Posted by (Visited 59769 times)  Game talk  Tagged with: , ,
Mar 162009

At SXSW, Susan Choe of Outspark stated (and was promptly tweeted as having said) that Outspark manages to get $60 ARPU, which is quite a respectable figure.

The challenge there, of course, is that you have to know on what basis this is being measured. I immediately asked “was that ARPU or ARPPU?”

Average Revenue Per User is generally calculated by taking all of your users and dividing your revenue. In a freemium model, that tends to yield (for successful businesses) figures in the 50 cent to $1.50 range. (For more on freemium models, I highly recommend Andrew Chen’s excellent blog Futuristic Play.

Average Revenue Per Paying User is calculated by dividing up the revenue amongst the users who paid anything at all. This yields a figure that is significantly larger. In the case of a subscription MMO, the ARPPU, measured by accounts, is the subscription price, diluted slightly by free trials.

Historically, the histogram of revenues per user have shown a (surprise, surprise) power-law distribution, with a few users willing to pay exorbitantly high amounts. The Simutronics games make much of their living from the dedicated users willing to pay a lot of money. During the pay-for-time days in the AOL era, these high-end users had a price tolerance of thousands of dollars monthly.

Even today, when so many users have multiple accounts, it is more accurate to say that the ARPPU for a subscription game is some multiple of the subscription price; if your average number of accounts per human is actually two to three, then your ARPPU is more like $30-$45.

If you add onto this the likely amounts being spent buying items and virtual currency on the black market, then it’s likely that the average spend per user in the MMORPGs is actually right in the same range as the figure Outspark cited. And in fact, follow-up questions determined that Outspark was indeed talking about monthly revenue per average paying user.

The big reason to go with a microtransaction model is because it opens up both end of the curve. People who would not be willing to pony up the full $15 a month are enticed to pay at least something, thereby hugely broadening your market. And on the high end, users who were getting a phenomenal deal at $15 a month are instead offered avenues to spend the thousands that they were actually willing to spend on their hobby all along.

The big reason not to go with a microtransaction model is that it doesn’t have the commitment fallacy in its favor. Once we make a choice, we tend to stand by it. A subscription can be seen (like buying a box, also a hugely powerful psychological buy-in tactic) as essentially making a commitment; you then attempt to live up to the commitment, which is why two games of comparable quality will have different first month conversion figures depending on whether the user had to buy a box to get in or not.

Every time you pay you have to make a fresh decision, whereas many customers typically let subscriptions ride; in fact, there’s considerably more psychological effort demanded to cancel participation than there is to join in the first place. There’s a lot of social proof backing up membership (friends, community, etc), and you tend to build mental justifications for remaining a member.

Check out this anecdote from Robert Cialdini’s wonderful book Influence: The Psychology of Persuasion. Here a group of folks attended a session trying to sell them on Transcendental Meditation. If they signed up, they had to leave a $75 deposit. A logician stood up and made a rational argument against doing so; as a result, more users signed up. The logic?

The spokesman put it best: “Well, I wasn’t going to put down any money tonight because I’m really quite broke right now; I was going to wait until the next meeting. But when your buddy started talking, I knew I’d better give them my money now, or I’d go home and start thinking about what he said and never sign up.”

As Cialdini describes it, this potential customer had a real need, and they saw an opportunity to commit to it. The argument against actually served as an argument to take the leap, because it undermined their hope. And hope is a powerful driver — exactly the driver that subscriptions make the most powerful use of.

There’s a reason most mature industries (phones, cable TV, even theme parks!) have evolved to support both subs and pay-as-you-go a la carte plans. Your ideal is someone who is committed enough to subscribe, and who also then accepts upsell offers to maximize their revenue potential. In the meantime, we have to listen for the subtle difference between ARPU and ARPPU.

  15 Responses to “ARPU vs ARPPU”

  1. I´d like to see more attention given to alternative forms of microtransactions that involve a different kind of congitive model (and therefore business model) than the current crop, which is generally using variations of the Diki dopamine dropper punctuated with racheting dings. You know what I mean. I designed a game that microtransacted a virtual currency used to enter games where players could steal a greater sum of that virtual currency, potentially gaining a profit. Incidentally, all markets in this modern, electronically settled era, operate on the same principle, and the operators of those games more more money when participants make more discrete actions. There´s tons of potential there, how ironic that a fledgling industry trying to quantize profits would ignore profit-seeking itself as a viable vehicle.

    Then there´s gotta be an approach that uses social reinforcement. I once advised a guy from to play up group dynamics and rivalries between the two, and sell people vengence/pranking items. But maybe there´s a more positive approach as well.

    Whether or not the game industry taps into these things, they will probably define the future of technotronic civilization.

  2. It is also valuable to consider how different monetary models react when deployed within a variety of atmospheres.

    My field of expertise, for example, is heavily in social web properties. We have found that money which a user spends, is in many cases perceived by fellow users as contributing to the community, or more importantly, to the common cause. Therefore it serves as yet another badge of social importance, and to be frank, social dominance over other users.

    In this case, a commitment to a subscription model seems to be the “cleanest” avenue, because it can be safely displayed and reported to others, while providing a list (in whatever form) of individual purchases, can generate substantial privacy liabilities.

    I would be extremely interested to hear any comments on backing a micro/item transaction type model into such an environment, where purchases can stand as social confirmations.

  3. Patrick, a possible issue could be running afoul of anti-gambling laws. It may not be so easy for western products to actually allow for payoffs like that. IANAL of course, but it strikes me as something that could crop up as an issue, didn’t SL have to shut down casinos in game for that reason? Maybe I’m remembering that incorrectly.

  4. […] micro­­t­ra­nsa­ct­io­­n g­a­me­s fo­llo­w a p­o­wer law. Th­at is­ th­e­r­e­ is­ a lar­ge­ […]

  5. I can attest to this same phenomenon in our company. We have used this “fremium” model (that’s a great term) for 13 years to great success.

    The top 10-20% contribute an overwhelming amount of the total revenue. I think they also get a lot more fun out of the game, but who can say for sure. The ones who pay less really do owe that top 10-20% a debt of gratitude. 🙂

  6. I think the man’s reason (“I knew I’d better give them my money now, or I’d go home and start thinking about what he said and never sign up”) was an inaccurate rationalization.

    There have been multiple studies showing that people are more likely to believe an argument after they’ve heard a counterargument. I initially read a study about this based on political viewpoints, but all I’m googling up now is something about toothbrushing. Still holds, though, even when it’s about flossing. 😛

  7. […] in the numbers because of varying definitions are in the orders of magnitude. As an example, in his ARPU vs. ARPPU blog post, Raph Koster estimates monthly ARPUs in freemium successful models as $0.50-$1.50, versus $30-$35 […]

  8. […] developers know to segment by age groups and target towards those with the highest demonstrated ARPPUs. The even smarter developers know that age ain’t nothing but a number – a single, […]

  9. […] developers know to segment by age groups and target towards those with the highest demonstrated ARPPUs. The even smarter developers know that age ain’t nothing but a number – a single, lonely metric […]

  10. […] Raph Koster, Metaplace, 3/16/08, Share this:TwitterFacebookLike this:LikeBe the first to like this post. Bookmark the […]

  11. […] Average Revenue Per Paying User (ARPPU): Calculated by dividing up the revenue amongst the users who pay anything at all for the app.. (ARPU vs. ARPPU) […]

  12. […] Average Revenue Per Paying User (ARPPU): Calculated by dividing up the revenue amongst the users who pay anything at all for the app.. (ARPU vs. ARPPU) […]

  13. […] 8. Average Revenue Per Paying User(ARPPU): 該当アプリに対して何らかの形で料金を支払っているユーザの数で、収益を割った値(ARPU vs. ARPPU。 […]

Sorry, the comment form is closed at this time.