This paper presents a model-based approach to guide the design of gamified applications. The models enable a structured representation and analysis of the social and intentional dimensions of the actors involved in the organizational context of the gamified system. The proposed notation (called GStar) makes use of the agent- and goaloriented features of the i* modeling framework. The main contribution of this work is to bring the social and intentional aspects, which are usually treated in an ad hoc manner in the design of gamified systems, to the core of a modeling approach. This approach supports the analysis of the consequences of the inclusion of each gamification element and facilitates the evaluation by the designer of different alternatives, contributing to a project more suitable to the organizational context and more capable of fulfilling the established goals.
There are many definitions for Gamification in literature. The most widely used
presents Gamification as the "use of game design elements in non-game contexts"
[
The fact that it deals with human motivation makes the project of gamified
applications a complex task, since it implies the need to understand the user’s
psychological perspective. The Self-Determination Theory (SDT) indicates, for
example, that details such as the meaning of a reward - whether it is perceived
or not as a symbol of competence - have a significant role in the reward’s ability
to motivate a person [
In this context, this work aims at proposing a modeling method to help
analysts and designers to deal with the complexity, especially from a psychological
and social point of view, of designing a gamification application. This method
is based on the use of GStar, an extension proposed by this work to the
agentoriented and goal-oriented framework i* [
The research methodology used to guide the development of the proposed extension was the Design Science Research. Successive suggestion / development / evaluation iterations were carried out, first to explore the possibilities of modeling gamified applications with an agent/goal oriented modeling approach (framework i*), and then to extend this approach to better suit the context of gamification. 2
The i* modeling language is flexible and satisfactorily represents the
dependencies, intentions, and potential conflicts between actors in an organization. It has
been extended to include concepts from different domains, including security,
software development, and autonomic computing systems [
Gamification can be understood as a way of influencing the user’s motivation
to manifest certain desirable behaviors through the satisfaction of psychological
drivers. These drivers are associated with the satisfaction of basic psychological
needs that are the precursors of human motivation [
The basic principle for the representation of the new concepts is the adherence to the original view of the i* modeling language. The i* framework is goaloriented. The three new concepts are represented by intentional elements already included in i*, being differentiated from the original ones by symbols attached to the visual elements. This allows the use of analysis techniques developed in the context of i* without the need for adaptations and facilitates the use of modeling tools originally developed for i *. Figure 1 shows the proposed representation for the new concepts: – Target behaviors: behaviors that the designer wants to incentivize through the implementation of the gamified experience. – Gamification elements: game design mechanisms selected by the designer to create the gamified experience. They are represented through a game related goal, from player’s point of view. For example, points can be represented by “ Earn points ”, and leaderboards by “ Improve on the leaderboard”. – Psychological drivers: The main drivers related to gamification are competition, status, achievement, self-expression, and altruism. 3
For illustration, we consider the design of the Beehive system, a private social
network implemented inside IBM reported in [
Users’ engagement in the experimental group increased after the deployment
of the gamification experience, but the result was not sustainable. In the second
week after the implementation of gamification, the activity reduced significantly,
and three weeks later the engagement measured by the average amount of shared
content per user was already at the same levels as the control group. The authors
of the study believe that the reasons behind the decrease on engagement in the
experimental group after the first week of implantation of gamification were
related to the fact that some users, after achieving an intermediate level, were
satisfied and stopped contributing [
The points system analysis was guided by two analysis questions: (1) What are the implications of the point system?, and (2) Assuming that users are engaged in the competition, and behave in a way that maximizes their chances to "win" the game. How would those users behave considering the designed point system? If users are engaged in the competition, they will behave in a way that raises their chances to win, i.e., they will seek to increase their score, and thus improving their positions in the leaderboards. As shown in Figure 2, there are three ways to earn points in Beehive: commenting (15 points), posting a list (10 points), or sharing a photo (5 points). In a competitive environment, engaged users will naturally prioritize activities that require less effort and reward with more points. As shown in Figure 2, in Beehive, the activity that demands the least effort is also the one that rewards with the highest number of points, which is Comment (this is highlighted through the dashed circles). Another important aspect easily seen in the model of Figure 2 is that, since comments are posted on photos and lists, they depend on them being previously created to exist.
Gamified experiences that provide the driver competition can provide a sense
of competence to its users, which is one of the basic needs postulated by the
Self-Determination Theory (SDT) [
The hypothesis that the point system may result in a disproportional
increase of comments is corroborated by data that shows the evolution of the
average amount of content created per user on the experimental group [
According to the SDT, performance-contingent rewards, as long as they are
perceived as a sign of competence, and therefore having an informational character,
positively affect people’s intrinsic motivation [
To understand the meaning of this visibility provided by the leaderboards, one needs to understand what it takes to reach the top in terms of performance. And from that, try to understand how people perceive the best performers. To guide the analysis, the following analysis question was outlined: What are the implications of the leaderboards? To improve their positions in the leaderboards, users need to earn game points. As shown in Figure 3, this is done when the user actively spends time on the social network, commenting, sharing photos or lists. Due to the triviality of these activities, the best-placed users will be users who simply spend more time on the social network.
The fact that to climb to the top of the leaderboards is only necessary to spend more time on the social network, together with the fact that it is possible to "cheat" by overdoing frivolous comments to move faster in the game severely affects the leaderboards experience. In addition, as shown in Figure 3, users who spend more time on the network, and therefore can secure the best places in the leaderboards, will have less time to perform their main work activities, which conflicts with the goal of advancing their careers. This may also be perceived by others as unprofessional behavior, affecting both the status driver and the soft-goal of advancing the career.
The analysis above suggests that the designed incentives in the Beehive gamification experience fail to provide the satisfaction of the basic needs which are precursors to human motivation. The SDT states that rewards, such as awarding points and visibility to the best performers, without providing a sense of competence both for the individual (self-esteem) and for their peers (reputation), are not effective in increasing people’s motivation. Perhaps the only positive aspect, as shown in Figure 3, is that when the user is visible among the most participatory users of the social network, the employee is showing himself engaged in a project sponsored by the company’s management. 4
The presented approach promotes a structured evaluation of the insertion of each
gamification element, supporting the analysis of the implications of the proposed
incentives concerning the user’s intentional perspective. The Beehive social
network analysis example shows how the model-based approach proposed by this
work supports the identification of challenges and conflicts of interest that
otherwise might not be noticed by designers. The identification and consideration
of these inconsistencies already during the design phase, besides mitigating the
project risks, may lead to significant savings to the organization and the
development of a tool more adherent to the organizational context. Due to space
limitations, this paper presents only an outline of the proposed approach, a more
comprehensive presentation can be found in [
Case studies have been carried out to illustrate and validate the usefulness of the proposed method. Still, a more formal and comprehensive evaluation would be an important contribution to investigate further the practical utility of the presented approach. A possible way to approach this is through interviews with gamification specialists. A structured evaluation by experts with experience and practical knowledge on gamification projects would further support the practical utility of the proposed model-based approach.