robert zubek / blog

Casual games are fun to play and fun to make, but can a developer make a living from them? Funding the casual business was one of the prominent questions at this year’s GDC. I want back and forth between the Casual Games and Worlds in Motion summits, and managed to gather a few interesting tidbits. Here are my impressions.

The popular funding models are: 

• Indirect ad sales - ad rotator networks, etc., which let people participate in a larger pool of ad space providers
• Direct ad sales - such as doing a deal with a specific client, for banner or in-game ads
• Shareware model - aka try-and-buy, typically with a content lock (e.g. first level free) or a time lock (first hour free)
• Subscription - the standard MMO approach, pay per month
• Virtual goods model - aka free to play, pay for items

Of these, advertising is the most popular entry-level option, but it has its pros and cons.

Indirect ads are the easiest to get into - just sign up with AdSense or one of the other ad networks, and watch the money come flowing in! Except apparently it’s not that much money. People were mentioning ad revenue figures on the order of a few cents per thousand impressions. ¹ 

Direct ad sales seem to fare better, since you can deal with clients who want to target specific demographics. In their Worlds in Motion session, the WarBook devs praised this model. But they also admitted to 700 million pageviews per month (across all of their games), and their access to players’ demographic info makes them very appealing to advertisers. A brand new project won’t be able to play in that league.

The other popular model is shareware - downloadable games with free trial versions, and full versions available for purchase. It’s a classic model (Doom and Quake come to mind), with the trial version typically limited in content (e.g. first level or chapter free) or game time (e.g. first hour free).

This model was the focus of much discussion, since it’s the primary model for the established players. It’s clear that it can do wonders for hit games, especially since the full game price is customarily set at $10 or $20. But standard games don’t fare as well. Only ~1% of players end up buying the full version, and even then, distributors take a significant chunk. Additionally, most downloadable games get their players from portals, which exhibit significant churn due to a large number of very similar games being created - so differentiation and getting to the top of listings are both difficult.

The other two models are subscriptions, and “free to play, pay for items.” The latter has worked very well for us, while the former works for others, including the 800-pound gorilla of MMOs. Both systems have a significant optimal case, however - they’re best for games where players are compelled to keep coming back to the game, over and over again. Fitting this against ”consumable” games is non-trivial.

So the dominant model is still content-limited shareware, but it’s not an optimal case: budgets can be brutal unless your game becomes a huge hit. But banking on blockbusters is a classic hit-driven studio funding scheme, with familiar drawbacks, only recapitulated on a smaller scale. A more stable model is something like subscriptions or virtual items, which scale nicely, but they require a larger, continuous kind of a game.

---

1. This roughly fits with announcements I’ve seen elsewhere, e.g. MSN Games offering developers 10% of ad revenue from the games they publish.

Viral Coefficient and Growth

The viral coefficient is a measure of how many new users are brought in by each existing user. It’s a quick and easy way to measure growth: if the coefficient is 1.0, the site grows linearly, and if it’s less than that, it will slow down. And if the coefficent is higher than 1.0, you have superlinear growth of a runaway hit.

In an invite-only situation (e.g. gmail closed beta), it’s easiest to calculate this directly, based on how many people are being invited by each new user, and how many of the invitees create new accounts themselves. The viral coefficient is simply:

v = new user invites accepted / new users
   = accepts_t / δp_t-1

where δp_t denote the number of new users who join in time slice t (ie, the increase in population between p_t-1 and p_t). 

But most sites have an open account creation policy. For those, we’ll have to estimate population acceleration from raw population deltas. Let’s assume that each accepted invite is quivalent to creating a new account at the next time slice. Then we can estimate virability as:

v ≈ δp_t / δp_t-1
   = (p_t - p_t-1) / (p_t-1 - p_t-2)

which is an acceleration metric, easy to compute from historical data. 
 

Population Forecasting 

Viral estimate calculated as momentary acceleration will fluctuate over time. But we can use it for some short term forecasting.

To calculate expected future population p_t some t steps in the future, given the viral coefficient and present population p₀, we first invert the above:

δp_t = v δp_t-1 = … = v^t δp₀

and plug this right back in:

p_t = δp_t + p_t-1
    = Σ_k≤t  δp_k + p₀
    = Σ_k≤t  v^k δp₀ + p₀

This describes a geometric series. When v ≠ 1, p_t  = δp₀ (1 - v^t+1) / (1 - v) + p₀ 

AIIDE starts tomorrow at Stanford. We have a very impressive array of speakers this year, and an intriguing collection of papers. Hope to see you there!