5. Assessment: Evaluating with(out) Points
5.1 Different Clocks, Same Time
5.1.1. Points for Consistency or Consistent Points?
We chose the use of points for our first proposal because points have the advantage/disadvantage of equating their award with specific behaviors; we can assess behavior, or even performance of a specific behavior, and award/deduct respective points.
Thus in this case we selected three specific behaviors from the employees:
Chapter:5. Assessment: Evaluating with(out) Points
1 1 7
1. Come to work on-time; logon before a specific time.
2. Be consistent; have the least late days as possible.
3. Show devotion and enthusiasm; try show up earlier at work than others.
Hence the formula calculating the rewarded points after logging-on to the workstation reflects the value of such behaviors.
Of course, one can clearly see the advantages of finity and fairness56 such a solution affords;
every user knows what to expect, in that the rules of the ―game‖ are clear, fixed and unquestioned. In that sense, the calculating formula takes up the role of the regulating authority, administering the process of checking in to work.
However, this type of points-for-behavior ―gameplay‖ will unavoidably lead to strategic planning and micro-management of the behavior; instead of motivating behavior, the application will motivate the collection of points. Thus the first problem arises when users will pursue the collection of points by all means; this translates into a manipulative attempt in order to twist the system against its static authority (the formula in this case)57.
Moreover, there is also an issue of options; in this solution, the user is only left to compete against other users on the leader board. Eventually this relentless strategy will reach a critical point where a leading small group58 will keep pursuing the top positions of the leader board, while the critical mass of users will just ignore and disregard the overall application59. The above problem was indeed an existing one for a gamified application; Foursquare, during its last production iteration (version 2.0) faced harsh criticism when many users started (ab)using
56 You know what points you may score, even before you logon.
57 A similar argument is made by Kohn connects the inefficiency of rewards with the governing nature of the rewarding authority and the loss of the user autonomy:
―rewards are usually experienced as controlling, and we tend to recoil from situations where our autonomy has been diminished.‖ (Kohn, 2001, p. 78)
58 In which the user types could either be competition driven individuals (see Bartle‘s achievers Bartle, 2003), or dependants; individuals who are tied to the reward of points (which clearly is not the case for most applications of gamification)
59 One could even claim that such an approach is utterly useless, since the initial motivation for designing such a solution is to motivate employees that are not a competitive by nature.
Chapter:5. Assessment: Evaluating with(out) Points
1 1 8
third party applications like ‗MayorMaker‟60 to gain advantage over the way Foursquare was developed and earn points easier. The result was the creation of castes between the top leaders of each leader board and the followers.
Finally, one even more significant problem with this proposal is that points must reflect a value of something noteworthy, either in direct monetary sense or in social value. With a significant number of users gradually dismissing these points over time, we see that any kind of social value is doomed to be reduced to zero.
We thus conclude that a points approach has limited effectiveness and life-span; it fails to address our problem in a way that is relevant for all users, it becomes less effective over time and is essentially rendered practically useless in depth of time. In that regard, the only option is to ―artificially‖ extend its complexity in order to mask the direct relationship between the points awarded and the behaviors exerted, by (over)complicating the points calculation formula such as the one used in this solution. Nonetheless, we regard it as a poor option, failing to alter significantly the user(worker)-artifact(logon system) relationship.
5.1.2 The Scout boy Worker
In the previous part we saw that our attempt to approach our design problem with a points-based solution proved to generate more problems that it solved. Thus, by deploying a badges-based proposal instead, we tried to overcome some of these problems, while maintaining the spirit of rewarding the user for specific behavior.
As such, our strategy was to expand both horizontally and vertically all desirable user behaviors in order to create a matrix of conditions. Thus, the initial three behaviors (on-time, consistency, devotion) had to be translated into more explicit classes which then could be combined to create a sufficient range of variable (micro) behaviors. Then, we scaled up these
60 See http://mayormaker.com/.
Chapter:5. Assessment: Evaluating with(out) Points
1 1 9
(micro) behaviors in order to add depth and duration to the overall process; Table 2 elaborates the process we followed. badge/achievement award and consequently afford for different types of users. Competition driven users can be addressed by setting competitive achievements, like being in the top-five or max-week-performance targets; for more adventurous users, we can ―hide‖ achievement conditions in complex behaviors like an employee that logs on first and logs off last while for enhancing social interactions we can include conditions that require some basic level of social exchange among workers like a group achievement for the best department61.
Hence, it becomes rather obvious that the whole user experience is more personal (single player) than our points approach; instead of competing against one another with the system regulating the winners/losers, the users are challenged and rewarded on a personal level with a system that is pushing them to be more serendipitous. Consequently, in order to induce social value to the rewards and enable some basic social exchange, we added the option of exhibiting/portraying the badges over social networks.
61 Again here, we refer to the four basic Bartle player types (see Bartle, 2003), but it obviously we could expand them to the extended 8-types model or a custom framework based on play-personas (see Canossa & Drachen, 2009)
Chapter:5. Assessment: Evaluating with(out) Points
1 2 0
However, even though our badges approach is less problematic compared to the previous point-based one, it does have the disadvantage of having a finite point of saturation. By this, we mean a state where most of the achievements are awarded to all (interested) users – thus rendering them inoperative. The solutions to this inherent problem are both costly and short-termed; constant support and development of (micro) behaviors and achievements is not always feasible or cost-efficient. For example in our problem, the desirable behaviors are so limited that after some point it is no longer possible to extend the behaviors-rewards matrix by combining (micro) behaviors; in this case scaling up is the only solution. However scaling is not always the best option, since linear upscaling very often turns the whole process into a boring grind.
Hence, we conclude that using badges is overall a better approach; it eliminates significant problems of our points approach and it maintains the main principle of conditioning (rewarding specific behavior). However, the cost and viability of such a solution are neither reduced nor expanded respectively. As such, we do not view a badge approach as a panacea for all design situations, but we reserve our opinion to claim that it can be both operative and efficient on some others.