Mojo Vision: AR Contact Lens Eye Gesture Activation Case Study

My role: Solo UX Researcher

Timeline: 6 weeks total

• Part 1: November 2021 - 3 weeks

• Part 2: December 2021 - 3 weeks

Stakeholders: Mojo Vision Executives and VPs doing investor demos; Product and Design Team

Methods Used: Semi-structured interviews, eye-tracking, A/B Testing (quasi-experiment)

Context:

Mojo Vision (a start-up)  was creating an Augmented Reality contact lens that would have been interacted with solely through eye-gestures. People’s first impression with anything is crucial and Mojo wants its lens to give a good first impression when users activate it. The design team came up with a way to activate the Mojo lens with an eye gesture. After implementing this eye gesture activation method, the design team noticed that there was a steep learning curve associated with this method. There needed to be research conducted to find a better way to introduce and teach this activation method. 

Business Goals:

• Increase the walk-up usability for demo testers and future lens users

• Figure out the best way explain the eye-gesture in a way that makes demo sessions run smoothly

User Goals:

• Figure out the best way explain the eye-gesture in a way that decreases confusion and frustration, while increasing success activating the interface

Primary Research Questions:

• What is the most effective way of communicating the eye gesture needed to activate the lens?

• How can we decrease confusion and increase success activating the lens during the users’ first 30 minutes of interacting with the interface?

Part 1: What is the best way to explain this eye gesture?

Part 1 Participants:

(Constraint: Can only test with employees due to confidentiality concerns)

This study had 10 participants total. I recruited employees who had not been a part of any past company user testing. I sent out a prescreener screening for age, gender, past VR experience, history of medical eye conditions, and if they wore glasses or contact lenses.

Part 1 Methods:

• Eye-tracking ,  A/B testing, semi structured interviews

I separated the 10 participants into two fairly even groups of five based on age range and past VR experience. I sent out a prescreener to these participants that helped me match up participants by age groups and their familiarity interacting with VR. I assigned one person from each pair to Group A and the other to Group B. 

I gave both groups the same introductory eye-movement test to establish a baseline accuracy in activating the lens. Then, I coached the participant with the method that was specific to their group to try to improve their activation accuracy. These coaching methods represented two different hypotheses of which explanation might be more intuitive. 

Group A: more detailed verbal explanation and a visual description of how to activate the lens 

Group B: short verbal description of how to activate the lens

The study was a mixed-methods study. Quantitative data was gathered through eye tracking and logging participant activation accuracy. Qualitative data was gathered through semi-structured interviews during the middle and end of the session. I analyzed the eye gaze data to see if the participants might have been experiencing eye fatigue, and which method reduced this eye strain. The semi-structured interview in the middle of the study was aimed to understand what confusions the participant still had about activating the lens. The interview at the end of the study was used to gauge overall experience and eye fatigue from a qualitative perspective. 

Part 1 Results:

The explanation Group B received led to a more successful activation experience compared to the explanation Group A received. 

However, the explanation that ended up being the most successful was a modified and refined version of the explanation Group B was receiving. The wording was modified based on the thoughts expressed during the semi-structured interview portion. I listened to the words Group B participants were using to describe the eye-gesture back to me. After the first few sessions, I was getting clear enough data that there were certain words they were using to describe the eye-gesture. I thought we should use their own words to describe it to them. This modified explanation was used for the rest of the participants in Group B to make use of the participant sessions lined up.  For participants in Group B, their success in activating using the eye gesture increased from ~48% to ~86% after the coaching they received. This “refined explanation” is what I went on to test and validate in Part 2. 

Part 2: Let’s validate the “refined explanation”

Participants:

This study had 10 participants total, and these participants were different from Part 1. These participants had never been a part of a user research study before and had no previous familiarity with this activation method. 

Part 2 Methods:

• Eye-tracking, semi structured interviews

All 10 participants received the same initial explanation on how to activate. The wording and nature of this explanation was the result of Part 1 of this study. I told the participant to try activating a certain number of times. They were then given a break and told to try activating the same number of times again. Their accuracy would be calculated on how successfully they activated before and after the break. The design team determined that they would be comfortable considering an 80% accuracy or above to be successful. 

Part 2 Results:

The “refined explanation” created from the research in Part 1 proved to be successful. Seven out of the ten participants in Part 2 met the 80% accuracy (“success threshold”) with just the “refined explanation” alone. These seven participants did not need additional coaching or help. Out of the three participants that did not meet this “success threshold percent”, two of them came very close to meeting them without any coaching. The last participant who didn’t meet the “success threshold” was able to get 100% accuracy after being told the same explanation a second time. 

Impact

The head of User Experience used the “refined explanation” resulting from this research during demos with first time interface testers. My research was used in investor demonstrations to improve their experience interacting with the Mojo Lens interface for the first time.

Challenges

1. Balancing trying to gather more data versus participant fatigue: Wearing a VR device is fatiguing especially when working with repeated, specific eye movement. I prioritized participant fatigue, while making sure I had just enough eye tracking data to analyze. 

2. Creating visuals for the whole company to understand: I presented these findings company-wide and in doing so I had to package the findings in a way that was succinct and easy to understand visually

   
   

Reflection

I learned a lot about research study design and interviewing during this Activation study. This was the first research study I did at Mojo Vision, and the first time I got to practice the skill of asking follow-up questions during an interview. 

Because it was my first company research study, there are definitely things I would have changed in hindsight. I would have benefitted from running a pilot session to test out the whole flow of the sessions and the wording of my semi-structured interview portion. In Part 1, my initial explanation of the activation method varied slightly from person to person. I should have been more rigid and standardized with this wording since I was aiming to construct a baseline.