Design a social game economy: how much should a virtual t-shirt cost?

Matthew Warneford

What have FarmVille, FrontierVille, and Facebook’s most popular game, CityVille, got in common? All made by Zynga, all very popular, all making barrels of cash, all have the suffix “Ville”… so lots of things in common really. But for this post, we’re interested because they share a game design built around a virtual economy. Zynga know that a well designed virtual economy will make a game more fun, players will stay longer, and more of them will pay. That’s the good news. Get the economy wrong and even a great game won’t make money.

In this post I’m going to show how to use a ‘scientific’ approach to design a game economy – where ever possible we try to use data models to set the prices of virtual goods rather than hopeful guessing!

Before we dig in deep, lets look at the process from high up. Step one is to identify the important variables and then create a model for how they all fit together. At first those variables are assumptions – we guess how long player will spend in the game, how quickly they’ll learn the rules, how long before they invite their friends, and so on. Step two, when the game goes live, is to collect actual usage data and plug those real numbers back into the model. Pretty quickly we can see where the problems are and what to do to increase conversions from playing to paying.

We don’t expect the first economy model to be perfect, but having a model means you can measure, improve, and get much closer to perfection far quicker than hopefully guessing.

But what do I know about building virtual economies? My company, Dubit, have been creating social games for over 10 years (we launched our first virtual world right before the dotcom crash – great timing!) Since then we’ve created the BBC’s first virtual world, Monty Python Ministry of Silly Games, The Wizard of Oz’s Adventure in Oz, and many others. We’ve learnt a lot, created our gaming platform, designed tools, and share our experiences on our blog.

Step One: Identify the Important Variables

The game economy sets the price of virtual goods, playtime between experience levels, amount of experience earned per hour, amount of virtual currency earned per hour, and even the number of virtual goods. Our free tool (screenshot at the top of the post) mashes together all the different input variables and produces four graphs showing the rate of progress through the game, and how the price of virtual goods change over time.

In this first post I’ll explain what the important inputs are, and how to use the tool. The next post will explain step two – how to use the graphs produced by the tool to improve your game.

There are about six different inputs that combine to make what we call the “progress curve”. Everything in the game economy hangs off the rate that we allow players to progress through the game. By progression I really mean how quickly players level up, because each time they reach a new level they unlock new items to purchase in the virtual stores, new powers, or even new features in the game. Levelling up is the measure of in game progress.

In the beginning we want players to level up quickly, we want them to feel like they’re progressing fast, and we want them to feel like they’re great at the game. This is known as False Achievement. But gradually the rate of progress slows down – it takes a little longer to reach the next level, get the next cool virtual tractor, or access a new part of the game.

Slowing down the rate of progression creates reasons for players to spend real money. Maybe they’re falling behind their college buddies, they’re impatient, or want to unlock that great virtual gift in time for a friends birthday. Whatever the reason, some players will want to progress faster than the game allows and will happily spend money to do so. This desire to progress faster is the number one reason players spend money in a social game. The difference between a successful game and a flop is that the successful game has a progress curve that finds the right balance between too fast and too slow.

Our tool needs only four input variables to define the progression curve:

  • Time before a player is primed to spend
  • Time saved by spending money
  • Experience levels progressed by spending money
  • Experience level after one hour of play time

Lets look at each of these four variables in turn, starting with “Time before a player is primed to spend”.

As we said earlier, the desire to progress faster is the number one reason players spend money in a social game. In other words, casual games sell time. The sales pitch is quite simple: play for 10 hours to get to the next level, or spend $5 to get the magic something or other and get to the next level much much quicker. $5 saves time.

However, this sales pitch only works when players actually care about saving time. After all, its a game, its fun, what’s the rush to save time!? Although some people are “completers” – they get their kicks from finishing the game – most people need to be given a reason to progress faster. This is where social kicks in.

Successful games encourage players to invite friends not just because people have more fun with their friends, and more friends mean more viral growth, but because those friendships create a little bit of competition. That competition, the social pressure to keep up, gifts from friends, and so on, all tip players into payers.

In summary. Players will spend money in a social game so they can progress faster than, or keep up with, their friends.

We can say that a player is “primed to spend” when they have friends in the game and feel a social obligation to keep up. At this point the game’s progression curve needs to flatten out – takes longer to get to the next reward – so that spending $5 to save time and progress faster becomes more and more attractive.

It takes time for players to get to the point that they’re ready to spend money. No-one starts playing a social games because they’re just looking for more ways to get rid of all their burdensome money! No, they need a reason to part with that cash. In our experience a little lite social competition will be the main driver. But, every game will have different steps and different features that get players from newbies into payers. Think of this as a sales process, split each step out, and design your game to move players through each step. This one variable could be an essay topic all by itself! I’ve already written a little on the subject here Converting players into payers in 5 steps, but if you’d like to discuss more do send me an email or a tweet.

Having figured out each of the steps, and the features that make up each step, it should be possible to estimate how many hours of gameplay before a player is ready to spend.

The second and third variable go hand in hand, “Time saved by spending money” and “Experience levels progressed by spending money”. Having got the player “primed to spend”, you now need to offer them a good deal. Spending $5 to save 15 minutes doesn’t sound like a great deal to me, but spending $5 to save 4 hours and progress 3 levels sounds good, but maybe too good!

The offer really depends on the target market. Children have more time than money, so they expect to progress further and faster for their dollar than busy working parents. You’re never going to guess these numbers perfectly, but thats OK. Remember, when your game is live you’re going to tweak and test what offers and what numbers produces the best progression curve for converting players into payers.

Lastly, the fourth variable, “Experience level after one hour of play time”, is quite simple. We’re looking for a measure of how quickly the player progresses after the first hour of gameplay. This is easy because you can look at other successful games targeting a similar demographic and play them for an hour. What level do you get to?

Remember, that first hour of gameplay is when we’re trying to hook our players – it’s that False Achievement again.

Plugging those four variables into the tool we’ve developed we get a curve that looks something like the one below.

The “Tutorial” and “The Sell” are marked by thick black dashed lines. Those two points are defined by the four variables we’ve just discussed. Our tool uses the Newton–Raphson method‘s_method for finding successively better curves to fit through the four points. That’s why we wrote this tool – it’s a real pain to write in excel!

From the progress curve our tool can also show how many experience points are needed to get to the next level, how many game play hours to the next level, and how many levels progress the player buys at each level. All we need to do now is work out the cost of a virtual item.

This step is actually fairly simple. Start by deciding what is a good reward for levelling up, should players get two virtual items, or 22? For each level the player should unlock new items in the virtual shops, but how many should they be able to buy?

However, virtual goods can usually be classified into two buckets, “wants” and “needs”. Goods in the “needs” bucket are required to play the game – these might be seeds in farmville. Whereas, those in the “wants” bucket are decorative – these are items purchased to show off, we want them because they show how well we’re doing.

When we talk about the rewards for levelling up we’re talking about virtual goods that fall into the “wants” bucket. Often each level will unlock more “wants” than the player will be able to buy. For example, if we think the player should be able to get 5 new virtual goods per level, then we might actually unlock 10 per level, but on average they’ll be able to afford 5 of them.

Now we know how many virtual goods a player can buy per level we just need to set the price. This is the easiest bit of all. For each level the player has to earn more experience points than the last. All we have to do is design the game to always award the player virtual currency whenever they earn experience points. The amount of virtual currency per experience point should always stick to a ratio, for example 5 virtual coins for each experience point.

By plugging that ratio into the tool (along with the number of virtual items the player can afford per level) we can calculate the average price per virtual item per level. As time goes by items become more expensive, but thats exactly what you want! Hopefully you’ll now have a curve that looks something like the one below.

Other Articles