Google - MHCI Capstone

Aria Opus

Over 7 months of user research and design, my team, Aria, designed a suite of technologies that augment today's instore shopping experience.

  • User Experience
  • User Research
  • Google Shopping
  • In-Store Retail
  • Service

MHCI Capstone Project
Team: Jonathan Brown, Eugenia Lee, Ramya Mallya, Pavel Samsonov, April Yang
Client: Google Inc.
Spring and Summer 2014 Carnegie Mellon University
Role: Team Co-Lead and Technology Lead
Click here to view the project website.

Project Scope:

The MHCI Capstone Project is a 32 week long user experience research and design project. My team, Aria, designed an experience for our client, Google. We aimed to use Google Shopping data to empower retail associates and help them guide their customers towards well-made-choices.

This project ran through spring and summer semesters. In the spring semester, my team completed the research phase of the project. In the summer semester, we used this research to inform our process as we designed We are now entering the design phase of our project, where we iteratively design, develop and test prototypes of increasing fidelity.

My Role:

While everyone needs to wear multiple hats over the course of the project, every team member has primary roles. My roles are Team Co-Lead and Technology Lead.

As Team Co-Lead, I am responsible for ensuring that our team works smoothly together, and that the quality of the work we produce is excellent. I need to make sure we achieve this, in such a way that all members of my team also achieve their individual learning goals. I provide feedback on design and process and make sure that our meetings and brainstorming sessions stay positive and creative. I manage task division, internal communication and scheduling.

Over the spring semester, my responsibilities as Technology Lead were limited to designing and developing our Spring Website. Hence, I was also actively involved in activities outside my formal role, such as executing our research plan and designing our report book. Over the summer I have been actively building prototypes. I helped build a low fidelity prototyope out of cardboard, built a mid-fidelity prototype using Axure, and was the primary programmer in building our high fidelity prototype using HTML, CSS and Javascript. I wrote testing protocols and conducted usability tests, playing the Wizard of Oz in low fidelity and mid fidelity usability tests. I took part in design meetings and worked with my team to design the interactions for our double sided countertop.

Exploring the problem space:

The spring semester consisted of extensive domain research and field research. Our domain research consisted of literature review, territory analyses and interviews with domain experts and retail executives. Our field research consisted of Guerrilla observations, contextual interviews and diary studies. We visited a total of 42 stores over the course of our field research phase, and spoke to 37 sales associates and 16 customers. We used a combination of Guerrilla research, contextual inquiry and diary studies to gather field observations.

Making Sense of our observations:

We consolidated our research by building flow models, sequence models and affinity diagramming. Using this process, we identified the key breakdowns and design opportunities that we shaped our solution around.

The Opus Experience:

We designed a unique way to take advantage of all the above opportunity spaces to produce a single experience.

Introducing the Opus Countertop:

We focussed on the Opus Collborative Countertop during our prototyping phase. We tested our design with shoppers and associates over three rounds of testing, using prototypes of increasing fidelity.

During low fidelity testing, we built prototypes made of cardboard to test the physical form of the countertop and the arrangement of the associate and shopper as they used it. Results from our tests indicated that the optimum configuration would be to have the shopper and associate stand across the counter from each other. This meant that the interface would need to have two orientations in order to be legible by both, the associate and the shopper. This proposed several unique interaction design challenges.

In our medium fidelity testing phase, we tested a the flow of a typical shopping scenario using an interactive Axure prototype and the Wizard-of-Oz technique on a 39" TV screen. Based on the results of this round of testing, we decided to minimize tap interactions and make drag-and-drop, and swipe the primary actions. We decided to also re-organize the layout.

In the high fidelity phase, we built an interactive web prototype using HTML, CSS and Javascript. We tested it using a 55" multitouch display, which we took into an actual department store.