Duolingo Feature Design

Personal Project // Fall 2018

Duolingo provides free language education so everyone has a chance to learn a new language in a fun, engaging way. However, currently there is no way for users to practice hand-written characters or stroke order — a key part of learning any character based language like Chinese, Japanese, or Korean.

My Role

Interaction Designer, Animator, Visual Design, User Interviews, Persona Development, Journey Mapping

Project Length

3 Weeks. Fall 2018.

Design Overview

Proposed Feature: Practise Character Writing

Currently, none of the character based languages include practice for character writing; learning how to write characters by hand and memorizing the order that you draw the strokes are key components to learning Japanese, Chinese, Korean, or Farsi. There are many character based languages, so focusing on a single language as a case study was the most feasible approach for this project. In our solution, we chose Japanese.

If this feature is successful, it would be scalable to all character based languages.

Lesson Plan Before Our Design:

Why Duolingo?

I was inspired to do a project for Duolingo after meeting several passionate users of the app, one of which went on to live in several Spanish-speaking countries using her near fluent Spanish skills — so, in truth, one of my goals was to simply sate my curiousity. However, our ultimate goal was to revamp a key feature, or possibly add a new one. This passion project aimed to improve the experience for existing users in a feasible, yet meaningful way, specifically by addressing frictions we identified through online research and testing with 3 users: a beginner, an intermediate, and an expert user of the app.

Key Findings

Something is Missing

Duolingo provides services to both individual learners and classrooms, we decided to focus on the more variable user: the individual. We started our research by scouring the Duolingo forums. From that, we discovered a reoccuring issue: there is no way to practise writing character based languages.

Key Frictions and Empathy Map

From our testing and interviews, there were 3 key frictions: 1) Unable to control lesson plans and some settings within the app (must go to desktop version), 2) Unable to match what they learn in class with what they learn in the app, and 3) Frustrating for users with learning disabilities (ex. our dyslexic user can't easily / didn't know they can adjust accessibilty settings). Although we were unable to address all of their frictions, we did use them to inform how we went about designing our new feature.

Getting the Design Right

Scenario: Alana, an intermediate Duolingo user, is riding the bus to school, and cramming for her kanji quiz. The bus is crowded and she couldn’t get a seat, so she can’t pull out her notebook to study or practice writing the characters.

In our design, learning to write kanji is seamlessly integrated into the already established lessons. The lesson structure is the same, but now our users are also asked to write the kanji by hand and in the right stroke order.

By utilizing swipe gestures they are able to practice kanji even with only one hand while standing on a crowded bus, a frequent setting all of our users used the app in. This also works for writing simple hiragana and katakana.

Try out the protoype

Visual Design Patterns

Type Treatment

The UNDO and CLEAR buttons follow the Duolingo typography standard ‘Secondary Button’, as it is a non-primary button. They also allow users to have more control of the lesson, letting them redo the lesson until they are satisfied with their result. This was meant to relieve some of the friction identified earlier where our users felt they lacked control of the app.

Colour Cues

Duolingo uses the color ‘Banana’ to introduce new concepts, so we applied it to the teaching animations. This also puts the kanji strokes first in the visual hierarchy.

Interaction Patterns

Swipe Gestures

Swipe interactions were designed for within the lessons. To stay true to Duolingo’s teaching style, the interaction has been implicitly, but not explicitly, explained to allow Alana to discover how to complete this question types.

Double Checking Your Work

Alana can check the simplified version of the kanji by tapping the underlined english translation. This interaction pattern was repurposed from the existing lessons to emphasize what the kanji she is learning means alongside learning how to correctly write it.


Following our new feature design, we re-interviewed our 3 users and had them test out the prototype. In the end, 3 out of the 3 users were genuinely excited about our new feature. One of the key responses for me was, "I'm surprised this doesn't already exist. It just makes so much sense!".

Although there are several ways to expand on this project, including creating versions of the lesson without any guides to aid with memorization, and possibly designing a seperate section within the app dedicated to practising writing, this was a successful proof of concept. I would love to share this project with Duolingo, and revisit this idea in a more actionable way in the future, but for now this project was a lovely exercise in following an interesting lead, and making something new.

Lesson Plan After Our Design: