Research and Design for On-Campus dining problems

Position: Rotate leadership; Focus on research
Key words: User study, Product design


In Georgia Tech, an approximately 5500 new international students arrive on campus each year. Good dining is an important part in overall adaptation to a new life abroad, but we have found the following experience widely shared among new international students (including ourselves):

“You are an international student attending Georgia Tech. While visiting the cafeteria you are confused by the many unfamiliar food choices and have difficulty understanding the menu. You cannot name the foods and do not know what foods you would enjoy. You are afraid to ask the server for information about the food because there is a long line and that would inconvenience those behind you.”

We are inspired to find a solution to help international students adapt to dining on campus in a better way.

User Research

Phase#1 Natural observation

We carried out four 30-minute oversrvation sessions in different on-campus dining places. We took notes of the behavior as well as oral communication and facial expressions of international students when they order food.

Phase#2 Task Analysis
Cafeteria Fast Food Store Buffet
This dining place requires people to interact with one or more servers. There is typically a menu, but the food is displayed behind the counter and serve as a more obvious reference.People could either communicate with words or by pointing to the food displayed. This dining place usually has only one window to place order and check out. The only reference is the menu. This dining place has open windows with no servers in most cases. Users could pick food on his or her own, with no need to communicate with others. There is often no menu, but there are often name tags of food and some information.

These three dining places have common characteristics:

  • Time is usually limited. This means that users have to take actions or make decisions under pressure.
  • The options are limited. This means that users only make decisions in not a lot many items.
  • Relative information about food and process is provided but the display area is limited. This means that users cannot get very detailed information unless they reach out to people or internet.
  • The environment is usually noisy. This means that the quality of verbal interactions is usually reduced.
  • The space is usually crowded with people, who are most likely in a hurry as well. This means that users have little privacy in the process, and have to care about others’ opinions.

The task analysis diagram is shown on the right.

Phase#3 Contextual Interview


After natural observations, we need to understand further about our users’ feelings and problems, as well as the reasons underneath. We also need to identify how users’ personal traits relate to their problems. Contextual interview gives us opportunity to choose different settings in which our task would take place: cafeteria, fast food store or buffet. It also enables us to fully probe into a few typical characteristics of people like backgrounds, attitudes and behaviors, helping us to see more beyond the limit of our previous assumptions.


Eight contextual interviews with first-year international students.

Phase#4 Affinity Mapping

An affinity map is created using the findings from CIs and observations. The standard procedure of creating affinity mappings was followed where we began with reading out aloud the note and sticking it on the wall. The subsequent notes were seen for matches and if not a new column was created. After this, we created blue labels highlighting the basis for column grouping. The pink labels were then introduced to group the various ideas grouped by the blue labels. Finally, the green labels were written to cover the pink labels giving us the final affinity diagram.

Phase#5 Surveys

We distributed the survey through social media groups of international students. There are 60 valid responses.

Some key findings
Summary of User research

Phase#6 Personas

For our personas we analyzed our design needs and recommendations and compared those with the feedback we received from our interviews and surveys. We matched our personas to the broad trends we perceived in our data. We were able to identify three different major personas from these trends.

The major personality factors included length of stay in America, introversion/ extroversion, and eating habits. The minor factors included wanting healthy foods and country of origin.

Primary Persona

Yan is a little introverted and less likely to ask questions in the process of ordering. She is much more aware of the social consequences of trying to order and doesn’t wish to take a lot of time in that process. Her response to the anxieties from trying to order results in her being more likely to skip a restaurant that she can’t understand the foods and menu for. Her personal solution to this problem is to rely on friends who have already gone through this process to either order for her or provide recommendations. However, she would like to try new foods, just that the difficulty in understanding what they are means that she would rather not without sufficient information.

We made Yan as our primary persona because she has the most difficulty in adapting to dining on campus. She represents a large group of people that we should be a major focus of our design choices.

Secondary Persona 1

The second persona is Ram, who has been in the United States for 1 and a half years. Ram is meant to represent certain trends that we noticed in our data. What we found in our research was that despite being in the US for a year or two, international students were still unfamiliar with American foods and uncertain about the process of ordering. This means that Ram shares many of the same problems as the first persona. The major difference is that Ram is more outgoing and less concerned with asking questions and taking his time ordering. The unfortunate consequence that we found was when ordering the talking back-and-forth to the server only reveals more issues and causes more confusing in the ordering process. Ram places a higher priority on knowing what foods actually are, and if he would find them tasty. While Ram is willing to try foods, he may not like them and is often disappointed when he orders something that isn’t like what he wanted.

Secondary Persona 2

Pablo, is a graduate student that has lived in the United States for over 3 years. He could be considered more of an “expert user” among international students. Pablo has been in the US long enough that he is familiar with common American foods and the cultural process of ordering. He still has several difficulties, but Pablo has different priorities than the other personas. Pablo is more of a connoisseur and wants to try new and tasty foods. His issue is that many times when he tries something new, the food is not good. Many of these experiences means that Pablo is not eating the variety of foods that he would like. Pablo is more influenced by price information, and coupons and sales may be more effective incentives. He already knows what many common foods are, and is much more interested in trying new tasty foods.

Design Process

Phase#1 Idea Divergence and Convergence

Based on the research findings presented above, our team is able to come up with a list of ideas that have the potential to solve the ultimate problem. We hold a brainstorming session, diverging with 40+ ideas, and organize them through Big Idea Vignettes. After converging the number of ideas to about 14, we further vote on the ideas according to two dimensions: feasibility and impact.

Phase#2 Idea Convergence

The feasibility-impact grid generated the top three design ideas: Food Game, AR Menu, Social Network of Food. The three of them solve similar user needs but from different approaches. We then developed low-fi prototypes for the three ideas respectively.

Here is one of the design : AR Menu. The key idea here is to make use of google glass to provide real time information of food. In this case, users won’t be embarrassed from asking questions to the server or not understanding the server.


  • This is the first time for Yan to dine at Tech Square. She has never ordered at a restaurant at Moe’s before, and she has little knowledge on Mexican food overall. Fortunately, she has downloaded the Food Detector App on her Google Glass.
  • On walking in the restaurant, the glass detects the location and surroundings and prompts her to start the app. Feeling the need to learn more about this new restaurant, she chooses the “I am feeling curious” mode. 

  • She looks up at the menu on the wall from a distance, acting as if she is just studying what to order like an acquaintance. However in her glass, she could already see the detailed information of every item of the menu. 

  • On seeing “Homewrecker”, she is confused about what the name means, but the glass tells her the origin of the name, as well as the real flavor of this kind of dish. She also finds out that 4 of her friends has LIKED this food on social media. These knowledge makes Homewrecker look more compelling to her. 

  • She has no idea about the procedure yet. But she doesn’t need to worry, because she could order along with the prompts on her glass. There are detailed steps listed in the second page for the food. She now feels confident enough to walk straight to the window for a fabulous adventure of trying new food. 

Phase#3 Final Product Idea

We collected feedbacks on the three ideas from classmates and teaching staff, and integrated features into one design. The final idea is an app that gamify the process of trying out (and sharing) new food within a social network of friends. It also provides clear information about the menu, food and instructions for the order procedure. The interaction flow is illustrated below in low-fi prototype:

*Flow chart made by Zhonghe Wen

Phase#4 Heuristic evaluation

We conducted a heuristic evaluation with two UX experts on the low-fi prototype. The key feedbacks include that the hierachy with the hamburger menu is not straightforward; there should be positive reactions to friends' recommendations besides the negative reaction. There are also some problems with the logic between pages. Based on these we made changes to the prototype and get ready to make hi-fi prototype on Sketch. We decided on using sea green as our primrary color to make the app looks friendly and health-related. This hue of green also goes well with real food pictures that we will use a lot later in the app.

Phase#5 Hi-fi prototype (inVision)


Action Plan

The two phases of evaluation mentioned above provided much insights into how to improve the current design. We prioritized the issues according to its serverity and impact, and generated this action list for future iterations.


This project is made possible by team Potluck: Phillip Roberts, Mahesh Ramesh Kumar, Zhonghe Wen and Yiran Ma. Every team member contirbuted equally to each research and design activity.

All graphics in this article are made by the author(Yiran Ma) unless specified.


Back to Portfolio   or, try the following links :