Retail dashboard study

Project type: Group
Timeline: One semester

Companies have been using customer data to improve marketing and operations for many years, but rarely does customer data serve to improve the customer's own experience.

During the spring 2014 semester, I worked with Anna Turner from Carnegie Mellon's master's program in HCI under the guidance of Profs. John Zimmerman and Jodi Forlizzi to explore ways that companies could use customer data to improve the customer experience. We wanted to be able to personalize customer-employee interactions by providing customer loyalty data to employees.

Our ultimate goal was to design a dashboard for use by customer-facing employees. Using this data dashboard, employees could alter their delivery of the service based on each customer's profile.


Phase 1: Context Exploration

We started out by observing customer behavior in malls and food service locations, as well as interviewing industry experts. I spoke with the head of customer experience at the Columbus, Ohio public library system, as well as expert service practitioners, like a personal shopper at a Charleston, SC, J. Crew store. These interviews gave us valuable insight into what experienced customer-facing employees do on their own to improve customer experience.

Phase 2: Concept Development

Anna and I quickly jumped into creating concepts based off our early research. We were trying to push the boundaries of what customers and employees might find acceptable in terms of privacy, technology use, and other social boundaries.

One of my favorite user research methods for testing new concepts is Speed Dating. Anna and I created storyboards of two to three panels, with each storyboard probing the participant for recognized needs. That is to say, we wanted to know if the participant felt that he or she had the need we were trying to solve for. We conducted Speed Dating rounds with multiple people at once in order to generate discussion around the concepts. However, we conducted testing sessions with customer populations separately from sessions with employee participants.

Our concepts came up against a lot of resistance from both groups--customers and employees. Although we knew that customers and employees are dissatisfied with the status quo, our scenarios didn't strike the participants as identifying the specific needs they felt. Our customer participants couldn't see much personal benefit in the scenarios we presented, and the employee participants were suspicious of them as well. Therefore, the reactions ranged from lackluster to irritated.

Phase 3: Concept refinement

Moving forward, we focused on designing for a type of customer experience that would allow us to push boundaries even further, while alleviating customer stress: the travel experience. Because airline customers are already used to disclosing personal information to the airline, as well as submitting to x-rays and following strict baggage rules during travel, we felt that we could get a better sense of future attitudes toward technology use if we situated our scenarios at the airport. Many customers find the airport experience quite stressful as well, so we also felt that we could make a big impact on customers by improving that experience.

Conducting "speed dating" sessions with a few interviewees.

Conducting "speed dating" sessions with a few interviewees.

Phase 4: Designing the dashboards

Finally, I drew some dashboard "wireframes" to be used by airport ticketing agents in the scenarios we had tested via Speed Dating. The most important factor was providing only the information most necessary to the scenario at hand; any additional information takes time and effort from the agent, and increases the amount of time he or she spends with his/her eyes on the screen instead of on the customer. My first wireframes were made on paper with markers. Then I refined the visual design in Adobe Illustrator.

A digital iteration of the information dashboard that a ticketing agent might use at an airport

Giving the employee an opportunity to collect data from the customer and enter it passively during the conversation was another goal we incorporated into the dashboard design. In the example below, previous employee-given suggestions are displayed with a thumbs up and thumbs down, so that the ticketing agent can ask questions like "How did you enjoy Toronto Island last year?", thereby improving the ability of the system to provide recommendations and gather data on the customer's preferences.

The final airport version dashboard.


Though we conducted a full cycle of research and design, I felt like Anna and I had barely scratched the surface of this socially and technologically complicated topic. Developing technologically feasible, socially acceptable, human-centered service features is difficult because it's hard to prototype and test these features with customers and employees both. Nonetheless, I believe that design will be increasingly involved in interpersonal interactions and relationships as well as single-user-to-device relationships over the next few years.

I'll be carrying on much of this research in my master's thesis during the 2014-15 school year, where I'll focus on personalization, service recovery, and customer-employee interaction in the travel context.