Abstract

Baked Potato is a mobile web service geared at addressing the imbalance of power between those who market and those who consume food products. Food marketers rarely provide a detailed range of information about products that would allow consumers to understand how a product and its company connect to their cultural values. The main goal of this application is to connect people in a way that celebrates their differences and gives them agency by helping them make better decisions about their food purchases.

Introduction

Activities surrounding food are some of the most fundamental activities that we do as human beings. Despite the variety of food in modern grocery stores consumers are not provided with the relevant information they need to make informed decisions about the food products they purchase that support their individual, social, or cultural preferences.
The information provided currently on food products is problematic in at least three ways:

1. Consumers are often powerless to control what information is provided on the labels, and how labeling is presented.
2. The information that is provided is regulated by a limited, yet powerful group of people who make decisions about how food is labeled, and consequently valued.
3. The information that is presented is not diverse enough to satisfy the different contextual needs of various populations of people. Research suggests that the information currently on food products is ineffective for consumers to make relevant decisions based on their own values [1]. For example Elise Golan suggests that food labeling is rarely effective in communicating environmental or other spillovers associated with food production and consumption, because the labeling reflects the interests of the company not the consumer [2].

For these reasons we believe that the information provided to consumers within the broader food system is complicating the activity of purchasing food.

Existing HCI Gap

Currently, there is no information technology that has been successful at changing this problematic situation with the food system. We believe this is because current apps do not combine relevant information with an incentive to act. For example GoodGuide.com provides information on the health, environmental, and social impacts of products including food. While these ratings place more agency in the hands of consumers, they offer no explicit incentive to act. Jane McGonigal argues that games provide scenarios and incentives for people to adopt new habits that foster positive social change. [3] We believe that social gaming will factor prominently in any HCI app focused on social change.

Research Process

Contextual Inquiry

Framing the problem is vital when designing for social change. It is impossible to understand the complexity of a system intended for change without a working structure which allows a designer both the ability to observe it happening and describe it to others. Activity theory is based on a model of a social world, which holds that larger social activities are made up of smaller actions [4]. This suggests that changing the injustice within the food system means changing the repeated mundane actions that reinforce the broader social system. In our case, grocery shopping is not the problem; however, it is a place to intervene, because the opportunity to change behavior through an information technology is much lower.

Figure 1: Despite the variety of food consumers are not provided the relevant information they need to make informed purchasing decisions.

Figure 2: Research suggests that the information on food products is ineffective for consumers to make decisions on their own values.

Our design team conducted a series of observations using a contextual inquiry approach to locate a particular opportunity for change. We used Grounded Theory to discover key decision points in the transcripts of our observations. These were instances of users making decisions based on a cost/benefit [4] criteria. We identified several key factors among shoppers that we grouped into categories, and understood these categories to be users’ shopping profiles. The resulting models from our observations show that different users rely on different key values to inform their purchasing decisions.

Key Food Value

1. Outside factors (e.g., shopping lists, coupons, mood, time of day)
2. Influence from other people (e.g., friends, family, co-workers, celebrities)
3. The price of food
4. How the food tastes
5. How healthy the food is
6. Trust in the brand
7. How the food tastes
8. The environmental impact (e.g., concern whether the production of the food does not harm the environment)
9. The economic impact (e.g., interest in supporting Michigan businesses)
10. The social impact (e.g., fair trade, animal welfare issues, production methods)

Analysis and Modeling

Additional observations helped us refine and confirm the key values of our shopping profiles. In the next phase we used polar graphs to map key values, and a users’ strength of preference for each value. Polar Graphs allowed us to visualize these profiles as a meaningful shape in order to compare across users, making it easier to identify preference patterns.

Figure 3. A sample of the User Profile Polar Graphs using the Key Food Values.

In addition, we developed a series of user-specific models that visualize individual products of users. For example, we were able to compare John’s overall user model against specific product choices he made (Figure 6). We found these models to be important because they helped use to see how specific products reinforced or deviated from Julie’s overall profile. We noticed is that often user product profiles deviate from their overall profile. This is important because it suggests that users are making different types of decisions across products, which we potentially need to support in our application.

Figure 6. A sample of John’s Profile Graph, which also includes specific profiles of individual products he purchased during a shopping trip.

Preference patterns are the secret sauce of our application because they provide a way for people to see their shopping in way they typically cannot do on their own. In addition these patterns are useful to tailor our application to the unique differences of each user including the incentive mechanism that nudges them to take action.

Figure 8. This is an example of early A/B usability testing using paper prototypes in order to discuss food preferences with users.

Results

Based on observations, modeling, and analysis we sought to design an application that allowed users to set achievable goals, make informed decisions, and connect with other people in several key ways.

• Create and Manage a Food Profile. Users create a food profile and sets food goals, which are tracked over time.
• Set Goals and Receive Rewards. Baked Potato rewards users for setting goals, taking challenges and making good decisions as a way to encourage regular use and forming positive purchasing habits
• Receive Product Suggestions. Baked Potato offers product suggestions based on a user’s food profile
• Manage Shopping Trips. Carts represent a unique shopping experience for a user. Users can create carts before and during a real shopping trip.
• Connect with Friends. Since friends have a tremendous potential to influence the shopping habits [5], Baked Potato supports users making connections with friends in order to facilitate action toward accomplishing their food goals.

Prototype and Usability Testing

Our approach to prototyping was influenced by Human-Centered Design practices which needs and abilities of those who are to use the software. Our prototyping method used story-driven approach influenced Behavior Driven Development (BDD). BDD is typically used by developers for testing during the coding stage of the process. BDD uses a simple syntax that all stakeholders can use describe a user’s relationship with the application. This story approach forces designers to make explicit all the behavioral assumptions of the application, which can be prototyped and tested by users.

Figure 4: A sample preference view from initial paper prototypes which shows how users can set preferences.

Figure 5: A more refined preference manager designed as a result of user feedback.

The BDD syntax is close to agile project planning, which smoothed the transition between design and development. BDD is compatible with Activity Theory because it is not about simply defining functionality, but rather describing a user’s relationship to the application, and thus a relationship to the change itself. Don Normal has argued that human-centered design methods tend to miss this aspect of behavior [7].

Figure 9: An example of the interactive prototype created using Axure RP Pro software.

Figure 10: An example of a user profile view, which allows users to easily navigate different features of the application.

In addition, since we used BDD during prototyping, the functional spec was a natural outcome of the process, not something tacked on at the end. Using the BDD scenarios (Figure 10), we initially created paper prototypes to confirm our features and refine our scenarios. We then moved into a more high-fidelity application Axure RX Pro to create interactive prototypes to test different versions of our application in a more real-world environment.

Figure 11. A sample scenario using Behavior Driven Development syntax our team used in order to create a narrative about a users relationship with our application.

Conclusion

How does Baked Potato intervene in the current problem?

Our initial goal was to build an application that has the potential to transform people’s relationship with the food system by intervening into the mundane actions of grocery shopping. Baked Potato combines relevant product information, goal setting, and social gaming in order to create an experience for users they typically cannot do on their own. Baked Potato is potentially a transformative product for users. It does not attempt to change the way people shop. Instead, it provides users with better decision-making resources, and ultimately more powerful and less marginalized shopping experience, by integrating additional information into the existing shopping experience. Feedback from our prototyping confirmed that pairing relevant information with social gaming was appreciated by our users.

Future Work

We realize there is much work ahead. Social change comes about when this becomes part of everyday practice. We see future development of this application existing alongside the usage life-cycle of users identified by Joshua Porter [8]. Usage lifecycle describes the different types of relationships people have with your software product. We recognize that people go through a progression as they use software, with unique hurdles to overcome moving unaware users to passionately engaged ones. We believe more work is needed to understand and design for first-time, ongoing, and passionate participation by users.

Long Term Goals

One of the most significant long term-goals that Baked Potato might help in dealing with the entrenched network of forces in the food system is by getting stores to change what is on their shelves, which is keyed by the moments of adoption by users.

Since grocery stores are in a position to benefit from the market forces of users, Baked Potato can work as a third-party source (either through APIs or other information technologies) in order to allow users to signal their preferences directly to grocery stores, ensuring the scalability and sustainability of our application.

Acknowledgements

Several people have made this work possible including our instructor and advisor on this project Dr. Bill Hart-Davidson. We also acknowledge the support of our fellow classmates who provided useful feedback throughout each phase of the project. Finally, our test users were indispensable in providing useful feedback during the prototyping phase.

References

[1] Bødker, S. Through the Interface – a Human Activity Approach to User Interface Design Hillsdale, NJ: Lawrence Erlbaum Associates, 1991.
[2] Consumer understanding and use of nutrition labelling: a systematic review G Cowburn… – Public Health Nutrition, 2005 – Cambridge Univ Press Y
[3] Economics of Food Labeling – Elise Golan, Fred Kuchler, Lorraine Mitchell, Cathy Greene and Amber Jessup Journal of Consumer Policy, 2001, Volume 24, Number 2, Pages 117-184
[4] Jane McGonigal: Gaming can make a better world
[5] Cost-Benefit Analysis by Richard Layard and Stephen Glaister
[6] Social Class and Life Cycle as Predictors of Shopping Behavior – Stuart U. Rich and Subhash C. Jain
[7] Donald A. Norman, Human-centered design considered harmful – ACM by DA Norman – 2005
[8] Joshua Porter, Designing for the Social Web, New Riders Press, 2008