Designing Mobile Technology to Promote Sustainable Food Choices Conor Linehan, Jonathan Ryan, Mark Doughty, Ben Kirman, Shaun Lawson Lincoln Social Computing Research Centre University of Lincoln, Brayford Pool Lincoln, LN6 7TS, UK. ++44 01522 837084 clinehan@lincoln.ac.uk ABSTRACT environmental sustainability and social sustainability. Indeed, This paper is an experience report based on challenges within environmental sustainability, there exist subtleties that encountered when designing scalable mobile persuasive HCI make it hard to define how sustainable any given item is. For applications to help users make informed choices over their food example, the question of whether it is preferable to grow fruit at a consumption. We recently developed Tagliatelle, a social tagging low carbon cost in the third world and air freight it to the UK, or system to help users to accurately monitor and assess their dietary to grow the fruit at a higher carbon cost in the UK, is a dilemma behaviour and to promote healthier food choices. In this paper we that currently appears to be a value judgment. Since the problem propose a similar system in order to help users understand the domain is so unclear, it is difficult at present to create meaningful sustainability of their food choices. We discuss the challenges applications that give judgement on an objective level. inherent in doing so, and extrapolate some important issues that Complicating the issue further, there is currently no requirement need to be addressed by technological developments that aim to for manufacturers to disclose where ingredients and components persuade users to adopt more sustainable behaviours. have been sourced (known as supply chain transparency [1]). Nevertheless, in order to design mobile tools to encourage more Categories and Subject Descriptors sustainable consumption, we must have some useable definition of sustainability. As such, in our recent work, we have adopted H.5.3 [Group and Organization Interfaces]: collaborative the goals of the “Slow Food” movement, which emphasises the computing. consumption of local and seasonal produce over that which is imported and/or out-of-season (see http://www.slowfood.com for General Terms more details). Hence, in the technology proposed here, users’ Design, Human Factors, Theory. food consumption will be evaluated in terms of how closely it adheres to the goals of the “Slow Food” movement. Keywords Eco-feedback, sustainability, sustainable consumption, tagging, feedback, persuasive. 1. INTRODUCTION 1.1 Background Recent studies have identified that topsoil erosion [17], depletion of fish stocks [13], tainting of meat products, depletion of oil reserves and climate change can all be linked to the method in which food is currently produced, distributed and consumed [17]. It is clear that reaching and understanding of, and improving the sustainability of, food that we purchase and consume is of growing interest [9]. As social computing researchers we are interested in how online mobile and social technology may facilitate these goals. In particular, we believe there is a need to directly engage the individual consumer in the process. Figure 1. In Winter the UK imports potatoes from Egypt, grown in the desert with seed from Scotland, water from 350m It is also clear, however, that, there is currently no all- deep wells, and packed in peat from Ireland. encompassing measure of sustainability that we can use to deliver feedback to users. For instance, there are a number of different issues that the term ‘sustainability’ can refer to; these include There are also a number of challenges facing any programme, technological or otherwise, that aims to change consumer habits. Copyright is held by the author/owner(s). MobileHCI 2010 September 7-10, 2010, Lisboa, Portugal. For example, although reports show that consumers are prepared ACM 978-1-60558-835-3/10/09. to pay more for eco-friendly items [5], and rate sustainable items . as of high importance, in fact they rarely purchase such items analysis may prove useful for a number of different tasks, [14]. It appears that in order to bridge this attitude-behaviour gap, including food sustainability. consumers need both access to sustainable produce and confidence in their ability to identify sustainable produce [11]. 2. A MOBILE APPLICATION TO We believe that significant potential exists with existing mobile technology to develop tools that allow people to identify the ENCOURAGE SUSTAINABILITY IN FOOD overall sustainability of their personal food purchases and take CONSUMPTION action to improve it. Indeed, the inclusion of motivational tools We are interested in exploring the possibility of harnessing the such as visual feedback, goal-setting and mini-games may help apparently intrinsically motivating activity of tagging images as a persuade consumers to make more sustainable choices. means of creating mobile applications that allow users to accurately monitor, assess and change the sustainability of food 1.2 Tagliatelle they consume. We believe that this type of approach may prove In previous work, we attempted to utilise the persuasive power of very effective in helping users to gain an overall picture of the social media as a means of facilitating dietary behaviour change sustainability of their own food choices. The main advantages of [7]. Specifically, we identified that the development of new and designing a system with a social tagging architecture are both the innovative methodologies aimed at helping people determine the lack of need for expert involvement and huge potential for nutritional content of their own food intake and motivating them scalability. to choose healthier options is an urgent goal. We proposed that Thus, we propose a system based on our experiences in the exposing participants’ eating habits to each other may act as design, development and evaluation of the Tagliatelle project. triggers [6] for motivating both healthier food choices and the However, instead of taking photographs of prepared meals, maintenance of those choices over an extended period of time. In participants will photograph their food at the point of purchase. order to examine this, we developed an application in which users In addition, as mentioned above food consumption will be uploaded digital photos of meals that they had eaten to a server, evaluated in terms of how closely it adheres to the goals of the which anonymously distributed these photos to other users for “Slow Food” movement. tagging. Each user was required to tag one photo that had been previously uploaded by another user before they could upload a The system will be composed of a mobile phone application and a photo of their own. In addition, users were free to visit the server-side database. Users will interact with the database website at any time in order to tag randomly selected images. primarily through the mobile application, although it is envisioned Thus, each photo uploaded was tagged several times by different that a standalone website will also be created. The application users, generating a rich history of tags for each photograph will allow three interactive experiences: uploading of photos, uploaded. tagging of photos and presenting of feedback. These are now discussed in turn. 2.1 Photo Uploading The mobile phone application will allow users to take photographs of their purchases and to easily upload these photos to their personal profile on the server with one button click. The server will anonymously assign all uploaded photos to other users for later tagging. One particular challenge lies in motivating users to photograph each individual item that they purchase and upload these items to the server. Failure to report a significant proportion of food items, or the selective uploading of only ‘good’ items would lead to inappropriate feedback. Exactly which tools are most effective at motivating honest participation is an empirical question that we Figure 2. Screenshot of image tagging in the prototype intend to pursue over the course of this and related work. tagliatelle application. 2.2 Photo Tagging An evaluation of a basic prototype of Tagliatelle [7] suggested Users will have the option of tagging photographs either through that although we encountered problems extrapolating valid the mobile phone application itself, or through a standalone nutritional information from the tags generated by participants, website. Specifically, a mini-game, inspired by [15] and [16] will the activity of tagging fellow users’ uploaded food photographs be created in which users rate the food content of the photos was very popular among participants. This finding is consistent presented in terms of sustainability. As in [16], ratings will only with work in the field of human computation ([15][16]), where be accepted if agreement is reached between independent raters. games are used to motivate users to tag digital images with Exactly what form these ratings will take is, at this time, an relevant content labels that can later be used in text-based image empirical question. There is no obviously superior option between retrieval. In effect, the players of these games function as a data numerical, visual or other methodologies. However, we do analysis tool. It seems that this type of crowd-sourced image recognize that a vital part of this research will involve educating users on how closely items do or do not adhere to the goals of 4. CONCLUSION “Slow Food.” A discussion addressing the problems facing any mobile application that attempts to promote sustainable food choices has 2.3 Providing Feedback been presented. We have proposed the design of a system based Each user will receive feedback on the overall sustainability of around the popular activity of photograph tagging that may help their food choices through a number of possible methods such as users to gain an overall picture of the sustainability of their own graphs and mini-games. This feedback will be reported both in food choices. We have also discussed how behaviour analysis can terms of personal goals and in comparison to the mean results for help HCI researchers design the way in which feedback is other users. delivered to users, in order to create applications that are both engaging and useful. 3. DESIGNING USEFUL FEEDBACK FOR PERSUASIVE APPLICATIONS 5. REFERENCES Apart from the very specific problems of ensuring that food is [1] Bonanni, L. Hockenberry, M. Zwarg, D. tagged validly and reliably, and that participants photograph and Csikszentmihalyi, C. and Ishii, H. 2010. Small business upload appropriate quantities of their food, there are some basic applications of sourcemap: a web tool for sustainable design and issues that need to be dealt with when setting out to design any supply chain transparency, In Proceedings of the 28th technology that promotes sustainable consumption. international conference on Human factors in computing systems, 937-946. One criticism that can be leveled at the vast majority of persuasive tools, mobile or otherwise, is that although these technologies are [2] Brown, B., Chetty, M., Grimes, A., and Harmon, E. designed with the specific aim of effecting change in user 2006. Reflecting on health: a system for students to monitor diet behaviour, very few have implemented empirically established and exercise. In Proceedings of the 24th international conference methods for doing so (see [8]). Indeed, very little of the extended abstracts on Human factors in computing systems,, 1807 published work on persuasive technology gives any specific – 1812. insights into the processes involved in behaviour change, nor [3] Catania, C. A. 1998. Learning (4th ed). Cornwall-on- specific examples on how to apply these processes. Fortunately, Hudson, NY: Sloan Publishing. however, there is an entire academic discipline that sets out to [4] Cooper, J.O. Heron, T.E. and Heward, W.L. 2007. examine precisely these questions. Applied Behavior Analysis (2nd Ed). New Jersey: Behaviour analysis is the scientific study of learning [3]. It is, by Pearson/Prentice Hall. definition, practical and pragmatic, as it presumes that all [5] De Pelsmacker, P. Janssens, W. and Mielants, C. 2005. behaviour is determined by interactions with and feedback from Consumer values and fair-trade beliefs, attitudes and buying the surrounding environment [12]. Successful behaviour is behaviour. International Review on Public and Nonprofit maintained, while unsuccessful behaviour is not. Crucially, Marketing, 2, 50-69. behavioural psychologists suggest that because behaviour is determined by the environment, it can be changed readily by [6] Fogg, B.J. 2009. A Behavior Model for Persuasive analysis and manipulation of that environment (see [10] for an Design. In Proceedings of the 4th International Conference on excellent introduction to behavioural interventions; [4] for an in- Persuasive Technology, Article 40. depth analysis). Hence, the field of behaviour analysis has spent [7] Linehan, C. Doughty, M. Lawson, S., Kirman, B. Olivier, decades investigating exactly how to deliver feedback in order to P. and Moynihan, P. 2010. Tagliatelle: Social Tagging to generate real and lasting behaviour change. We believe that, Encourage Healthier Eating. . In Proceedings of the 28th regardless of the target behaviour, in order to create effective international conference extended abstracts on Human factors in persuasive technologies, the science and methodologies of computing systems, 3331-3336. behaviour analysis must be employed as an integral design phase. [8] Kirman, B. Linehan, C. Lawson, S. Foster, D. and Indeed, assuming that the principles of behaviour don’t apply Doughty, M. 2010. There's a Monster in my Kitchen: Using when a person is interacting with a computing device is a stance Aversive Feedback to Motivate Behaviour Change. In that is uninformed, and will lead to a large amount of duplication Proceedings of the 28th of the international conference extended of effort in addressing questions that have already been abstracts on Human factors in computing systems, 2685-2694. comprehensively answered. [9] Pachauri, R.K. and Reisinger, A. 2007. Contribution of In the application introduced in section 2, the way in which Working Groups I, II and III to the Fourth Assessment Report of feedback is delivered to participants will be informed by the the Intergovernmental Panel on Climate Change. Geneva, methods of behaviour analysis. Specifically, we will endeavour Switzerland: IPCC. to provide consistent, regular and specific feedback, regardless of [10] Pryor, K. 1999. Don’t Shoot the Dog. New York: whether participants reach their goals or not. This will, at times, Bantam. necessitate the considered use of aversive feedback [8]. We will also design the system so that there is a range of available reward [11] Robinson, R. and Smith, C. 2002. Psychosocial and structures, such as mini-games, social networking and competitive Demographic Variables Associated with Consumer Intention to leader boards, and will ensure that the system is adaptive enough Purchase Sustainably Produced Foods as Defined by the Midwest to recognize and utilize the types of rewards that are most Food Alliance. Journal of Nutrition Education and Behavior, 34, effective for each participant. 316-325. [12] Skinner, B.F. 1974. About Behaviorism. New York: [15] Von Ahn, L. Liu, R. and Blum, M. 2006. Peekaboom: a Random House. game for locating objects in images. In Proceedings of the [13] Thurstan, R.H. Brockington, S. and Callum, M.R. 2010. SIGCHI conference on Human Factors in computing systems, 55 The effects of 118 years of industrial fishing on UK bottom trawl – 64. fisheries. Nature Communications, 1, Article 15. [16] Von Ahn, L. 2007. Human computation. In Proceedings [14] Vermeir, I. and Verbeke, W. 2004. Sustainable Food of the 4th international conference on knowledge capture, 5-6. Consumption: Exploring the Consumer Attitude Behaviour Gap. [17] Weber, K. 2009. Food, INC. Participant Media. Journal of Agricultural and Environmental Ethics, 19, 169-194.