As I mentioned in an earlier post, I just started the Networked Life course taught by Michael Kearns from the University of Pennsylvania on Coursera. Part of the coursework has been about “heavy tail distribution,” a phenomenon typical in a large-scale network (like Facebook). To summarize, a heavy tail distribution (see image below for an example) means there is a clustering of vertices (users or in the case of Facebook, friends) with a very low number of connections but a long tail of vertices (users) who have a lot of contacts. Thus, in the case of Facebook (or any other typical network) the large majority of users have only a few Friends but there is a long tail of people who have a lot of friends (say over 1,000). These latter people are called “Connectors” in network theory. What is particularly interesting is that this structure, which again is typical of large-scale networks, is effectively the inverse of a traditional bell shape distribution, which would show a few users at the low end, a peak (the top of the bell) and then quickly trail off. The heavy tail is typical of all social networks, Facebook, Twitter, LinkedIn, etc.
What I find interesting is that you find a similar curve in virtually all free-to-play social games when it comes to users’ monetization tendencies. Depending on the game, you normally 90-97 percent of the players do not monetize. You then have a heavy tail where some players monetize at a low level but others (the high value players often referred to as “whales”) will spend thousands of dollars in a game. Thus, these high value players are serving the same role in making a free-to-play game viable as Connectors do in making a social network work.
Why the heavy tail is important
This finding is important for several reasons. First, many critics of the social game industry argue it is not viable because the games rely on only a small number of players to drive the game’s revenue. By understanding heavy tail distribution theory, though, it becomes clear that this clustering is a natural phenomenon and does not show weakness in a game (or business model). Saying Chefville is flawed because it relies on four percent of its players to drive profitability (no inside information here, just an educated guess) is like saying Facebook or LinkedIn are fundamentally flawed because they need their few users who have 1,000+ friends/contacts for the network to work. Second, for many social game designers and companies, their holy grail is to increase the number of people who monetize, even if they get them only to monetize in small amounts. Just as you cannot deny the laws of physics, you cannot deny the heavy tail. Efforts to increase the number of percent of players that are monetizing is likely an uphill battle if not impossible and you are better off using your resources to focus on the high-value players.
9 thoughts on “The heavy tail of monetization”
Excellent post! I like to think of the monetization curve exhibiting a pareto distribution over a continuous range of revenue values (with the origin being $0). Not only is this a little easier to use in forecasting revenue, but it also helps break people from the “payer / non-payer”, binary line of thought.
Interesting ideas on heavy tails but I think you made a few unsupported logical leaps. Gaming is dependent on the whales because they provide the revenue. Connectors don’t make direct payments like whales, at least not in their role as connecctors. Maybe you could show that the connectors are important to ad revenue but of course that still is not that same contribution as the whales. Your article doesn’t state how FB would be significantly weakened by loss of the connectors. …Doesn’t FB also profit from gaming, even if thru 3rd party vendors?
What I was saying is that the underlying dynamic that shapes a social network is the same as one that shapes monetization in a social game. Virality is one function, monetization another. I did not mean to imply that Connectors on Facebook directly pay revenue to Facebook (though they do generate higher ad revenue). To me, the fact that this phenomenon is typical in all large networks suggests that they are imperative to the network’s survival or growth to becoming large scale. If Facebook would not be weakened, we would see more large scale networks (or at least some) that had a bell shaped distribution rather than a heavy tail.
I also did not mean to discuss gaming versus Facebook. Gaming generates anywhere from 15-25% of Facebook’s revenue and creates a very healthy ecosystem.
WHat my post was intended to do was show that the heavy tail of monetization in games is not a problem or an indicator of problems with the underlying model.
Having run F2P and subscription games in the past for varying audiences, I don’t argue with the “long tail” model. Greg Satell has a great article about the same underlying mathematical issue (http://www.digitaltonto.com/2010/justin-bieber-social-networks-and-how-numbers-can-lie/). However, I do argue that it points to whale-based gameplay. When searching for the money (the area under the curve, in the above diagram), decisions about gameplay can fundamentally affect the curve’s shape in dramatic ways.
I believe that game design drives the “steepness” of the distribution curves, and the degree of difference between the mode and the mean. Gameplay that punishes players for not monetizing produces a game where most of the money comes from the “whales.” The result is a game where the curve is much like Lloyd’s above (i.e., very few low spenders). I have seen many of these games: in many strategy games, after X days you’ll get clobbered by PvPers who have spent “diamonds” (or similar RMT currency) to acquire vastly superior forces, unless you also buy such forces, or buy a “peace treaty” to keep you little kingdom safe. Many ville-like non-combative strategy games don’t allow your city, farm, or whatever to progress without purchasing something using the “diamond” currency. In FPS-style shooters, you find yourself blown away by players with superior weaponry purchased using “diamond” currency. It is no surprise that games of these types earn most of their money from “whales.”
However, I have seen very successful games earn most of their money non-whales, simply by revising the design to constantly encourage but never require monetization. These games never punished players for not monetizing. We discovered that the right side of the curve did not dive down to the low values anywhere near as fast. Instead it slowly declined, yielding a satisfying large “area under the curve” to the right of the whales.
A key correlation was gameplay time. Many people who kept playing for a number of months.did eventually monetize. If we had aggressively pursued monetization and whales, a “diamond wall” that virtually forced monetization would have chased them out of the game before they began monetizing. This would have deprived us of what ultimately became a major source of revenue.
– Arnold Hendrick
Very well written, clearly understandable except one thing: what about the graphic, why didn’t you add text that indicates which axis represents what? Like this, you can only understand the graph when you read the text as well, which is not so nice.
Thanks for the comment. Unfortunately, I did not feel that the graphic on its own could convey the point I was making in my post, I needed to put more context around it than just labeling the axis.
Yeah, of course the graph can’t stand alone, of course you need text. But without axis labeling, one simple has NO clue what is depicted in it. Could be the evolution of apples on trees over time as well as the amount of jokes about Apple copyright infringment depending on one’s monthly salary 😀 .