I know one thing for sure--the problem is NOT the Duolingo TikTok account. Duo the owl, Duolingo’s spunky and sometimes slightly unhinged mascot, has taken the app by storm, much to the entertainment and delight of TikTok users.
But aside from Duo’s antics on TikTok, Duolingo is also known as one of the most popular language learning apps in the world. While the app is thorough and research has proven it to be a useful tool, one thing it lacks is that there isn’t a way to easily look back on vocabulary, phrases, and concepts that users have learned in past units.
Adding a feature that allows users to review what they’ve marked for review in the app would allow for further personalization of the learning experience, which has been found to be an important part of language learning.
However, much like the unpredictable nature of Duolingo’s TikTok account, my research led me to a new and unexpected direction.
I wanted to gain a better understanding of how users would engage with a feature that allows them to review past units, vocabulary, and concepts without having to complete each unit.
+ Market research
+ Competitor analysis
+ Customer interviews
“Overall, I like Duolingo because it’s easy to use and accessible. I also like that it’s free.”
Duolingo user during customer interviews
I interviewed 4 Duolingo users on their experiences with the app, thoughts on e-learning, and language learning preferences, and that's when I learned that perhaps I wasn't addressing the right problem.
Fortunately, I was able to come to this realization early on because I did my best to avoid leading questions and check my biases. Check out my research plan here. You can view my notes from the interviews below; I've categorized the responses and indicated patterns with the dashed line.
.png)
These interviews revealed another challenge that was more pertinent for users--namely, that users learned best by listening to or speaking with native speakers of the languages they’re learning.
There wasn’t really any mention of wanting an easier way to look back on past vocabulary and concepts like I had initially hypothesized.
I recognized that this was an important and humbling lesson to learn in the UX research process, and was intrigued by this new problem: connecting language learners with fluent speakers.
.png)
Initially, my idea was to re-work the dialogue in the existing “Stories” feature and create a different but similar set of dialogue for a new “Speaking” feature. "Speaking" would mimic real conversations and prompt learners to translate into the mic. View the task flow and user flow on Whimsical for visual reference.
However, after my mentor and I reviewed my first prototype, he challenged me to evaluate whether or not I had truly addressed my users’ need of connecting with fluent speakers. I decided the answer was "no."
After that conversation, it was back to the drawing board! And that’s when version 2 came together. This version integrates voice and video calls so that language learners can connect directly to fluent speakers about an established theme (like the supermarket as seen in the wireframes).
“I like to be in the culture of the language I’m learning as much as possible. Going to markets was really helpful for listening and speaking practice.”
Duolingo user during customer interviews
While I was pretty confident that version 2 was the stronger design and better response to the problem, I decided to conduct usability tests via Zoom for both versions, just to make sure I wasn’t jumping to any conclusions. Each participant only saw one version of the design and wasn’t told about the other version.
After evaluating the responses from the 6 usability tests (3 for each version), version 2 came out as the best method forward. If I had more time, I would have aimed to recruit additional participants to gather more definitive data.
.png)
Firstly, one of the v1 participants expressed that she would have preferred talking to live speakers and felt limited in only translating prompts. While with v2, most of the users commented on the value of connecting with fluent speakers. Their challenges and critiques were really about improving the trust and quality of the feature, rather than seeking an entirely different solution. (View affinity map in Whimsical.)
Secondly, I reflected on the problem again as well as the original user interviews from the discovery phase. All of the interviewees had indicated that their most effective way of learning a language is by connecting with fluent speakers in some way.
Check out version 1 here. Which version is a better learning tool for you? (The version 2 prototype is in the next section.)
This project was not only a way to practice conducting interviews and build wireframes, but it was also a reminder that checking my biases and making room for feedback are skills that also require practice and continuous refinement.
These skills are just as integral to the design process as the deliverables themselves. While I was aware of this from the start, as an emerging designer, I think it was important for me to experience it firsthand and reflect on the significant impact they had on my final design.
