Putting virtual makeup to a real world test
Ulta Beauty · GLAMlab design and usability research
I helped design Ulta Beauty’s augmented reality makeup experience, then researched how customers actually used it. Every piece of qualitative feedback entered the development roadmap, and the improvements were implemented across multiple releases.
My role: UX design collaboration, study planning, participant recruitment, test scripting, synthesis, and presenting findings to the team and leadership.
The poster got the makeover
One participant was trying to apply virtual makeup to her own face. GLAMlab kept applying it to the face of an athlete in a poster behind her instead.
“This is pretty good,” she said.
It was entertaining. It was also a useful reminder that what people say about an experience can be very different from what we see them doing. A satisfaction survey would have missed quite a bit of that story.
An experience that had to earn trust
GLAMlab launched within Ulta Beauty’s mobile app in 2016. I worked with a team of UX and UI designers to make the experience realistic, accurate, and easy to shop. We had studied other AR experiences, but testing our own with customers was new territory for us.
We wanted to understand whether people could find GLAMlab, use it comfortably, and trust what they saw. Did the makeup land where it should? Did the color, coverage, and finish seem right? Would someone feel confident enough to buy a product—or return to try something else?
Make room for different faces
I set up remote, unmoderated studies in UserTesting, recruiting from a customer panel I had previously built from our customer database. Screeners helped us include participants across a variety of skin tones, face shapes, and eye shapes.
The tasks left room for people to make their own choices: download the app, find the experience, and try a product in a color they liked. Rating questions captured their perceptions of ease, confidence, and accuracy. The recordings let us see what was happening while they answered.
Watch together and compare notes
As sessions came in, I monitored for technical problems and confusing instructions. Once the recordings were collected, I held watch parties with the UI designer and product owner. We watched, took notes, and worked through the findings together.
I then prepared a readout for the broader team and leadership, connecting what we had observed to opportunities for improvement. Watching someone struggle with a design you helped create is uncomfortable. It is also a very good reason to keep watching.
What the research changed
The study revealed differences in usability across skin tones, face shapes, and eye shapes. Recruiting a variety of participants helped us see problems we could have missed with a narrower group.
The poster incident made the value of observation especially clear. The participant was working around a conspicuous problem while still describing the experience positively. Listening mattered; so did paying attention to what she was tolerating.
The findings also prompted the team to define measures for future AR research, giving us a basis for benchmarking the experience as it evolved.
A roadmap followed by releases
The development team added every piece of qualitative feedback to the roadmap. Those improvements were eventually implemented across multiple releases.
The study became part of an ongoing cycle of research and refinement. As GLAMlab grew, customer feedback continued to guide the experience. Its usefulness took on a different significance during COVID, when people needed ways to explore beauty products from home.
The outcome I can trace directly to this work is the path from observed problems to implemented improvements. A convincing virtual lipstick needed more than the right shade. It needed to reach the right face.
A look inside GLAMlab
The earlier screens below show the choice of a selfie or model and facial mapping in the 2016 experience. The later interface, preserved in my 2020 manuscript, shows product and shade selection with a path to purchase.