Before Adding Rewards, We Asked Which Behavior We Wanted to Increase
A look back at the education platform's points and bubbleBible's reward design: separating behaviors that can be recorded from the values the product aims for.
Joohwan Yu · Founder, Full Stack Developer
Toris · Published
Contents
At work I handled point processing tied to content and activity, and in bubbleBible I worked on a reward structure that helps reading and sharing. We used the same name, points, but the behavior each product wanted was different. I felt that carrying an existing feature over unchanged is not good design.
Behavior the system counts versus value that stays with people
Clicks and the number of posts are easy to count. Whether someone understood what they read, or whether there was meaningful sharing, is not that simple. If you attach rewards to behavior that is easy to measure, the product can drift from the behavior it originally wanted to help. This is why you can’t say the reading experience improved just because the click count went up.
In bubbleBible’s early retrospective, we left a direction of restraint so that rewards would not replace the meditation itself. Looking back, that was a more important judgment than deciding the amount of points, because we first had to think about where a feature directs the user’s attention.
Payout rules were also operating rules
Once you add rewards, you also have to decide how to handle duplicate actions, cancellations, deletions, and switching to private. My experience handling point payouts and deductions at work connected to this question. If you only implement the case where someone participates normally once, it becomes hard to explain later-edited content and canceled actions.
If I were designing it again now, I would write the rewarded behavior in one sentence, along with the number of times it can repeat and the cancellation conditions. The goal is not to complicate the feature by imagining abuse cases. It is to check that the rules promised to users remain consistent after behavior changes.
The way of checking the effect has to be different too
Even if the number of events rises after turning rewards on, you have to look separately at whether that is a good outcome. You have to distinguish whether the same person entered repeatedly, or whether the number of people who actually keep reading increased. I won’t claim in this retrospective that such an effect was measured. Implementation and effect verification are left as separate stages.
This experience made me ask about the product’s purpose again before casually attaching game elements. If I introduce points next time, I’ll answer “if this behavior increases, will users get closer to what they really want?” before “how much should we give?” The calculation logic had to come after a clear purpose.
Related posts: bubbleBible’s reward design, Point state transitions
Advertisement