Metrics That Matter For Social Gaming Investors

by Guest Author on June 9, 2010

This post is written by Guest Author Byrne Hobart, a marketing consultant at NYC-based Blue Fountain Media. Blue Fountain Media helps clients with website design & development, online marketing, graphic & logo design and more.  In this post, Hobart explains in detail the metrics driving social gaming companies like Zynga and Playfish.
Traditional investment analysis tries to boil a company’s value down to some simple numbers, like earnings or free cash flow. On a slightly more advanced level, some industries have particular ratios that give a more detailed picture of a firm’s operations—in insurance, for example, the “combined ratio” measures a company’s operating efficiency; in trading, income divided by Value at Risk shows whether or not a firm is making safe bets.

In social gaming, the relevant number is different ratio: the company’s churn rate compared to their viral coefficient.

Churn rates are familiar to anyone who has analyzed a subscription-based business like cable television or phones: the churn rate is the percentage of customers who will stop using the product in a given month. In social gaming, that means the percentage of users who will stop playing a given game in a month. From the analyst’s perspective, every new customer is a rapidly ticking time-bomb; they’ll get bored fast, and move on to another game. Some social games have churn rates of 50% per month, or more; if their revenue per user is $5, and the churn rate is 50%, the expected value of each new user is $10 ($5 + $2.50 + $1.25… technically, you would discount these future values based the time-value of money, but since the relevant numbers are in the first few months, it’s not worth it).

The viral coefficient is in some ways the opposite of the churn rate: it’s average organic growth rate in users in a given month. If 100 Farmville users are likely to cause five of their friends to join in a given month, that’s a viral coefficient of 1.05.

Combine these two numbers, and you have the “Expected user-months per new user”—for each user, the total number of months you expect to be played by that user, plus the people they recruit.

In the case of a 50% churn rate and a viral coefficient of 1.05, this means each new player is expected to generate 2.10 months of play time. At a viral coefficient of 1.10, that number jumps to 2.22. At a churn rate of 40% instead of 50%, and a coefficient of 1.1, the number of player-months per new player is 2.92. Keep in mind that these increases can be multiplied by the game’s average monthly revenue per active user. In other words, a bump in player-months per new player adds directly to revenue, at a gross margin of basically 100%.

One thing this illustrates is the strong economies of scale in the social gaming industry. Zynga can afford to invest millions of dollars in making Treasure Isle marginally more addictive; going from 2.10 months of play to 2.92 months of play on a userbase in the millions will more than pay for itself. And the techniques a company develops in one game can be easily applied to others (while the biggest games are months or years old, the biggest launches are more recent).

Finally, quantifying this number can also show a company when it makes sense to start spending money to acquire new users. There’s a reason Zynga can afford to blanket Facebook in ads—once they’ve figured out the average value of buying one new user, they can fine-tune their bids to maximize profitability.

A more robust model would also consider the total size of the market. Even a massively popular game like Farmville can eventually reach all the people it’s able to reach; while it might take several Farmville requests before a given user signs up, but after a certain point they’ll be desensitized. Thus, a site with high virality and high churn will peak faster and fall faster than a site with low virality and low churn, simply because it will be seen, used, and then discarded by more of its total audience.

(It’s likely that the narrower a game’s target audience is, the lower its virality and churn rate are. Tens of millions of people play Farmville, Zynga Poker, and Treasure Isle, and they get bored of it quickly. But a niche game like Fashion Wars will not get shared as aggressively, but will provide deeper interactions between the friends who do play together.)

There is a wide range of actual values that can make a game a winner, for example:

  • A game that spends $5 to acquire a user with a 50% churn rate, a viral coefficient of 1.05, and revenue per user of $5 will earn about $5.52 in profit per user acquired.
  • A viral site with the same revenue per user, but a viral coefficient of 1.2 and a churn rate of 20% would earn $51 in revenue for each new user acquired (this kind of math explains why Groupon can spend so much on Adwords and Facebook ads).
  • A site with revenue per active user of $5 and a churn rate of 40% with no viral characteristics will lose money paying $13 per new user.
  • A more traditional subscription-based business with a lower churn rate and a very modest viral coefficient is also worth considering. For example, TheStreet.com’s monthly churn rateis 3.8%. If their viral coefficient is 1.005, and their monthly subscription averages out to $80, their expected revenue per subscriber is $856—in the first year alone. But viral effects account for only $23.15 of this, meaning that TheStreet won’t significantly benefit from enhanced viral effects.

When social media companies start to go public, I believe that investors should call for them to prominently disclose these numbers. Giving investors an idea of how viral a product is, and how high its churn is, can tell them how quickly to expect it to grow, how soon it will peak, and whether or not they should expect the company to lose money early on in a race to acquire customers. In the case of a viral, low-churn game, it’s irresponsible not to run at a loss early on in order to acquire players. But once a game starts to peak, and the churn rate increases while the viral coefficient approaches 1, paying for growth becomes a similarly poor decision.

Social gaming companies have shortened the fuse and narrowed the range of outcomes for subscription-based businesses. Someone operating an online game can use simple measurements like viral coefficients and churn rates to determine exactly where they’ll earn the most money. A traditional game company (not to mention a phone or cable company) could spend years trying to determine the average value of their customers; for Zynga and the rest of the industry, knowing this information is almost automatic. For their shareholders, disclosing this data will make it easy to tell whether the company’s growth represents free cash flow in the future, or merely an impressive number of users in the short term.

{ 2 comments… read them below or add one }

1 Karen Sweet November 9, 2010 at 9:13 pm

Can you please tell me how you calculated the number of player-months per new player in paragraph 5? How did you combine the Churn Rates and Viral Coefficient?

Thanks for your help.

2 Social Media Colombia November 16, 2010 at 8:48 pm

Great post, valuable information. The key of this business is Customer Adquistion Cost and Life Time Value of a costumer, now I undertand why Groupon ads are everywhere

Leave a Comment

{ 2 trackbacks }

Previous post:

Next post:

 
 
 
© 2010 SecondShares