UX/UI Designer
2 weeks
Case study
Duolingo’s goal is to make language learning fun, accessible, and rewarding. However, learning a language is difficult to do in isolation. Duolingo already allows its users to make friends and participate in leaderboards, but I felt that its social capabilities could be enhanced to make learning more collaborative.
I wanted to explore designing an integrated chat - complete with a chatbot - as well as other improvements to the ways users can interact with each other.
The research for this project consisted of looking at market trends, evaluating competitor’s strengths and weaknesses, and conducting a survey and one-on-one interviews with Duolingo users and multilingual individuals.
My competitor analysis focused on branding, UI, and main features of other language learning apps, as well as the strengths and weaknesses of a couple of popular chatbots.
of users were interested in talking with somebody fluent in their target language.
of users watch videos or listen to music or podcasts in order to learn the language outside of the app.
of users do not use Duolingo's Friend or Leaderboard features.
of users find practicing common conversations the best way to learn, while the other half prefer learning grammar and vocabulary.
Practicing with a speaker of the target language helps people become braver and recall words that would express their thoughts.
Media like movies and songs help to become familiar with the sounds and accents of the language.
Topics related to areas of interest ties the language to other useful applications.
Competitive environment and ability to see progress are excellent motivators.
Variety of exercises keeps learning interesting and fresh.
Theory-based courses don’t help with being able to put the language to use.
Intimidating to have to use the language with a native speaker.
Lack of options to practice industry-specific terms.
Conversation structure is hard to grasp without a lot of practice.
Based on the interview results, I created a persona as a way to summarize common user goals and frustrations so I could keep them in mind throughout the project.
I also began thinking about the Information Architecture of my features, and how they would fit into the app. The site map helped narrow down what I wanted to include, while the user flow gave me a more concrete idea of how certain tasks could be performed.
The core challenge of this project was to make sure that any new or existing design patterns I use blend in seamlessly into the existing application. Duolingo has a very identifiable brand and descriptive design guidelines, which I referenced constantly throughout my process.
Along with the new chatbot and friend chat features, I added the Gift section to the Shop and two new achievements to the Achievement page.
Since I was adding a small number of new features, I sketched out a couple of screens for each one. I wanted to brainstorm how chatting with Duo or a friend would look, and note down minor additions like new achievements and shop items.
Because Duolingo has a very clean interface, my wireframes were pretty close to the final look. This stage helped me work out spacing, navigation, and design patterns for the chat.
Duolingo has a very distinctive brand - the design guidelines had a lot of details about exact spacing, shades, and font styles. My style tile was a compliation of these elements.
The UI Kit grouped Duolingo's style direction with the design patterns used in the application. I ensured that the buttons, text boxes, and other elements adhered to Duolingo's design guidelines.
Blending new features into an existing app was complex, but through a combination of following the design guidelines and comparing my screens with existing elements, I felt my project was complete enough to put it to the test in front of some discerning users.
Users were given the following tasks:
All users were happy with the final design, with only minor critiques relating more to personal preference. In order to make modifications with any confidence, I would conduct usability testing on a larger group, possibly with A/B prototypes.
Reporting function, additional safety features for chatting with others.
Matching with fluent language speakers, ability to indicate user's own fluent languages.
More chat variety: different error states, career-related topics.
