That’s the other piece. If you can ensure that there is a fee for that transaction, no matter where it sells, whether it’s player-to-player, in the game, out of the game, you can use blockchain to ensure that transaction fee is collected, and it’s recorded and transmitted. You can do some cool stuff.

That’s the big piece we love. We also love the idea around bringing content into the game. We talked about that. Everybody goes immediately to UGC. Nobody likes UGC in their game. But it’s not really UGC. We’ve talked to a lot of big groups. It could be esports teams. For our former employer, it could be these $20 million franchises. If they spent $20 million on a franchise, maybe that’s who the audience is. Let them bring their content. Maybe it’s part of that franchise fee for a game. They have the exclusive rights to bring in content of their own design to a game. It could be a lot of different things in terms of bringing content in, whether it’s brands or influencers.

Allowing influencers to give away items to their top followers–even though they don’t get money from that primary sale, they get a cut of every secondary transaction that comes off that. That’s suddenly going to create really amazing marketing opportunities. It’s going to create loyalty between fans and influencers. It’s going to create loyalty between influencers and the game, because now they’re able to profit from the games they’re promoting.

These are the things we love, and I think it’s a whole different style of economy. We hear a lot of this, “Blockchain will be revolutionary for gaming!” Eh, bullshit. It’s going to be evolutionary for gaming. It’s going to allow us to do some things with a new set of tools to design the game the way we want to design it. That’s what we’re excited about, what we’re up to. It’s a bit different, but I think we’ve all seen a lot of these principles.

Audience: I love this concept of managing scarcity by being in the know. What makes the sneakerhead thing work, you’re in the know. You know where to go to a specific shop when a specific model comes out at 2AM. I’m curious how you think about that aspect of feeding the economy through communication. How much do you do?

Linden: This is where the blockchain stuff becomes interesting. You can now have a limited set of items that are individually numbered. If you think about physical collectibles, they have these certificates of authenticity. That little hologram that says you have number 14 of 62. Sometimes that has value in and of itself. My favorite number is 22. I want number 22 of everything. There’s more value in that particular item. You can almost let your partners determine the time.

The first game we’re looking at, we can bring in brands. They can sell into our economy. They get a cut of primary and secondary. The brand can help decide some of that, which is really interesting. We’ve talked to a lot of brands, and they’re really tied into that time. You might have some brands where all they want to do is maximize primary sales. Warner Bros. could say, “We have a new movie coming out, so we want to push as much as we can.” But other brands will look at rarity, or even a combination.

What if you had a Batman-branded item that’s a primary sale — it’s on sale for the entire season, three months — but then you have a limited edition, one of a thousand, that’s fixed and on sale for one day? Then you have another one that’s on sale for two hours and you sell as many as you sell. You can start mixing these economies together with the same brands and the same game economics.

I think it’ll start shifting our game economies into how retail thinks about their businesses. We saw that a bit with Call of Duty. I was a studio head at Activision for a while. At one point there was a Marvel deal. They were trying to get the Winter Soldier or somebody like that in the game. It’s tough to get Activision to do those kinds of deals. They might be less tough these days on this stuff. But there was a lot of discrepancy on pricing. Marvel thought the value of their character was greater than a traditional soldier.

What if you could have the toolset to let them take the risk? With this type of model, charge a wholesale price and split the upside. If they want to take a risk and say that a Call of Duty character costs $9.99, we collect the $9.99 and you price it at whatever you want. You take the risk. It turns it into almost manufacturing. We get paid no matter what, and if the brand thinks they can make more money, you give them the upside. So many new sets of tools come out of this transition. I think it’s going to be awesome.

Audience: It’s an interesting evolution of the license model. What you’re proposing has a lot more in-between.

Linden: You have a lot of options. The other one we’ve seen–we saw this with Marvel again, oddly. We did this last game with them, and again, they want to push the value. They say, “If we’re going to push this type of asset, we think it’s going to be the most valuable in the game two years from now.” They think through some of that stuff. And again, you can do that. Go ahead and sell how you want. If it skyrockets in the secondary market, if it becomes worth $1,000 apiece on eBay, Marvel gets a cut of that. There’s a lot of cool options you can start playing with.