This capstone project was a proposed donation system for surplus food on campus at the University of Washington. My team and I worked with campus stakeholders to build an app that catalogues cafe surplus and tracks its delivery to local shelters.
Problem Space
Seattle restaurant food waste and donation practices
Nearly 40% of food is wasted in the United States, and much of that food is still edible or usable. This fact is particularly troublesome in light of the 41 million people in the US who face food insecurity. Ideally, the US food system would neither overproduce food nor overlook hungry communities, but the unfortunate reality is that there is no "US food system" at all. Our country manages food with a tangle of incompatible and uncoordinated local distribution networks. Our team chose to look at the local food systems in our city of Seattle to investigate the possibility of using these two food problems to solve each other. ​Redirecting surplus goods to communities facing food insecurity benefits hungry individuals while also reducing the amount of wasted food overall.
We hoped that bringing a design approach to this complex problem space would yield a thoughtful and holistic food waste solution.
​
This is where Mittel comes in!
User and Stakeholder Research
Learning about logistic and communication barriers
Our primary research question centered on the logistic and communication barriers faced by restaurants and donation organizations in the process of donating food. After interviewing local chefs, reviewing state donation laws, and analyzing the approach of other food donation programs, we generated user personas that anchored the ideation phase of our project to a specific set of user needs.
​
Our team visited restaurants that participate in donation programs as well as donation distribution sites. We learned that food donation programs face problems that include: varied standards for food quality, compliance with state and local law, uncertain profits and funding, and liability.
FIELD STUDY
​
Our team visited Washington's largest charitable food distributor, Food Lifeline. At their warehouse we learned about how they organize and process food donations. Food Lifeline partners with the University of Washington to redirect surplus food from dining halls to local food banks and shelters. Food Lifeline and other donation programs tend to deal with large scale food surplus coming from grocery stores and food industry partners. They do some small scale work with local restaurants, but these relationships are difficult to coordinate because they rely entirely on human-to-human coordination via email.
​
​
INTERVIEWS
Based on findings from our literature review and competitive analysis we devised a set of interview questions. Our team interviewed four local chefs at their place of work with each interview lasting 20 - 30 minutes. Each interviewer carefully followed the same protocol as a method to constrain the findings. After all of the interviews were conducted, they were transcribed and coded. By coding the common themes in these interviews, our team outlined some of the prevailing issues with food waste and food donation from the perspective of a commercial kitchen:
Time and labor are primary concerns among chefs. Food donation requires taking extra time to package and store food for donation. These additional steps are a low priority in a fast paced kitchen, and are not voluntarily undertaken after intense work shifts.
The liability presented by donating food is also a universal concern among chefs. Most of our interviewees were not well-versed in the laws surrounding food donation, but they all noted that donation was a very careful process in commercial kitchens because of the potential for incurring legal risk.
​
​
LAW REVIEW
​
When donating food, kitchens must consider regulations at all levels of the process. Federal and state law encourage food donation with regulations such as the Bill Emerson Good Samaritan Food Donation Act which protects donors from any civil or criminal liability in case the donated item caused harm to a recipient. (There are however exceptions for “gross negligence.”) Locally, the King County Health Department specifies a number of holding, heating, and cooling steps that must be taken before food can be donated.
Most charities receiving donations have their own unique set of regulations regarding the time frame in which food can be accepted and consumed. While most are highly tolerant of accepting food outside of its expiration window, the multi-layered and often ambiguous policies surrounding food donation discourage many potential donors.
​
Ideation
Scaling the solution from city to campus
Our team completed a series of ideation activities including affinity diagramming, brainstorm sketching, and storyboarding. This led us to narrow our scope to prepackaged surplus food on The University of Washington campus. We felt this offered the best opportunity for generating realistic and fleshed out ideas within the given time constraints of the project.
​
After scaling our vision to the University of Washington campus we focused on two user groups: food delivery drivers and HFS employees working at University cafes.
​
NARROWING OUR SCOPE
​
To elucidate the findings from our research, we laid out a loose affinity diagram on a whiteboard with sticky notes. From
the board work we realized how just how enormous and multifaceted this project was, and we struggled to process all
the information in a comprehensive way. Our research left us at a wide crossroads: we felt we could work on educating food workers on safe donation practices, organize a volunteer based system to transport donated food, or design a more structured line of communication to fit in with Food Lifeline's efforts.
​
We ultimately arrived at the consensus that we needed to more clearly define the scope of our project and the problem we were addressing. Based on the time constraints of the project we agreed to adjust the scope of the project to address only prepackaged food waste on campus. From first hand accounts, we know that food on campus is thrown away on the day of its expiration, without any opportunity for donation or reduced price sale. Knowing that prepackaged food would be easier to handle in a distribution system, we felt we could achieve the most concrete results with this line of attack.
​
​
COMMUNITY APPROACH
Our ideation exercises yielded two possible directions our design could take. Scenario 1 depends on a volunteer based distribution system of items that are recovered from a cafe. In this scenario, we imagined a network of individuals communicating through a smartphone app, letting each other know when and where food would become available. Items would then be stored in a local fridges hosted by individuals or groups around campus, for example in student lounges or club rooms.
Scenario 2 is imagined operating under the consent of HFS. In this system, cafe employees or volunteers would be encouraged as part of policy to distribute rescued items to a centralized storage, for example, a public fridge in the HUB. In this scenario, users would be able to access information about when and where food is available, but they would not need to act as volunteers to keep the system in place.
​
​
A LITTLE MORE RESEARCH
​
Initially, we imagined this kind of system operating without the explicit consent of the entity in charge of food distribution on campus, HFS, but it became clear that the most fruitful approach on campus would require their approval. We had a very encouraging meeting with HFS directors and head chefs, and with their input we adjusted the stakeholder interests for our project. With HFS backing our mission, we were able to consider ideas that did not rely solely on volunteers. We were already considering cafe workers as part of our design, but now we had the potential to include campus delivery drivers in our system. To gather more information about this new user group I personally conducted a contextual inquiry with a delivery driver. This additional user group had a huge impact on the flow of food through our system, and we ultimately cemented the direction of our design around these crux users.
​
Design
Designing for two distinct use scenarios
Driver View
Pickup logistics
Establishing Communication
Addressing liability
HFS Employee View
Integrating our features with existing HFS tasks
Creating donation receipts
Testing
Collecting feedback from drivers and HFS employees
We conducted user testing to get feedback on how this app might fit into daily closing duties for student employees who are in charge of pulling expired items off shelves and documenting these items in waste logs. We received valuable design feedback resulting in the following points for improvement:
​
Labeling: The labels for buttons and action options were not clear to the participants. Our system included too much specific language that our team has become accustomed to, but was confusing to a third party.
​
Implied Interactions: Visual cues indicating features such as the barcode scanner were too subtle and left some users confused.
​
Fidelity of Prototype: Some of the misunderstandings around the system centered on the current fidelity of the prototype. Advancing the functionality smoothed out the user’s experience and allowed them to focus on the tasks.
Final Prototype
Polishing and presenting a final product
We made changes to our initial prototype based on user evaluations and polished some of the aesthetics. The project won runner up in the Capstone Symposium for Human Centered Design and Engineering.
Driver View
HFS Employee View
Reflection
Challenges and future implications
Overall, this project was a success. We started with a massive problem space, and in the matter of a few short weeks created a viable design to redirect surplus food to hungry communities near the University of Washington.
​
That said, there were a few points where we could have improved our project. Firstly, the initial project idea was too big. We could have gotten much further in the design and implementation if we had not devoted so much of our research and ideation phases to organizing the tidal wave of information that came along with such a broad topic. However, much of our initial research yielded key considerations that made it into our final solution, so perhaps this roundabout design path was necessary. Secondly, our user testing was quick and dirty. While there is value in any amount of user feedback, we could have further improved our design with a second round of user testing. Unfortunately, deadline pressure had become a real factor at this point in our process, otherwise we certainly would have sought feedback for the delivery driver side of the design.
​
If this system were to be implemented on the University of Washington campus, HFS administration's feedback would need to be given great weight in the final design. We attempted to involve HFS donation leaders during the project, but their lack of timely response and our short timeline forced us to forego their input. Alas, HFS is still the most important stakeholder as far as green-lighting the operation, so revisiting this project with enough time to involve campus administration would be necessary to get it off the ground.
TIMEFRAME
March - June 2017
​
TEAM
Camille Mitchell, Quinn Mau, Joanna Bailet, Kyle Musselwhite
​
MENTORS
Mark Zachry, John Porter, Doug Divine, Justin Hamacher
University of Washington Human Centered Design and Engineering Department and resources
​
TOOLS
Google Survey, Sketch, Adobe Illustrator, Adobe Photoshop, Adobe XD