Why gamification doesn’t always lead to the desired result
High scores, badges, leaderboards: what motivates us in video games can also be used in other areas. Various studies show that the principle known as gamification can definitely work to get people to take certain actions. Software development is no exception.
A good example of this is the streak ad removed from GitHub in 2016. It tracked how many days at a time a developer contributed code to a project without a break. A second advert showed profile visitors the person’s longest series to date.
According to the authors of this papers did it actually work. At least in the sense that before the streak counter was removed, developers tended to write code every day if possible in order to be able to show a corresponding series in their profile.
The question at this point is whether the mechanism really motivated a type of behavior that could be described as meaningful. If you just want to prove that you contribute to an open source project every day, you probably won’t be trying to solve the most difficult problems. After all, a simple problem can be solved much more quickly. After all, it had no effect on the series.
Editor’s Recommendations
According to the paper mentioned above, after the feature was shut down, there was a very significant drop in the number of developers who contributed code to projects on a daily basis. Interestingly, however, the behavior of those who had streaks of 200 days or more did not change.
The GitHub example shows a very fundamental danger of gamification mechanisms: game elements function as a motivational aid, but can also quickly lead to false incentives.
Studies show that certain people, when confronted with game mechanics, optimize their behavior precisely to do as well as possible in the “game”. Of course, that’s exactly a problem when it comes to work and not a high score in a video game that is irrelevant in and of itself.
Anyone who uses gamification approaches should therefore always keep a very close eye on what behavior the users actually show in the end. After all, achieving a target metric while an equally important metric suffers cannot and should not be the goal.