Gaming at work to save energy – Learnings from workers playing a cooperative game Yi Lou, Anders Lundström, Cristian Bogdan KTH Royal Institute of Technology Stockholm, SWEDEN [yiluo, andelund, cristi]@kth.se ABSTRACT consumption of buildings [4, 5]. One cause for this gap is Workplaces constitute a substantial part of the world’s that energy estimations fail to account for occupant behavior, energy consumption. In this paper we study how a potentially highlighting the potential for occupants to lower cooperative game, EnerSpace, could be used to increase their consumption. Therefore, increasing occupant energy awareness and support energy conservation. In the awareness of energy consumption and persuasive design study, after a one-week baseline period, 8 workers played becomes an interesting option to promote behavioral change. EnerSpace for one week. They competed in two teams that each powered a spaceship with their energy saving on a To increase energy awareness and promote behavioral resource exploitation trip to Mars. Each worker also had an change through persuasion, researchers have introduced avatar in the spaceship that reported personal consumption gamification and social networks into eco-visualizations in data using face recognition. The participant efforts and order to engage users through competition and interpersonal abilities to conserve energy affected the speed of the relations [6-8]. Studies have also focused on providing spaceship and the living condition of their avatars. The energy education [9] and gamified energy feedback [10, 11]. results show that workers decreased their overall energy Overall, these studies have focused on using either individual consumption, especially on the weekend. Based on our and cooperative approaches to energy conservation (e.g. results we discuss 1) designing for energy conservation when [12]). Researchers have also designed gamified missions, absent from the workplace, 2) how using the coffee room competitions and cooperation to motivate energy created energy conservation narratives but hampered local conservation [13, 14]. For instance, Power Agent was a experimentation and learning, 3) team formation to support persuasive game designed to encourage energy conservation energy conservation and team spirit, and 4) challenges of a family by team competition [14]. However, while many concerning games centered on baseline comparison for of these attempts have demonstrated short-term positive energy conservation. effects the long-lasting effects are questionable as people tend to disengage over time [15]. Author Keywords Persuasive technology; gamification; visualization; energy In this paper we turn our attention towards workplaces by consumption; energy awareness; behaviour change. exploring and evaluating how a cooperative game could support energy conservation. We do this by presenting the ACM Classification Keywords design, implementation and evaluation of a cooperative H.5.m. Information interfaces and presentation (e.g., HCI): game designed to motivate workers to reduce electricity Miscellaneous. consumption in a commercial office setting. 1. INTRODUCTION 2. BACKGROUND The rapid growth of energy consumption has raised increasing concerns over the last decades. According to the 2.1 Theoretical Model of Behavioural Change International Energy Agency there has been a dramatic In order to design for behaviour change, it is important to increase in primary energy use (49%) and CO2 emission understand how occupants’ current behaviour is shaped and (43%) from 1984 to 2004 [1]. Moreover, energy influenced. A well-known behavioral theories is the Theory consumption is estimated to keep rising due to economic and of Reasoned Action [6]. They believed that individuals’ population growth [1]. The U.S. Energy Information intention to perform a certain behaviour led to the actual Administration has reported that the building sector behaviour and the behavioural intention was determined by consumes 20% of the worldwide energy [2]. It has similarly the attitude towards behaviour and subjective norm. been found that around 40% of energy was consumed by Subjective norm is the social pressure that individuals buildings in Europe in 2015 [3]. Those numbers proved the perceive for a certain behaviour and it is decided by urgency to cut down energy consumption and increase individuals’ normative belief and motivation to comply. energy efficiency in buildings, which calls for the attention The Theory of Planned behaviour extended this model by of both organizations and individuals. adding perceived behavioral control [16]. A perceived Several studies have shown that there is a gap between behavioural control means how much an individual has estimated energy consumption and the actual energy control over a specific behaviour. The Theory of Planned behaviour has been shown to explain 46-61% of the variance education by asking users to run a sustainable city [9]. The in occupants’ intention in performing pro-environmental results showed that the experimental group who played the behaviour [17]. The findings highlight that in order to game had better knowledge about the impact of their understand the occupants and design for persuasion, it is consumption behaviour than the control group [9]. beneficial to take occupants’ attitudes, subjective norms and Regarding providing gamified feedback, Energy Chicken is perceived behavioural control into consideration. a game that relates the energy consumption of each device in According to the Theory of Planned behaviour, social factors the office environment to an object in a virtual farm [26]. The play a crucial role in influencing occupants’ behaviour. size and condition of the chicken in the virtual farm was Social comparison serves as a trigger for motivating influenced by the users’ energy consumption. In a 12-week sustainable behaviour. Also, comparisons with others who evaluation with 49 occupants after collecting the are slightly better can have a positive effect on performance consumption baseline for 4 weeks, an energy reductions of [18]. What this suggests is the importance to ensure that there 13% was achieved [26]. is always a chance for each side to win in a competition. The Another game called Coralog visualized energy consumption facility also plays an important role in promoting an energy as the health status of a coral reef, whose size and colour conservation culture by organizing relevant events and changed [7]. While users were motivated to change workshops, which was proved to be related to the energy- behaviour using Coralog, they also expressed wishes to see saving behaviour of occupants in the building [19]. the data of daily energy consumption besides the more Moreover, another study revealed that public recognition of abstract representation of the coral reef. This highlights the energy-saving behaviour led to a higher decrease in energy importance to keep a balance between abstract and factual consumption than solely self-monitoring behaviour [20]. representation in persuasive system. Cost reduction has also been reported to be a motivation to In addition to providing energy consumption feedback on an save energy at offices [19], but the actual effects depended individual level, researchers have introduced social network, on the personal sensitivity to price [21]. In addition, competition and cooperation on a group level in order to psychologist Richard Katzev and his team found that achieve better user engagement through interpersonal commitment to conserve resources is more influential for relationships and the sense of belongingness [7, 9, 13, 14, occupants’ behaviour than monetary incentives [10]. This 27]. Art project Nuage Vert was a city-scale light installation might suggest that intrinsic motivation is stronger than in Helsinki that lasted for a week [12]. The artist illuminated extrinsic motivation. However, a survey with 229 students vapour emission with a lime-hued laser animation that on campus showed that the majority of participants agreed visualized the real-time electricity consumption of citizens in on the importance of energy conservation, but they reported Helsinki. This collective visualization invoked a sense of that they lack clear motivations [19]. Contextual factors, community and contribution. Boork and her colleagues such as perceived behavioural control, have been found to introduced a collective prototype called Super Graph that partly explain this low level of motivations [19]. In an office employees in the office crafted together [28]. The Super building, the energy consumption is mainly influenced by Graph was hung in the lunchroom and moved vertically energy planning and management, which results in according to the collective energy consumption. This served occupants’ low sense of control and insufficient engagement as a symbol that represented the employees’ efforts and in energy conservation. created the sense that every employee was a part of 2.2 Persuasive Technology something bigger. Climate Race is a persuasive game that Persuasive technology aims at using technology to achieve was co-designed by workers in the office, whose team goal behaviour change without coercion [22]. BJ. Fogg has was to reach 1000000 points collected by energy saving and introduced his behaviour model that stressed three elements by completing missions [8]. While Climate Race achieved of persuasive technology: motivation, ability and trigger the change of workers’ motivation, it also revealed a [23]. Motivation is the reason why a person shows a certain provoking finding that workers preferred explicitly behaviour. Ability indicates whether the person has the requesting feedback over immediate notifications. This required ability to complete a behaviour. A trigger can be a finding is especially crucial for how to provide feedback in a reminder, deadline or other elements that cause the person to non-intrusive way. perform a behaviour. Those three elements are of necessity 3. METHODS to achieve the persuasion of a certain behaviour. A study that followed this model claims achievement of the energy saving 3.1 Participants and Procedure goal [24], which suggests that taking those three elements In the study, 8 engineers (6 males and 2 females) working at into consideration could be beneficial. a corporate research center volunteered to participated in a two-week-long study with a one-week baseline period Many studies have worked on applying gamification to eco- followed by a one-week cooperative gaming period with our visualization in order to provide energy education and game EnerSpace. The age distribution was four between 31- feedback on energy consumption [7, 11, 25, 26]. One study introduced a game called EnerCities to promote energy Figure 1. The main interface of EnerSpace: competition between two spaceships, team red and blue. The dotted lines show the route towards the next planet to exploit, on the line there is a text showing the distance left for each spaceship. In the upper right corner there is a ranking list. 40 years old, two between 41-50 years old, one between 20- how the participants self-rated (5-point Likert) the 30 years old and one above 51-year-old. All participants EnerSpace impact on energy awareness. signed an informed consent. We analyzed and compared the quantitative hourly At the beginning of the study we installed an energy meter in consumption data between the baseline and the game period. each workers office and collected a one-week hourly average In this process, we divided the consumption data into three energy consumption (Wh) baseline. Data logging was categories according to the participants personal logs: designed to collect 5-6 points per minute. All the devices working hours (7:00-19:00), non-working weekday hours around each personal workspace were connected and in the (19:00-7:00), and weekends (not working). Six of the control of the participant. This included desk lights (not participants logged their exact time of arrival and departure ceiling/room armature), computer, monitor and other while two provided a timetable of their daily routine. Due to peripherals such as smartphone chargers. Prior to the study instable data logging of the energy meters, in particular each worker also completed a pre-questionnaire about their during the night when we sometimes had hourly dropouts, energy awareness. average hourly consumption was calculated using only hours with constant data. After the baseline period, the workers competed in two teams (red and blue team) using EnerSpace for 7 days. The teams 4. GAMIFIED VISUALIZATION: ENERSPACE were formed randomly by the researchers. During the study 4.1 The Design of EnerSpace we did not prompt for team meetings, everything that The design of EnerSpace went through the stages of field happened in the study was on the participants own initiative. study, defining personas and scenarios, design iteration and At the end of the study, each worker completed a post- developing a working high-fidelity prototype. The field evaluation questionnaire about their energy awareness and study included an observation of energy consumption user engagement and a 20-min interview concerning their behavior in a corporate research building and an interview experiences with EnerSpace. During the study, participants with 15 workers who work in the building. The findings from also logged the time when they came to and left the desk. the field study suggested that they had low energy awareness, 3.2 Data Analysis knowledge and engagement in energy conservation. This study collected qualitative data from the interview and Although the workers generally considered themselves quantitative data from the questionnaires as well as the motivated for energy conservation, they did not actively experiment. As for qualitative data, we performed a thematic monitor or take actions to save energy. content analysis for analyzing the transcripts from the The coffee room was selected as the setting for the game as interview [29]. In this process we generated codes from the it is the commonplace where co-workers have a break and transcripts, defined themes and interpreted the results. As for chat. Assuming this would be a natural setting for teams to the questionnaires, we compared the pre- and post- discuss and follow their efforts. questionnaire to assess whether the cooperative game increased the energy awareness of occupants and looked at On the basis of the field study, 8 ideas were generated and a brainstorming was conducted together with 2 UX designers Figure 2. Individual page (a,) animation where the avatar talks about the estimated time of arrival. (b) a postcard sent by the avatar when arriving to Mars. (c) The heating system is turned off due to over-consumption. (d) The team is ready for a for a space jump since they have saved energy for 5 consecutive days. to converge the ideas to a competition between departments chosen to remove the need for login and make the experience in a space trip context. more ubiquitous. On the individual view, the participant’s consumption data is visualized as the living condition of an 4.2 Gamification Mechanics The narrative of Enerspace is that two spaceships are sent to avatar and also displayed on a control panel on the interface. the Universe for a resource exploitation to cope with the Each user has an avatar on the spaceship who serves as a resource shortage on Earth (see Figure 1). Users are divided concierge to introduce the user’s personal consumption data. into two teams (team red and blue) and each team owns a In this animation, the avatar first greets, then talks about the spaceship whose color corresponds to its team color. The user’s personal consumption data and the estimated time of spaceships are powered by the energy that the team members arrival to the destination (see Figure 2(a)). The consumption save in the office by comparing the team’s current hourly data shown on the control panel include an average hourly average consumption (Wh) the baseline period on working consumption (Wh), a percentage of saving compared to the hours and non-working hours respectively. This was user’s baseline and a ranking within the team. When the visualized in the gamified by the speed of the spaceship and spaceship arrives at a destination, the avatar will send a rotation of the flame behind it. The more the team members postcard or souvenir to the user as a reward (see Figure 2(b)). saved, the faster the spaceships moved and the flame rotated. Gamified Feedback If the energy consumption exceeds the baseline, both The energy consumption of the whole team influences the elements stopped. A dotted line shows the route of a living condition of their avatars. When the consumption of a spaceship with a number indicating the distance. The dotted team exceeds the baseline, there are three levels of power line of the leading spaceship is presented slightly clearer than failures inside the spaceship that move from running out of the other in order to show the ranking in a subtle way. When coffee to shutting down the heating system (see Figure 2(c)). arriving at a planet they can start exploiting recourses. The Increasing Gamification Engagement amount of collected resources is shown on a ranking list at Special events are triggered when specific conditions occur the top-right corner of the interface (Figure 1). When the to increase gamification engagement. When a team keeps resources on a planet are depleted, the spaceships continue saving energy for 5 consecutive days, the spaceship to the next destination. accumulates enough energy for a space jump that shortens Besides the main view, an individual view was shown when the distance to the destination (see Figure 2(d)). Other events a participant’s face was recognized. Face recognition was may also occur in order to award the leading team and motivated for energy conservation. All participants took encourage the other team. actions to save energy either intentionally or by habit. Situation-based Saving Tips Turning off the lights was mentioned by all participants as The avatar gives situation-based saving tips occasionally. an action to save energy, while only 25% of the participants For instance, the avatar suggests “I heard the weather in turned off the computer or made it sleep when not in use. [town] is great. Why not turn off the lights if it is bright Post-questionnaire outside?”. Those tips provide users with approaches to Results about energy awareness showed that 25% of the increase energy efficiency. The content is based on current participants knew the amount of their energy consumption situations including but not limited to weather and place. while 25% of the participants had a general idea of the 4.3 Implementation amount. Also, 75% of the participants thought they were The EnerSpace was set up with a 1080p monitor and a more aware of their energy consumption after the game. All Logitech C920 HD Pro Webcam in the coffee room in the participants thought it was fun to play the game and reported building. The energy consumption of each participant was being engaged with EnerSpace. They were interested in measured using a FIBARO Wall Plug. The data was knowing how the competition and space trip goes and 87.5% collected by a VeraPlus Advanced Home Controller via Z- of them stated that they thought of the game in their spare wave protocol and then sent to a server. The software of time. 60% of the participants felt more motivated by the EnerSpace was developed in Python Flask, HTML5 and game while 50% of them thought they saved more energy JavaScript. The animation and 3D model in EnerSpace were due to it. When asked whether they thought that they saved made in Cinema 4D, Adobe Fuse CC, Adobe Mixamo and more energy during the game, the responses were neutral, which was followed up for how the participants viewed their Adobe Premiere CC. energy conservation during the game in the interview. As 5. RESULTS many as 75% of participants were interested to keep playing 5.1 Consumption Data the game. In summary, besides a positive experience, Participants’ baseline average hourly consumption (Wh) EnerSpace appears to have helped to increase the varied from 16.99 Wh to 58.45 Wh (M = 34.33, SD = 15.17). participants’ awareness and knowledge towards energy to The average hourly consumption decreased by 21% during some degree. the game (M = 27.26, SD = 9.14). 5.3 Post-experiment interviews In the analysis the hourly data was categorized as working Impact of EnerSpace hours, non-working hours in a workday and weekends. When The results about the impact of EnerSpace was in-line with comparing the average hourly consumption between, before the post-questionnaire. All the participants thought they and during the game for these three time periods gained a better understanding of their consumption after respectively, all the participants consumed less while playing playing the EnerSpace. Although the company has had a goal the game except for the non-working hours in a workday (see to reduce energy use for a long time, they generally felt more Figure 3). For the non-working hours in a workday, the motivation and engagement as an individual because of the average hourly consumption increased by 48% during the game. Some participants reported that they sometimes gameplay week (M = 7.32, SD = 8.08) than before (M = 4.96, thought about the game before they left the office and that it SD = 6.29). Conversely, results indicated a large decrease of reminded them to shut down the computer and turn off the 81% in energy consumption during the game (M = 1.73, SD lights. However, even though they were motivated for saving = 2.15) than before (M = 9.18, SD = 13.93) during the energy, they mentioned that some of their regular work tasks weekend. Even though both periods are off-duty hours for (e.g. running simulations) took time to run and that they had the participants, the different results might suggest a to keep those simulations running during the day. Therefore, different behaviour pattern for the weekend and non-working they tried to save on the occasions that they could, but were hours in a workday. Besides, the average hourly consumption during working hours was reduced by 23% during the gameplay week (M = 56.80, SD = 21.29) than before (M = 74.13, SD = 36.57). These results showed a general tendency of reduction in energy consumption during the gameplay week with different degrees of decrease for the three periods of time. 5.2 Questionnaire results Pre-questionnaire The results indicated that 87.5% of the participants did not know how much energy they consumed and the rest only had Figure 3. The line chart shows how the average hourly a general idea of the amount. All the participants showed consumption changed after the game. interest in energy consumption data and 87.5% of them felt aware that some energy had to be consumed in order to do some participants who preferred to know how much money their job. they saved. Gamification Mechanics Another complaint was that the participants did not know As for the mechanics, all the participants agreed that it was how much their personal baseline was and how the personal fun to play and the competition motivated them to save saving was calculated, indicating that a more detailed energy to different extents. Some participants felt motivated explanation might have been needed. Also, some participants by the sense of competing with another team, while others wanted a clearer representation of the connection between were mostly attracted by the comparison between their their energy conservation and game, since they wondered current consumption and the corresponding baseline. how their saving influenced the speed of the spaceship. Knowing their consumption data and how much they saved provided sufficient motivation for them. These participants Regarding the in-team ranking, some participants that were appeared to consider energy saving as a competition with ranked low found themselves thinking about what they did themselves although the competition still added some wrong compared to other teammates. This indicated that the additional value. in-team ranking could serve as a reminder for those who over-consume energy. As for the cooperation, participants who were sitting Coffee room as the context physically close to each other typically liked the idea of All the participants liked the idea of using the coffee room as competing as a team because of the sense of community. This the context for EnerSpace. In addition to creating a local made them have in-team discussions around energy narrative for energy conservation, it also raised awareness to conservation and created a social pressure. However, one other workers in the building. The participants were participant mentioned a shortcoming in the study that they frequently asked about what was going on and mentioned did not experience a process of making up a team, which that it created a discourse around energy conservation. resulted in a weaker sense of cooperation and belongingness Another reason that was mentioned was that it created a during the week, physical proximity was also mentioned to positive “interruption” for work. Note that this was taking be an issue for a sense of team. place typically in the coffee room during breaks and did not Gamified Data Visualization interfere much with their work. As for the game elements, most participants liked how the movement and position of the spaceship represented their Nevertheless, some participants wanted the game both in the energy conservation, since it was easy to see what changes coffee room and in their office. The advantage mentioned their behaviour brought. Having a personal avatar that was that it would allow checking their consumption more introduced the progress of the space trip was perceived as fun frequently and understand the real-time effects of their and made the game more interactive. Although all behavior in their office context. Also, the coffee room could participants cared mostly about the consumption data, they be hard to reach for some participants who spend most of the would prefer to have both the data and the avatar, which felt time on meetings or in the laboratory. more interesting and engaging than solely factual data. Suggested Improvements According to one participant this was because it made it more In addition to the suggestions mentioned above, other desired into a game rather than facts on a screen. As for the saving features were proposed by the participants. One future advice that the avatar provided, some participants felt improvement was to expand the scale of the game and persuaded by it and thought “why not?”. Also, some included more equipments into the game, such as the heating participants reported a sense of proudness when receiving a system. Since the heating system is one of the main compliment and postcard from the avatar. contributors of energy consumption, some participants thought they would be more motivated for saving a larger When talking about the consumption data displayed on the amount of energy. Another possible improvement mentioned control panel (see Figure 2(a)), all participants thought it was to add the game to a desktop or mobile platform, making gave them an idea of how much they consumed and saved. play possible in a wider context. However, some participants found that only one number of their current consumption was not enough and it was hard for 6. DISCUSSION them to get a general idea of their consumption from it. All Combining the results from the data analysis, interview and participants wished to add a line chart of how the energy questionnaire, results indicate that EnerSpace could help consumption changed over time, from which they could be increase energy awareness and support short-term behavioral aware of how their consumption changed at a different time change. Although long lasting effects are unknown, a period of a day and compare the consumption between knowledge acquisition on energy consumption is still an weeks. Additionally, some participants wished to know how important first step to enable behavioral change. much each device consumes and their total consumption According to the results, the participants achieved a total instead of an hourly average consumption. A monetary decrease of 21% with 23% decrease during working hours indicator was suggested to provide a stronger motivation for and 81% on weekend. Those results indicate a positive effect of EnerSpace on energy conservation. However, the energy appears crucial to carefully consider the physical proximity consumption in the non-working hour in a workday of people in the team and the formation process of teams in increased by 48% during the game. One possible reason was order to promote cooperation and discussions around energy that the experiment period was too short. Another reason conservations at workplaces. according to the interviews might be due to a limited ability Regarding the setup context, having it in the coffee room, to change. Although they were aware that keeping the this was perceived as positive as it naturally created a center computer on when absent from the workplace was a waste, point and it prompted discussions and comparisons for the in this context they had to keep some simulations and workers at the department. Furthermore, the workers experiments running. However, if they were leaving for a emphasized that its central and obvious location caused longer time, such as over the weekend, they would consider curiosity from other groups and departments at the company turning everything off. This reasoning might explain the and in this way, it created a larger discourse on energy substantial decrease of energy consumption over the conversation at the company. On the downside, it was weekends compare to non-working hour on workdays. This difficult for the workers to experiment in real-time with their hints towards a potential to design for the ability to save office environment and therefore it was difficult and tedious energy when absent, rather than when present, in to explore and understand how they could improve their professional settings where workers may have a limited energy efficiency and position in the game. Therefore, a personal space (ability) for energy conservation at work. reasonable suggestion put forward by some workers was to This leads us to a suggestion of remedies to the problem of extend the game to their smartphones to allow for personal participants being unable to save much energy on non- experimentation and improvements. Extending it to a working hours on workdays. The main reason mentioned personal medium, in addition to personal performance was that participants had simulations running – this was feedback, could potentially also be used in many ways to beneficial from a work efficiency perspective – after leaving improve the competition through notifications about the office so that they would be completed in the morning. advances, ranking, and events. Regardless of the simulation took a few hours or all night, Another aspect of how the mechanics worked that was the computer was on all night. A potential remedy would be questioned by the workers was how their personal baseline to investigate if contexts such as this might make better use was calculated, what it was in comparison to others, and if of automatic shutdown of computers and peripherals on idle the competition was fair. This could be particularly or after completed simulations. interesting information for low performers who struggled in Results from the interview data analysis reveal three the game, as this might have been due to good performance influential motives that drove the participants to energy during the baseline week causing them to chase a difficult conservation, two group level motives and one individual. goal during the game week. This has caused us to wonder On the group level, the main motive was the competition, how baseline-based competition could be setup in a fair way where the participants strived to act more energy efficient to and how the baseline should evolve over a longer period of get an advantage. Another motive was social pressure and time. For instance, should it be static (fixed on the baseline), the responsibility of belonging to a team, which motivated to or recalculated daily, weekly or monthly. The choice would contribute and also created a sense of guilt for poor most definitely affect long term game dynamics and tactics performers. Although creating negative feelings may have and the question is what way would be the most energy implications for the game experience and overall motivation, efficiency promoting approach. Our setup used a static no participants reported lower engagement due to these baseline, but with a changing baseline performing good one aversive feelings. This rather caused them to reflect on what day in the game might lead to low game performance the they did wrong and what could be improved. This could consecutive day. This approach would require well designed suggest that aversive feedback does not harm the fun of and graspable game mechanics so energy conservation gamification as long as the participants have the ability to performance does not appear as random, which might disturb change. This implication is in line with the findings of study learning and have negative long-term effects due to faulty BinCam [30] and study Power Ballads [31], which both mental models. Another issue that was highlighted was the indicate that evoking aversive feelings, such as guilt, does difficulty to assess what of their equipment in the office not necessarily deter user engagement. On the individual side, contributed mostly to energy use, suggesting de-aggregated some instead focused on competing with themselves. energy feedback. Managing to save more energy than before created a sense of In EnerSpace, data was visualized using both abstract and achievement that satisfied and motivated. factual approaches inspired by Coralog [7]. As suggested by Although some participants reported that they checked the the Coralog, this study improved the abstract gamified game together and discussed it with teammates, this visualization by displaying an animated avatar and combined happened mainly between colleagues whose workplace was a factual and abstract visualization by adding accurate factual physically close. For the others there was less team spirit and data to the animation such as the current energy consumption a low sense of belongingness to the team. Therefore, it and saving. All participants liked this combination and considered it better than displaying solely factual data, participants reported an engaging and fun experience with similar to previous studies [32]. On the other hand, many EnerSpace in the coffee room, which created a local and participants asked in the post-interview for historical charts, extended narrative around energy conservation at the de-aggregated data, and personal baseline data, so some workplace. However, as the central placement was degree of tension between abstract data factual data was disconnected from their office environment it hampered captured in the study. This might suggest that factual data is learning and experimentation locally in the office. needed for creating trust and learning of consumption, but Furthermore, the formation and composition of the teams perhaps this could be hidden but accessible for the curious might need to better account for the location of the workers user to not stand in the way for the overall game experience. and social connections to enhance the discourse on energy conservation and team spirit. Regarding game mechanics, The interest to keep playing the game was high after the some workers struggled grasping the baseline comparison study although that would require that the game was and questioned its fairness, which poses challenges for how expanded with new complexities and additions, for instance to design fair and graspable baseline-centered games for by adding a heating system to manage in the game and energy conservation. scaling up the game. Another interesting idea would be that the team could gain more detailed data and game tools by REFERENCES behaving well. For instance, if they have saved energy in [1] L. Pérez-Lombard, J. Ortiz, C. J. E. Pout, and consecutive days or reached a goal of collecting resources, Buildings, "A review on buildings energy they gain a view showing more detailed consumption data as consumption information," vol. 40, no. 3, pp. 394- a reward. 398, 2008. [2] EIA. International Energy Outlook 2016 [Online]. As regards the choice of the game theme of resource Available: exploitation, some might argue that this communicates https://www.eia.gov/outlooks/ieo/pdf/0484(2016). values counter to sustainability. In the design of the game, pdf this was discussed and our thinking was that people could [3] Ec.europa.eu. Consumption of energy - Statistics distinguish this as fictional and as a satire of the absurdity of Explained [Online]. Available: resource exploitation and the current needs of mankind. http://ec.europa.eu/eurostat/statistics- However, we did not follow up on the interpretation of this. explained/index.php/Consumption_of_energy 6.3 Limitation [4] T. Maier, M. Krzaczek, J. J. E. Tejchman, and The connection between the FIBARO Wall Plug and the Buildings, "Comparison of physical performances VeraPlus Controller was not fully stable during nighttime, of the ventilation systems in low-energy residential which resulted in a low frequency of consumption data. houses," vol. 41, no. 3, pp. 337-353, 2009. Although this issue was accounted for by the formula of total [5] V. Fabi, R. V. Andersen, S. P. Corgnati, and B. W. consumption, the accuracy probably was affected. Also, the J. B. S. Olesen, "A methodology for modelling participants’ presence in the office was not fully taken into energy-related human behaviour: Application to consideration. If the participant left the desk in working window opening behaviour in residential hours without shutting down the devices, their consumption buildings," vol. 6, no. 4, pp. 415-427, 2013. data would still be compared to the baseline of working hour, [6] M. Fishbein and I. Ajzen, Attitudes and voting which lowered the accuracy of energy saving. Furthermore, behavior: An application of the theory of reasoned although EnerSpace had a positive influence on the energy action. 1981, pp. 253-313. awareness and behavior during the short test period, the long- [7] T. Kim, H. Hong, and B. Magerko, "Coralog: use- term effects are unknown. Lastly, the fact that the aware visualization connecting human micro- participants where all engineers may have influenced their activities to environmental change," in actions and behaviors, however, their engagement in energy International Conference on Human Factors in use was prior to the study not remarkable. Computing Systems, 2009. [8] J. Simon, M. Jahn, and A. Al-Akkad, "Saving 7. CONCLUSION energy at work: the design of a pervasive game for In order to understand how to increase worker awareness of office spaces," presented at the Proceedings of the energy consumption and motivate energy conservation, we 11th International Conference on Mobile and have studied a cooperative game called EnerSpace. After a Ubiquitous Multimedia, Ulm, Germany, 2012. two-week experiment with EnerSpace, participants achieved [9] E. Knol and P. W. D. J. S. S. E. P. Vries, "EnerCities a 21% decrease in overall energy consumption during the - A Serious Game to Stimulate Sustainability and game week compared to the baseline. Surprisingly our Energy Conservation: Preliminary Results," 2011. intervention lowered the energy consumption during the [10] K. Neuman, Personal Values and Commitment to weekends substantially, suggesting a potential to design for Energy Conservation. 1986, pp. 53-74. the ability to save energy when workers are absent. Along [11] D. Geelen, D. Keyson, S. Boess, and B. J. J. o. D. these lines better support for automatic or planned shutdown R. Han, "Exploring the use of a game to stimulate of office equipment can be an important implication All energy saving in households," vol. 10, pp. 102-120, [26] B. Orland, N. Ram, D. Lang, K. Houser, N. Kling, 2012. and M. Coccia, "Saving energy in an office [12] T. J. X. C. t. A. M. f. S. Holmes, "Dialogical environment: A serious game intervention," Energy encounters: art and energy awareness via eco- and Buildings, vol. 74, pp. 43-52, 2014/05/01/ visualization," vol. 17, no. 4, pp. 52-55, 2011. 2014. [13] A. Gustafsson, M. Bång, and M. Svahn, "Power [27] B. Fogg, Persuasive Technology: Using Computers explorer: a casual game style for encouraging long to Change What We Think and Do. 2002. term behavior change among teenagers," in [28] M. Boork, A. Gustafsson, S. Dijkhuis, and C. International Conference on Advances in Computer Katzeff, "OfficeWise: Energy feedback in office Entertainment Technology, 2009. workplaces," in Eceee Summer Study on Energy [14] A. Gustafsson, C. Katzeff, and M. Bång, Efficiency, 2015. "Evaluation of a pervasive game for domestic [29] V. Braun and V. J. Q. R. i. P. Clarke, "Using energy engagement among teenagers," in thematic analysis in psychology," 2008. International Conference on Advances in Computer [30] A. Thieme et al., "We've Bin Watching You: Entertainment Technology, 2008. Designing for Reflection and Social Persuasion to [15] V. Nisi, D. Nicoletti, R. Nisi, and N. J. Nunes, Promote Sustainable Lifestyles," in Sigchi "Beyond Eco-feedback: using art and emotional Conference on Human Factors in Computing attachment to express energy consumption," in Acm Systems, 2012. Conference on Creativity & Cognition, 2011. [31] D. Foster, C. Linehan, S. Lawson, and B. Kirman, [16] I. Azjen, "The theory of planned behavior", "Power ballads: deploying aversive energy Organizational Behavior and Human Decision feedback in social media," in Chi 11 Extended Processes, Vol. 1991, pp. 179-211. Abstracts on Human Factors in Computing [17] S. D’Oca, C. F. Chen, T. Hong, Z. J. E. R. Belafi, Systems, 2011. and S. Science, "Synthesizing building physics with [32] J. Ham and C. J. H. J. I. J. o. S. R. Midden, "A social psychology: An interdisciplinary framework Persuasive Robot to Stimulate Energy for context and occupant behavior in office Conservation: The Influence of Positive and buildings," vol. 34, pp. 240-251, 2017. Negative Social Feedback and Task Similarity on [18] C. Grevet and J. Mankoff, "Motivating Sustainable Energy-Consumption Behavior," vol. 6, no. 2, pp. Behavior through Social Comparison on Online 163-171, 2014. Social Visualization," 2009. [19] E. Azar and H. Al Ansari, "Framework to investigate energy conservation motivation and actions of building occupants: The case of a green campus in Abu Dhabi, UAE," Applied Energy, vol. 190, pp. 563-573, 2017/03/15/ 2017. [20] R. G. Bittle, R. Valesano, and G. J. B. M. Thaler, "The Effects of Daily Cost Feedback on Residential Electricity Consumption," vol. 3, no. 2, pp. 187- 202, 1979. [21] J. H. Yoo, K. Yuasa, and H. J. J. E. S. R. Hwang, "Evaluation of measures to improve residential energy policies considering occupant characteristics," vol. 15, pp. 33-43, 2017. [22] B. J. Fogg, "Motivating, influencing, and persuading users," in Human-computer Interaction Handbook, 2009. [23] B. J. J. R. J. Fogg, "The new rules of persuasion," vol. 155, no. 5538, pp. 24-29, 2009. [24] M. H. Lester Shen, Carl Nelson, Jeenifer Edwards, "Human-Building Interaction (HBI): A User- Centered Approach to Energy Efficiency Innovations," ACEEE Summer Study on Energy Efficiency in Buildings, 2016. [25] H. M. Chen et al., "Persuasive feedback model for inducing energy conservation behaviors of building users based on interaction with a virtual object," vol. 45, no. 2, pp. 106-115, 2012.