Duolingo Case Study: Introducing Study Partners

Harini Kannan
Jul 20, 2017 · 10 min read

The goal of this case study was to propose, ideate and create a new feature for Duolingo, a popular language learning-gaming application. Through market and product analysis and a thorough UX Design process, I explored what feature introduction would address user and business needs.

Note: This case study was done around September 2016 before new features of similar ideas rolled out.

Disclaimer: This was my first project in Product Design. The way this is ordered and structured here is reflective of the process I took when I first started. It is now July 2017 and through project-based classes, research and my internship as a UX Designer at Blackstone, I have grown in the way I approach the product design process. Along the way in italics, I suggest changes in my approach if I were doing this project again now.


Idea

We want to provide a “study-buddy” interface in the context of a language-specific, level- specific community, in order to have two learners participate in level-appropriate study sessions about the material. This simulates similar learning and studying patterns of students in high school and college, rendering a more engaging learning environment for Duolingo users.

Objectives

  • Bolster the educational value of Duolingo by facilitating and aiding a user’s learning experience.
  • Retain users by engaging users in social interactivity and further motivating users to complete lessons
  • Strengthen Duolingo network by fostering a communal learning experience

Desired Impact

Objective 1

If Duolingo were to increase the educational value of its app, the desired impact would be improving upon Duolingo’s values as an education provider and creating a better user experience for the students. The integrity and ethos of Duolingo as the leading educational app in the market would be furthered by integrating more ways of productive learning.

The success of this objective is measured by the user’s success rate of proceeding to higher levels.

Objective 2

If Duolingo retains users by motivating them through engaging learning interactivity, then the desired impact is that users will feel compelled to continue through new lessons (by the success rate of completing previous lessons) and feel committed to finishing their language. This commitment and motivation prevents a user’s decreased usage of the app or dropping of the app altogether.

The success of this objective is measured by the usage of the app. If there is no decrease in usage, or the user hasn’t stayed on the same level for over a week, then those are indications that the user is engaged with the applications. This is especially effective if Duolingo obtains data about the user’s past usage (how often a user completed a level or logged into the app).

Objective 3

If Duolingo is able to create a strong network of interacting users through any sort of communication between users, then that will increase commercial visibility to prospective students. Providing novel sources of learning will be ways of appealing to new users.

The success of this objective is measured by any sort of increase in the users of the application. The measurement is most effective if the new users are utilizing the communication functionality.


Reflections and Changes:

I pause here to comment on my design process thus far. Over the course learning and practicing product design, I have learned that you never start with a solution. Starting with a solution presupposes what the user wants before understanding what the user needs. As a designer, you create visions not for the product, but for the users. And to do so, you must do extensive research about where they feel the product disservices them.

However, understanding the overall objectives of the application and the return on investment (ROI) of each objective is important in narrowing where the product can enhance and what aspect coincides with business needs.

If I were to do this again, I would start with the objectives and ROI and then conduct user research and interviews (while looking at relevant activity statistics and data). From this research, I would synthesize the data (through affinity diagrams or similar methods) to narrow down on problem areas within selected parts of the product. From there, I would create personas and user stories to further solidify my understanding of user needs.


User Profiles

User Profile 1

Jaelyn Moore is a 20 year old student at Cornell University studying Linguistics and Art History. She took four years of French in high school, but chose not to continue in college because she wanted to begin learning Japanese. However, Jaelyn has not lost her interest in French and wants to keep afresh with the language, as to not lose her skills. She downloaded Duolingo in order for her to get a casual refresher in the basics. She is quite quickly making her way through the levels seeing that her pace quickens as her familiarity grows with the material. Often she is going through the levels when she is by herself, taking a break from studying, or walking to class.

However, Jaelyn loses her ease when she is at the more advanced levels. She has to spend more time on each level in order to really understand the information. As she attempts to do the harder levels, she does not isolate herself anymore. She finds public spaces to work in, often talking to her friends (who are not French speakers) about the levels. She does not know any native French speakers nor does she know anyone currently learning French. She pinged her high school friends who learned French with her, but they had forgotten much of the material that she infact had already gone through in previous levels. She wants to keep working on the harder levels, but with homework and extracurriculars it is hard for her to dedicate the time. She has not completed her current level which she has been on for two weeks, irregularly going into the app.

User Profile 2

Amanda Garcia is a 26 year old Associate Manager at Verizon. She lives in Basking Ridge, New Jersey, and often visits the city to visit her friends. Amanda is a team player, often learning and working with other people since she was in High School. She found the best way for her to study was in groups of one or two people studying the exact same material at the time she was (ex. Both her and her friend would be reviewing the same chapter or same concept). Even at work, she often bounces ideas off of one or two engineers and designers as she is brainstorming new features for her current project.

Amanda has a passion for gymnastics and was even on the gymnastics team during her time at the University of Florida. She really enjoys watching gymnastics as well, and follows many of the famous olympic gymnasts. Her favorite gymnast is Aliya Mustafana from Russia. As she followed Aliya and the rest of the Russian team closely during the Rio Olympics, she became deeply interested in Russian culture, and wanted to learn the language as a serious endeavor. Upon her friend’s advice, she downloaded Duolingo and began the lessons. Amanda is bilingual, fluent in both English and Spanish, but she still struggles with Russian given its different linguistic structure. Even with the basic levels, she takes some time to learn get through the lessons. She is attempting to get her friends to download the app and join her in learning Russian, but the ones interested in language learning downloaded the app to learn a different language. Amanda continues to get through the levels, but wishes she had someone else to work with her.

User Goals

User Goal 1

To be able to get through each lesson understanding and retaining the information learned.

User Goal 2

To obtain a thorough learning experience that does not intrude (in terms of time and effort) on daily life.

User Goal 3

To utilize resources to increase the speed, accuracy, and understanding of the material.

User Tasks

User Task 1

To finish a level in one time-manageable setting.

User Task 2

To contact someone for help when he or she is struggling with a lesson.


Reflections and Changes:

While user profiles (personas), user goals and user tasks are all pertinent aspects of the product design process, they are misplaced in this part of the process. As stated before, before creating personas and defining user goals, it is important to understand what the users and what the business defines as problem or feature-lacking areas of the product.

While the user profiles, user goals and user tasks are attempting to address the problem of “lack of person-to-person communication system”, they are following a solution rather than following a grounded and verified problem. That aside, I think the user profile effectively creates a story of a user experiencing a problem that could be solved with a feature introduction. Additionally, the user goals and user tasks appropriately draw upon what the user desires from their product experience.


Features and Functionality

Feature Summary

This feature would be an interactive chat feature facilitated through a messenger-like app. The purpose of the feature is to enhance learning experience by creating spaces for co-user interactivity.

Round 1 : Single Chat Feature

In the first round of designs (with just low fidelity mockups), the chat feature is designed particularly for users who dedicate more time each day for the app.

Low Fidelity Mockups

Round 2 — High Fidelity Mockups

In the second round of designs, I revise the feature by bifurcating the functionality between low-intensity high-intensity users.

Study Sub-Feature

  • Converse with a known friend from a study network
  • Long term conversation/study session to hone in on specific skill
  • Geared towards 30XP and 50XP users (highest user-identified intensity levels)

Add people you know to a network

  • Inspired by “leaderboard” feature
  • Allows users to choose people they feel comfortable conversing with for long periods of time

Accessibility

  • Only in the layer for a specific skill
  • Studying should be relevant to the ”unit” user is on
  • Can not utilize feature in previously completed tasks
  • Can only access people within your network
  • Time constraint — 30 min

Language Usage Reward

  • At the end of each conversation, calculate the percentage of total words used in the language being learned
  • Each skill level will be assigned a percentage goal of amount of learned language should be incorporated in the conversation
  • If the users cumulatively make that goal, it will count towards daily progress

Audio Communication:

  • Just like in a lesson, audio “tests” can be sent from one user to the other
  • One user can send this test to the other user to record
  • Can only be 1 sentence to record
  • Must be in the language being learned
  • Can only send 2 audios in the entire session

Help Sub-Feature

  • Randomly match with someone in the same lesson to ask a question
  • Accessible and quick question asking tool
  • Geared towards 1XP, 10XP and 20XP users (lowest user-identified intensity levels)

Accessibility

  • Must activate availability (either request or provide help)
  • Located only within the lesson
  • Targeted for those who need quick help or can provide it (chat is optional)
  • Get randomly matched with anyone online on the same lesson, native language
  • This type of user has no preference for partner

Details

  • Can toggle in and out of chat
  • Can mute or prematurely end chat
  • Notifications for chat bubble will be in different color
  • Time — Each chat is capped at 3 minutes
  • Communication limited to text

Reflections and Changes:

Upon examining the two rounds of designs, I realize my design took very little consideration into both the optimal user experience and technical feasibility. The chat feature itself overcomplicates a problem that is not even verified in legitimacy and takes multiple steps and multiple minutes for the user to receive help for a question that may take upward of 1–2 minutes for the user to solve. For either feature, the access points are unintuitive and the learning curve for understanding the feature is too steep and can be jarringly complicated for the user. The design introduces too many different sub-features and functionalities which are inconsistent with the rest of the functionality. The accessibility of the chat feature hinders the user’s ability to adapt to it because it drills down all the way to a specific level.

Essentially from a user perspective, the chat feature forces the user into a long-winded workflow that belies the original intention of the game- a fast and easy way to learn a language within 10 minutes or less everyday.

From a technical perspective, there are a lot of constraints, especially with the help sub-feature, with matching people randomly. At any given moment, thousands of users for a specific language on a specific lesson can be online- random matching at that scope can be technically infeasible or remarkably slow on the system.


Conclusion

While this project has many points on which it can be improved, the overall goal was to get experience and exposure to the product design process which I succeeded in obtaining. Through a year of research, side projects and work experience, I feel confident in my critique of this project and hope to grow further as a product designer.

)

Harini Kannan

Written by

Aspiring Product Manager & UX Researcher

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade