One of the biggest challenges for a social or mobile game company is managing the challenges of an industry that is hit-driven. Understanding the laws of probability is a good step in overcoming this hurdle. Since the beginning of time, game companies have been struggling with the hit-driven nature of our industry. Although new markets and technologies (e.g., casual, social, and mobile games) held out the promise of not being as hit-driven, they all ended up having a few titles drive the majority of revenue. The number I use is that 25 percent of games will break even or make a profit after launch (without accounting for development expense); this means that there’s a 75 percent failure rate. (I have heard estimates that as many as 90 percent of game projects fail and I would not argue strongly against that number, but for the sake of analysis I will use a success rate of 25 percent.)
The argument for taking a portfolio approach
Basic probability theory shows the importance of having multiple independent game projects to improve your company’s chance of success (more on independence later). By using probability, the benefits of taking a portfolio approach become clear. The probability of either one event (one game being a hit) or another occurring is the sum of the probability of the first event occurring plus the probability of the second occurring minus the probability of both events occurring together. Assuming there is a 25 percent chance of a game being a success, if you develop two games the chance of creating a success rises to 43.75 (.25 + .25 – (.25 * .25). If you develop three games, the chance of success increases to 68.75 percent (.25 + .25 +.25 – (.015625*4 (or the four possibilities where there are two or more successes)).
The importance of independence
One crucial element to the example above is that each game development project is independent, meaning that the success or failure of one project does not change the odds for the other games’ succeess. If the games are not independent, then the odds of success can stay as little as .25 even with creating three projects.
Some ways the projects may not be independent:
- Same dev team or key personnel. If all the projects are created by the same team and the team is not very good and that is the reason the first game fails, the others will probably also fail.
- Same core game play mechanic. If the underlying gameplay is the same, and that does not resonate with the market on game one, it is not likely to resonate on future games. For example, if you create a hidden object game and the audience on your platform does not like hidden object games, the second and third products have a much greater than 25 percent chance of failing.
- The licensed IP does not work. If the first game is based on a brand or property and it fails, then there is a strong chance that subsequent games using the same IP will also fail even if they are very different games. For example, if a Star Trek Sudoku game fails, the odds on your Star Trek city builder will be less than 25 percent (but not zero since they are different genres.
- The themes are too similar. If all of the games have the same theme (say politics or football) and the first one fails, the chance that the market is not interested in that theme reduces the probability that subsequent titls will be successful.
Don’t fight the odds
Almost everyone believes they are in the 25 percent that will succeed, but even the best companies have to deal with this baseline reality. Even if you are that much better, maybe you have increased the odds to 30 or 35 percent. By looking at probabilities, you can triple your chance of success by developing multiple projects. I actually do not understand how anyone would bet their company when you have a 25 percent chance of success (you might as well take your investors’ money to Vegas, where the odds are better).
The same holds for the importance of making sure your projects are not too closely related. You increase your probability of success by having the projects as independent as possible. This rule also holds for publishers: The benefit of publishing multiple titles are negated when the titles are all virtually the same (such as the rush to mid-core titles). If you really want to increase your probability for success, the math tells you to have as many independent good projects as possible.
12 thoughts on “Using probability to build a game portfolio likely to succeed”
I think your message is valid, but even accepting the 25% figure, you’re being optimistic Lloyd! For 3 games, each with an independent 25% chance of success, I’d put the odds of at least one success at 1 – the chance of all failing (.75 x .75 x .75) = only about 57.8%.
You are right on both points, David. Math mistake on my side, I should of done the math your way (I neglected a possibility and double counted). I may have been optimistic but it really depends how you define the group. If you take it very broadly to include poorly capitalized devs working out of a basement, the odds are much lower. I was thinking if you do everything right, you are closer to that 25% number. Regardless, having several independent projects has a huge impact on your chance of success (or survival).
Great article Lloyd. I would add something that we have found is also an argument for the portfolio approach and it is that it forces you to think in terms of toolchains and the capability to have strong live ops capability.
When you only have one type of game, or even worse, one game, you might not feel the need to invest in developing these set of tools. But once you are operating 4-5 different games live, you will need to have the ability to make changes, and insert content into the game more effectively.
Completely agree, that’s how to build a company, not a project.
Interesting article. Though it’s missing the one almost guaranteed advantage to taking a portfolio approach: cross-promotion and creating your own internal network between games. Having a network of games on iOS right now decreases CPI substantially through optimizing cross-promotion between multiple titles.
To me, that’s more important than the ‘odds of having a success’. Great post though =) I always love your content here.
In my experience (on iOS) a diversified portfolio approach actually works against cross-promotion, since each title is targeting a different audience, minimizing the chance of cross-over appeal. In fact when targeting different ages, it can even be inappropriate. I still like Lloyd’s argument because you’re nowhere without those initial successes – but it’s not the road to maximizing cross-promotion.
Unless you use cross-promotion with a service like Applifier so you can at least generate “free” users who are relevant.
I agree. It is actually missing a lot but I wanted to focus on the math side of things. Cross-promotion is huge, as is having a community.
this is an awesome topic you have raised. So how will you define “success” or “being a hit” in your example? Does it mean that one “successful” game will cover the costs related to development and marketing of all three + some additional margin? I don`t know the rate for casual titles but i remember that for traditional console games the number everyone was using a couple of years ago while defining “successful” was x6-x8 times compared to budget. If you mean something similar to that (at least) then consider this – to be THAT successful all three games in your example need to be either of significantly higher production quality compared to competition (costs go up, more difficult to break even) or somehow riskier in terms of design/mechanics in order to be distinguished in the crowd (which correspondingly reduces that .25 chance for success for each particular title). In the end, won`t you end up in a situation where the more you do to increase your chances the lower they get? And is there any balance here?
This model is not meant to be that precise. As David Hamm pointed out, the 25 percent number to begin with is probably too optimistic. What I had in mind was actually at a minimum a game that supported itself, the cost of running the game (and acquiring users) was less than the revenue it was generating. In my mind, it did not necessarily even recoup development expense. The goal of my post was to show the benefits of multiple projects over one game and the math behind that. The actual P&L has many other variables and probably has to do more with the discussion on LTV.
Yeah, thanks for clarification, in this case the model completely makes sense. I was initially thinking if one can use this model to potentially bring risks down to zero. While using David`s math, you will have to release approximately 15 titles to take your chance of success with at least one of those close to 99%. In real world it sounds like a very true statement for a game that just needs to cover its ongoing costs (as you explained) but not necessarily true for a game that would need to have 15x ROI (as I initially thought). That`s why I was asking about your definition of “success”.