This article displays a design ethnographic case study on an ongoing machine learning project at a Scandinavian gamification start-up company. From late 2020 until early 2021, the project produced a machine learning proof of concept, later implemented in the gamification start-up´s application programming interface to offer smart gamification. The initial results show promise in using prediction models to automate the cluster model selection affording more functional, autonomous, and scalable user segments that are faster to implement. The finding provides opportunities for gamification (e.g., in learning analytics and health informatics). An identified challenge was performance; the neural networks required hyperparameter fine-tuning, which is time-consuming and limits scalability. Interesting further investigations should consider the neural network fine-tuning process, but also attempt to verify the effectiveness of the cluster models selection compared with a control group.
TopIntroduction
This chapter aims to highlight opportunities, challenges and future research on an ongoing development project - the AUtonomous TailOred gaMificATiON (AUTOMATON) - at a gamification start-up for gamified machine learning acting in Scandinavia. Gamification refers to using game elements in non-game situations (Deterding et al., 2011) and has been discussed in various domains, such as crowdsourcing, tourism, computer science, sustainability, software development, health and wellness as well as in business (Helmefalk, 2019). A considerable amount of literature emphasizes these elements being defined as game mechanics (e.g., badge, level, leaderboard, points), supporting various processes while being engaging and motivating (Looyestyn et al., 2017; Reiners & Wood, 2015; Sailer et al., 2017; Wee & Choong, 2019).
During its relatively brief existence, the multifaceted concept of gamification has gained much attention in the last decade in both business and academia (Nacke & Deterding, 2017). Regarding search trends (Figure.1), it has also surpassed its older sibling Serious games, a game designed for purposes other than entertainment such as learning. Ironically, the last decade of gamification hype might be to gamification's detriment. This is because even though several established advisory firms and research institutes (see Burke, 2012; IEEE, 2014) have predicted a promising future for the use of game elements in the non-game context, several of these predictions have failed to actualize.
Figure 1.
“Gamification” surpassed “Serious Games” in search trends in 2012 (Google Trends, 2021)
TopBackground And Problem
Gamification researchers have suggested that gamification, probably due to its rapid rise to fame, has largely failed to evolve with the fields that elevated it (Raftopoulos, 2020). Also, even though research on gamification shows promise in several fields (Koivisto & Hamari, 2019), the gamification industry has not yet refined any rigorous and verified standards (Nacke & Deterding, 2017), resulting in an uncertain market in the hands of various gamification consultants – with slim possibilities of developing and validating industry standards due to the many different design approaches (Koivisto & Hamari, 2019). To illustrate, while one industry may apply best practices, standards or solutions that generate promising results, other industries have their own. Consequently, inputs, outputs and data become too eclectic to (dis)confirm the wider efficiency of gamification. Last but not least, gamification scholars and practitioners alike have argued that gamification designs are context-dependent (Palmquist et al., 2021). Transferring a successful design from one context to another can have serious consequences in the gamified situation, affecting the users negatively. A gamified solution that have showed promise in increasing engagement among students in education, may not only be different from other contexts such as health and fitness services, but may also differ from other courses. Needless to say, context is important. This predicament has made gamification demanding in terms of resources such as capital and time; likewise, it makes gamification hard to generalize and scale up effectively. Solving these problems is not a simple matter, but requires long-term methods, designs, systems and technical solutions that aid our understanding how people think, feel and use gamified services.