Improving design through behavioral research

Imagine a company where any employee can provide feedback or suggestions for another, from a new hire all the way up to the CEO. A team of product developers, designers and data scientists at Bridgewater Associates, a hedge fund based in Westport, Connecticut, is working to do just that: creating artificial intelligence tools to help companies foster and operate within such a culture.

"Usually, people who are the loudest—either vocally or because of their position in the company—tend to be heard, and their ideas get to the top," says Jennifer Romano-Bergstrom, PhD, director of user experience (UX) research at Bridgewater. "But in an idea meritocracy, everyone’s opinions are weighted on believability and other qualitative traits—not based on title or age."

Romano-Bergstrom joined Bridgewater’s Systemized Intelligence Lab last year to work as part of a team of developers, designers, data scientists and the company’s leading business strategists to develop tools to support an idea meritocracy, which Bridgewater can use to improve workplace culture and exchange feedback among its 1,500 employees as well as sell to other companies.

She is the first UX researcher at Bridgewater, the world’s largest and oldest hedge fund. In addition to investing, Bridge­water designs products such as Dot Connector, a software program that helps team members view and better understand the opinions, strengths and weaknesses of their colleagues.

"My psychology background comes in when I design experiments to test whether people understand our products, and then if they don’t, to understand why," Romano-Bergstrom says. "I’ll determine what kind of study we need to conduct and how best to conduct it. Anybody can ask questions, but can they ask them well?"

Fostering teamwork

UX and usability researchers aim to answer two basic questions: How do users interact with a product, and how can that product’s design be improved to enhance the overall user experience? A fundamental understanding of human behavior and experimental design allows researchers like Romano-Bergstrom to answer those questions systematically.

The goal of the idea meritocracy product is to create an artificially intelligent system that learns the strengths, weaknesses, goals and interests of employees based on past projects they have worked on, their performance assessments and feedback provided by their colleagues. The system then uses that information to make personalized suggestions for every member of the organization. For example, it might recommend readings based on a person’s strengths or weaknesses. Or it could advise someone on how to think about a situation differently. Based on information about an employee’s personality characteristics and past work, the system might remind the employee that making mistakes is OK, as long as one learns from them, and then challenge the employee to explore what he or she learned from a particular mistake.

Romano-Bergstrom’s job is to collect, sort and prioritize questions and hypotheses about the idea meritocracy tools, design and carry out experiments, and use the results to modify the product’s design before it’s ready for market. If the suggestions are too conceptual, for instance, Romano-Bergstrom circles back with the product team to brainstorm ways to incorporate more concrete advice into the tools.

"It’s important to make sure people understand how to use the product and that it’s matching their expectations. That’s where my research comes in," she says.

Romano-Bergstrom also ensures that data—such as the number and timing of clicks and errors users make when testing a software product—are collected with precision. She designs surveys, eliminating or rephrasing questions that are leading, biased or out of order. When recruiting participants, she ensures that sample sizes are large enough and demographic groups of interest are included.

She also teaches designers about psychological concepts such as cognitive load and working memory. If designers understand that human working memory capacity is limited, they can create user interfaces that require fewer mental resources—ones with familiar layouts and less visual clutter.

Romano-Bergstrom learns a lot by observing users and gathering their opinions, but some aspects of UX research are less explicit. Her expertise in biometrics, especially eye-tracking, allows her to study these implicit responses. She uses eye-tracking technology to map where people look when they’re performing a task, how long they attend to various images and the order in which they process information. Other tools for UX research such as heart-rate monitors and facial-recognition systems allow her to monitor pulse and facial expressions, which are indicators of frustration and other emotions.

In addition to working closely with the product design team, Romano-Bergstrom collaborates with senior-level strategists at Bridgewater. Through such interactions, she learns about the company’s vision and objectives, which can be at odds with the results of her UX research. When this happens, she works to find solutions that will both serve users and accomplish the company’s goals.

"Fundamentally, product design teams are serving people," says Hans Anderson, the product design director at Bridgewater, who hired Romano-Bergstrom. "In order to create value for our users, we need products that are both well-formed and intuitive. Jen helps us do that in an objective, evidence-based way."

Experience for the job

Romano-Bergstrom first entered the world of UX design through an internship in the Human Factors and Usability Lab at the U.S. Census Bureau. At the time, she was pursuing a PhD in applied experimental psychology at the Catholic University of America in Washington, D.C., researching cognitive aging.

Cell phone as an example of UX psychologists bringing an understanding of behavior to product design.

While there she studied how people interact with census forms and websites. She developed expertise in eye-tracking and other biometric tests, learning how UX researchers use such tools to quantify participants’ physiological responses. And she found that she thrived on an inter­professional team of sociologists, developers, designers and survey methodologists.

"Working with an interdisciplinary team allowed me to interface with people who approach problems from different angles," she says.

She later joined Fors Marsh Group, where she formed and led a UX research team, before moving to Facebook, where she helped design tools like Safety Check, which allows users to "check in safe" during a crisis and can even facilitate delivery of resources to affected areas in a disaster. She came to Bridgewater in 2018.

Romano-Bergstrom has also maintained partnerships with academics, including co-authoring journal articles and textbooks on usability and eye-tracking. Through such partnerships, she has continued her research on cognitive aging, which helps her improve the usability of products for older adults. For example, she found that older users spend less time looking at the periphery of web pages than younger users do (Universal Access in the Information Societyopens in new window, Vol. 15, No. 2, 2016). She also collaborates with academics on privacy-related work.

Her doctoral training and ongoing contributions to peer-reviewed research give her an edge in her UX research.

"It’s really important to me to help bridge the worlds of academia and industry," she says, adding that she regularly organizes panel discussions, talks and workshops to that end. "The benefits go in both directions. Academics get to see firsthand the problems industry practitioners are trying to solve, while industry folks benefit from the wealth of knowledge academics have and the rigor they bring to their work."

Usability Testing for Survey Research
Geisen, E., & Romano-Bergstrom, J. Morgan Kaufmann, 2017

The UX Careers Handbook
Lebson, C. CRC, Press, 2016

How Are Online Health Messages Processed? Using Eye Tracking to Predict Recall of Information in Younger and Older Adultsopens in new window
Bol, N., et al., Journal of Health Communication, 2016