A key component of game design, that can be traced all the way down to operant conditioning, is the concept of a rewards schedule. The idea that an individual is more likely to engage in an activity if he expects to be rewarded for it.
When a mouse presses a lever, if it is sometimes rewarded with a treat, the mouse is likely to continue pressing the lever, even if it doesn’t always release a treat. The more early successes the mouse experiences, the longer it is likely to continue pressing the lever even as the rewards become fewer and far between.
In the same way, a human trying a new experience needs frequent encouragement earlier in the process and less after the process becomes familiar. Unlike the mouse, humans often require only recognition of a job well done to continue – bribes of food are unnecessary.
In videogames, recognition comes in many forms including power-ups, free-lives, unlockables, in-game currency, badges and achievements. In a gamified experience, this can also take many forms, from real rewards and privileges to simple badges and status indicators. Regardless of context and content, all successful implementations will necessarily adhere to a reward schedule.
A Moment in Time
It’s probably obvious why rewards encourage, but why do they need to be scheduled?
There are two reasons for scheduling: aspiration and recency
Aspiration
The user needs something to aspire to:
- A goal
- Something obtainable
- A reason for participating just a little bit more
Recency
The user needs something to defray doubt:
- A recent history of success
- Reassurance that the next goal isn’t unobtainable
At any given moment with the activity, the user should not be too far from his next success and not too far from his last victory.
But How Much is Too Much?
The concept of too much is relative; early in the user experience, the timespans need to be shorter, but as the user becomes more experienced and more confident, time between rewards can (and should be) padded out. A new player may need reinforcement every few minutes, while a veteran player may go multiple sessions without recognition of a victory.
Often content limitations dictate the total amount of rewards that can be given out; eventually users will earn all the rewards and reach the end of the content. Ideally this moment should be postponed for as long as possible, but doing that means either adding more content or longer waits between rewards.
At Badgeville, we use a tool to model rewards distributions, ensuring an optimal rewards curve. Badges are plotted over the expected time to earn them. The resultant curve of points should end up resembling something like this:
The area under the curve also roughly corresponds to the distribution of rewards earned over player lifespan. By the time players are considered ‘experienced’ they have earned well over half of the available rewards (but those rewards remaining are significantly more difficult to obtain and carry higher status).
Maturing Community
As a dedicated community matures, new rewards can be added to extend the tail end of the rewards frequency indefinitely. This is important because you want to ensure that aspiration continues to exist for your most advanced users.
Super Users
There will always be the occasional super-user that consumes content astronomically faster than everyone else. When building a rewards schedule, be careful not to fall into the trap of trying to keep ahead of the lone one or two super-users. Instead, give them a handful of nearly impossible super-achievements to keep them busy and focus the rest of your efforts on optimizing around the bulk of your advanced users. If there is anyone you don’t have to worry about losing, it’s the super-users.
The Anti-schedule
Random Intervals
Any graduate of a Psychology 101 course will tell you that the most effective conditioning schedule is actually the random schedule. While this may be true, it is essential that you don’t run with this sharp little factoid and impale yourself on it.
The key is that the user must anticipate the reward. If the reward is completely random, the psychological effectiveness is lost. Think about it this way, when you pull a slot machine lever, you know the desired outcome of 3 bars pays out the jackpot, 3 cherries pays out x10, three plums x8, etc, and you also have some sense of the likelihood of each of these occurrences. The excitement comes from the anticipation of the random result. If there’s no anticipation, there’s no point. In a way, gamifying with random intervals is a chicken and egg problem: you need engagement to make random intervals effective, but you need Gamification for engagement.
The solution: look for naturally-occurring unexpected outcomes. Put yourself in the mindset of an obsessive gambler: what can you place a bet on? Where’s the natural, semi-predictable uncertainty in your system? Gamification can train your users to watch the system like hawks, waiting for those natural occurrences to claim their rewards. Leverage random triggers, not random reward definitions.