Freitag, 30. August 2013

Gamification – What is the benefit for community management?

Gamification is a very young and hot debating topic. The term was not existent before 2010 [1], a fact that might explain that there are so few empirical studies on this subject. Furthermore, the available studies show mixed results [2]. Apparently, the effectiveness of gamification depends on a whole set of interrelated factors (community topic, community type, community size, cultural factors, members' motives and appraisals, the design of the reward system itself etc.). This multitude of determinants makes it difficult to decide whether gamifying one’s own community site may be such a good idea after all.

But, what is gamification?
Gamification can mean two completely different things. According to the definition given by Wikipedia: "Gamification is the use of game thinking and game mechanics in a non-game context to engage users and solve problems." [3]
In this context, gamification means the use of game concepts in order to solve problems of a non-gaming-domain with the help of engaged online-gamers. Very interesting examples can be found in the last four minutes (16:00) of Jane McGonigals speech [4]. Another good example is with its claim: Solve puzzles for science.
But the popular discourse omits the game concepts and the problem-solving aspect. Here, gamification simply refers to the application of game design features to non-game applications in order to make them more fun and engaging. [5] The most common features are points, badges and leaderboards. If you feel like having a déjà-vu now, then you are right. These are features of a non-financial incentive system which we find in customer loyalty programs or in the job context. And there are many studies dealing with the mechanisms and effects of such reward systems stressing either the benefits or the dangers. Only the fun factor is new.

A basic behavioural model

Social psychology and learning psychology offer a lot of models and theories that try to explain how motives, attitudes and intentions shape our actual behaviour. The common denominator of these theories is the basic behavioural function in which the (observable) behavior is determined by personal factors (e.g. traits, dispositions, abilities) and by environmental factors (e.g. physical and social resources).

When we add a behavioural norm such as „avoid pain – seek pleasure“ (avoidance – appetence) we get a very simple model with which we can try to predict behaviour.
Of course, the basic function is just a black box. We do not know which (unobservable) psychological mechanisms lead from (observable and unobservable) personal and environmental factors to observable behaviour. So, we need further assumptions. It is in these assumptions that those theories differ. And needless to say, the more complicated the organism is, the more complicated it will be to predict/trigger a certain behaviour because the internal mechanisms are more complicated, the range of possible actions is wider and more factors from the environment have to be considered.
This in mind, let us try to hypothesize some of the factors that might be relevant in an online community context.

Which factors might determine the effectiveness of gamification?

Community Topic
According to the Gamification Wiki, gamification can be applied to any industry and almost anything to create fun and engaging experiences, converting users into players. [5] But honestly, would you gamify an online community for cancer survivors and their relatives?

Community Type/Community Size
According to Ren, Kraut & Kiesler [6] online-communities can be divided into two groups in terms of social psychology:

  • In common-identity groups, members feel more attached to the group as a whole. Identity means that a member feels a commitment to the group’s purpose or topic. Seeking and providing information is a primary driver (e.g.: a movie-talk group).
  • In common-bond groups, a member feels socially and emotionally attached to particular members as well as to the group as a whole. Here, the focus lies on social interaction (e.g.: pupils who meet after class in a social network).

Please note that both types can coexist especially in large communities: The most active members of an interest-based community may form a bond-based subgroup for instance. And community type can change over time.

Dholakia, Bagozzi & Klein Pearo [7] offer a similar distinction. They posit that, in general, participants of

  • a network-based online-community, where members show identification with the online site itself and not so much with particular members, are more purpose-oriented, they seek information and expect the community to bring them together with others who will provide this information.
  • In small-group based online-communities, where members maintain a dense web of relationships with other members and identify with them in the first place, interpersonal connectivity and engagement in social interactions are the drivers of participation.

Here is an example of what can happen when gaming elements are introduced in a community where both groups coexist:

„In a community context, an e-learning community I worked for added a few gaming elements in 2011 such as individual points, organizational points and giving thanks to other members who gave them tips/shared a good course etc. We used the points system a couple of times to get members to complete certain actions in return for bonus points which worked well but it was those members who already enjoyed participating in the community who responded to this the most and most frequently used the gaming elements.

A small group of active members who regularly communicated in the community and outside the community via email and events utilized the points system most as they were already close, felt comfortable enough in the community to publicly compete against each other and loved gloating to each other when they were in the lead.

It was all in good humour and feedback from other members who perhaps didn't use the gaming elements in such a competitive way was that it was fun to observe.“ [8]
It seems that the underlying motivation for taking part in the community was not affected. Those with the common-bond attachment were happy to get just another chance to express their close ties.

A similar observation was made by Hamari [9]. The mere implementation of gamification mechanics didn’t automatically lead to significant increase in use activity of sharetribe members. But those who actively monitored their own badges and those of others showed increased user activity. (N = 3.234, experimental setting).

Cultural Background
In a context where open competition and „ gloating to each other when they were in the lead“ is deemed inappropriate gaming elements are detrimental.

Members' Motives and Appraisals
Cognitive models of human behaviour make several assumptions about the person's factors such as

  • motives e.g. affiliation, 
  • achievement, 
  • power (dominance, control)
  • self-determination: Do I pursue the goal for its own sake (intrinsic motivation) or do I pursue the goal because I will get a reward/I want to avoid a sanction (extrinsic motivation). Note: Motivation is not static it can shift from extrinsic to intrinsic (internalization) or from intrinsic to extrinsic (overjustification, see further below).
  • appraisals
    • locus of control: Will my behaviour be causal to reach my goal (expectation of attaining a goal)? How much do I value the goal (how dear is the goal to me)? Can I control the attainment of the goal (internal locus of control) or does it depend on luck or other factors I cannot control (external locus of control)?
    •  self-efficacy: Will I have the know-how necessary for the task? Will I have the resources
    • norms: What will the other members think of my behaviour?
That these factors are indeed relevant shows a study on lurking behaviour [10] from 2006. It offers interesting details why community members don't participate. In checkbox questions and open-ended questions lurkers were asked to tell more about their reasons (for details see The options were derived from original statements of a previous study. Most of them can be attributed more or less to the above mentioned factors, for instance:
  • lack of motivation: “didn't need to post”
  • lack of self-efficacy: “still learning about the group = I don’t know whether I already have enough knowledge to participate, couldn't make the software work”
  • low expectation for attaining one's goal by participation: “didn't like the group (poor dynamics, fit)”
Of course, in 2006 the concept of gamification was unknown. And so none of the interviewed lurkers could have possibly answered: No badges, no leaderboard. On the other, hand it is difficult to imagine how gamification could fit into this picture.

Social factors like norms or reciprocal benefits
Hamari et al. [11] show the importance of social factors. Social influence (the individual’s perception of how important others regard the target behavior), recognition (social feedback the individual receives from other users), reciprocal benefit and network are strong predictors for attitudes towards gamification, continued use intentions and intentions to recommend the related services (n=107). Note: Intentions are only a predictor for behaviour.

Design of gaming elements
Hakulinen et al. [12] found that achievement badges can affect student’s behaviour depending on the badge type. (Unfortunately, only the abstract of this study was available to me ).  Liu et al. [13], after having conducted two case studies, argue that the design of the main functionalities has a much greater impact on behavior than the additional gamified components. But the study had only a small number of participants and the experiment was of short duration.

What are the possible risks of gamification? 

Overjustification - Do extrinsic rewards undermine intrinsic motivations?
The overjustification effect occurs when an expected external incentive such as money or prizes decrease a person's intrinsic motivation to perform a task. According to self-perception theory, people pay more attention to the external reward for an activity than to the inherent enjoyment and satisfaction received from the activity itself. The overall effect of offering a reward for a previously unrewarded activity is a shift to extrinsic motivation and the undermining of pre-existing intrinsic motivation. Once rewards are no longer offered, interest in the activity is lost; prior intrinsic motivation does not return, and extrinsic rewards must be continuously offered as motivation to sustain the activity. The overjustification effect is controversial because it challenges previous findings in psychology on the general effectiveness of (behaviourist) reinforcement on increasing behavior, and also the widespread practice of using incentives. In fact, a 2001 meta-analysis showed that rewards can increase intrinsic motivation for tasks that initially hold little intrinsic interest. [14]

Mean maximization distorts the decision outcome
In their article from 2003, Hsee, Yu, Zhang & Zhang [15] demonstrate in several experiments that, when people are faced with options entailing different outcomes, the presence of a medium — for example, points or money — can alter what option they choose. The effect (= medium maximization) occurs because the medium presents several illusions. A medium is a token people receive as the immediate reward of their effort. It has no value in and of itself, but it can be traded for a desired outcome.

No guarantee of sustainability
Typical online games give the gamer a chance to develop her/his skills and to reach new levels, a simple reward system with badges doesn't. So introducing such a system may achieve only a quick spike.
[1] Xu, Y. (2011). Literature Review on Web Application Gamification and Analytics. CSDL Technical Report 11-05. Google Scholar
[2a] Witt, Maximilian; Scheiner, Christian; Robra-Bissantz, Susanne. Gamification of online idea competitions: Insights from an explorative case. Informatik schafft Communities, 2011, S. 192.Google Scholar: Participants (n=30!) were asked about their motives for participation, flow, enjoyment and task involvement. Explorative study (no experimental design, small sample size, mixed results).
[2b] Thom, J., Millen, D., & DiMicco, J. (2012, February). Removing gamification from an enterprise sns. In Proceedings of the ACM 2012 conference on Computer Supported Cooperative Work (pp. 1067-1070). ACM. Google Scholar: When the points deployment was introduced the contribution of employees’ content rose sharply. The experimental removal ten months after the introduction reduced participation. The study focused on 3.486 active members.
[3] Wikipedia. (8.8.13)
[5] Gamification Wiki. (8.8.13)
[6] Ren, Y., Kraut, R., Kiesler, S. (200?). Identity and bond theories to understand design decisions for online communities. Google Scholar
[7] Dholakia, U.M., Bagozzi, R.P., Klein Pearo, L. (2004). A social influence model of consumer participation in network- and small-group-based virtual communities. International Journal of Research in Marketing 21 (2004) 241 – 263. Google Scholar
[8] Community Geek. (8.8.13)
[9] Hamari, J. (2013). Transforming Homo Economicus into Homo Ludens: a field experiment on gamification in a utilitarian peer-to-peer trading service. Electronic Commerce Research and Applications. Google Scholar
[10] Nonnecke, B., Andrews, D., Preece, J. (2006). Non-public and public online community participation: Needs, attitudes and behaviour. Electron Commerce Res 6, p. 7-20. Google Scholar
[11] Hamari, J., & Koivisto, J. (2013, June). Social motivations to use gamification: An empirical study of gamifying exercise. In proceedings of the 21 st European conference in information systems. Utrecht, Netherlands. Google Scholar
[12] Hakulinen, L., Auvinen, T., & Korhonen, A. Empirical Study on the Effect of Achievement Badges in TRAKLA2 Online Learning Environment.
[14] Liu, Y., Alexandrova, T., & Nakajima, T. (2011, December). Gamifying intelligent environments. In Proceedings of the 2011 international ACM workshop on Ubiquitous meta user interfaces (pp. 7-12). ACM. Google Scholar
[14] (8.8.13)
[15] Christoper K. Hsee, Fang Yu, Jiao Zhang, Yan Zhang (2003). Medium Maximization. Journal of Consumer Research,  Vol. 30, June 2003

Keine Kommentare:

Kommentar veröffentlichen