Early last week a paper appeared in the Proceedings of the National Academy of Sciences arguing that the huge discounts we usually see attached to large quantities of virtual goods in so-called "free-to-play" games have almost no impact on profitability, nor do they have any impact on the number of customers buying those goods. Fascinating stuff, at least if you're a game developer learning how to best implement the model.
Yet in some ways the information isn't quite as interesting as how it came about. The data from the study came from King Digital Entertainment, the studio behind the mobile hit Candy Crush Saga, who then shared their data and worked with University of Chicago economists Steven Levitt and John List to draw conclusions. Susanne Neckermann of Erasmus University was listed as a coauthor, but so was David Nelson of King itself.
Even though data like this affect many of our lives, companies the size of King tend to keep them locked behind their accountants' doors. Such collaborations are thus unfortunately rare. The very fact that it comes from King is something of a surprise, as in 2014 the company seemed to go out of its way to spread ill will when it sent a cease-and-desist letter to the makers of The Banner Saga, a game inspired by Scandinavian myth, for its use of the word "saga," which King had trademarked. (They eventually settled, and there's now even a Banner Saga sequel.)
Now, though, not only is it working openly with university researchers but it's sharing the fruits of its research with the wider community.
"I think there is a lot to gain from partnerships between cutting edge firms and academic researchers," Levitt said in a release. "Many firms are now doing randomized experiments, but mostly testing incremental changes. This project was an example of using experimental methods to test a much more radical shift in strategy."
As for the research itself, if you're not familiar with how these discounts work, consider this. You can usually download and play much of games like Candy Crush Saga for free, but King (and many, many other companies) make money by selling items that can help you advance more quickly or efficiently. These items can come in small or large bundles.
Typically, King doesn't discount larger bundles that much. But over the course of three months they and the University of Chicago researchers experimented with offering around 14 million players with as much as 60 percent discounts for mid-range purchases and 70 percent discounts for large ones.
Those discounts, they found, had almost no effect on how many people actually bought the items. The people who bought the heavily discounted items were typically players who usually only bought small quantities, and players who normally spent a lot of cash tended to buy fewer items during the discount period.
"The net result was no impact on revenues or profit," the release said. In fact, the data seem to show that customers who normally might have bought small purchases were actually discouraged from doing so while the large discounts for the bigger bundles were still available.