Language App
Retaining Users by Serving Advanced Learners
Problems
Less than 5% of language learners reach the upper levels of popular language learning apps, and claims of practical language skill gains from app learning alone are questionable.
Solutions
Low-wait language lobbies where users can converse in-app.
AI chatbots customizable according to user interests.
Extended reading and listening in the main curriculum path.
Outcomes
I predict that implementing my recommended features will increase the number of advanced users by 100%, such as those who reach ACTFL’s intermediate-high level.
Research Process
1. Discovery Interviews
Objective: I wanted to test my hypothesis that users would use and benefit from the same app features I would find useful.
2. Academic Literature Review
Objective: To gather information from credible sources that would challenge or confirm the data from discovery interviews
3. Competitor Analysis
Objective: To test my assumption that current market offerings do not already have features similar the ones I had in mind, and to uncover other shortcomings that might be exploited to gain market share.
4. Triangulation of Data
Objective: To find a common thread among the insights from my other research methods and strengthen the legitimacy of my recommendations.
User Interviews
Methodology: I interviewed potential users and made an affinity diagram of quotes that stood out when listening to recordings of our sessions.
Findings:
Learners want Customization
Learners want Customization
Language learners use many different methods and resources. They often do so simultaneously, and they are wary of the effectiveness of any one method alone.
Customize! When we went to Costa Rica, I wished there was an option for construction terms.
-Participant 1
Learners want to speak and listen
Learners want to Speak and listen
Both respondents mentioned that they would want help with what is most difficult for them, speaking and listening. Real-time, authentic conversation improves both abilities simultaneously.
Those classes got me to a place where I could comprehend, but I wouldn’t say I was conversational.
-Participant 1
Learners Want to Connect
Learners want to connect
Language learners use many different methods and resources. They often do so simultaneously, and they are wary of the effectiveness
of any one method alone.
YES! I think of that as like gaming lobbies.
-Participant 2
Learners are curious and wary of AI
Learners are curious and wary of AI
Both participants expressed misgivings about AI’s impact on society, even participant 2 who works in the tech industry.
However, both said that it could be a powerful tool if leveraged to individualize instruction for learning apps.
AI is not “genrative,” it’s plagiarative…but I’d try it in a langauge app.
-Participant 2
Academic Literature Review
Methodology: I read 5 articles on tech-assisted language learning and made a master list of my highlights. Further examination and annotation of my notes revealed three key insights.
Findings:
personalization
“The features they liked best about using apps to support their learning outside class were flexibility, convenience, portability, and the ability to personalize.” (Rosell-Aguilar, 2017)
Interpersonal communication
In oral interaction, there is “more communication breakdown and negotiation” and therefore more opportunities for growth. (Chappelle, 2009)
Weak claims of efficacy
Many claims by language apps are often based on flawed, in-house studies. (Jiang et al, 2021) These studies are used to imply general fluency is a common outcome while failing to mention their product’s shortcomings in teaching interpersonal communication.
Competitive Analysis
Methodology: I made a feature comparison chart in a spreadsheet to determine each of the top three language app’s strengths and weaknesses.
Findings:
Points/Gamification
All of the top three apps have points and rewards systems. Rosetta Stone is weaker in this area. Many users will expect some form of gamification based on past experience.
Input vs grammar
None of these apps feature extensive input (reading/listening) or authentic production in their core curriculum as recommended by many second language acquisition (SLA) researchers. (Thorne et al, 2011)
chat/user interaction
All apps underutilize video chat to connect people on the app, though Babbel and Rosetta do have some limited offerings. Developing communication channels between users may yield a significant competitive advantage.
Ai
At the time of this case study’s publication, only Duolingo had an AI chatbot, and it’s features were limited.
Data Triangulation
I synthesized findings from data collection into three main recommendations:
Learners want choices
Our literature review and interviews both yielded the insight that language learners want to customize their learning experience.
Listening and reading
All three research methods informed us that the approach of many language apps should be replaced with one that focuses on comprehension.
Communication practice
All three methods confirm the need for better interpersonal communication practice options in language learning apps.
Pre-wireframe Ideation
I experimented with an AI wireframing assistant in Figma.
After giving a couple of different prompts, I realized that the outputs were far from perfect, but I took a few ideas from them into the sketching process.
AI Wire 1
AI Wire 2
Hand-drawn Sketches
User Flow Diagram
To finalize my decisions about which screens to wireframe, I made a user flow diagram. Based on the result, I decided to add a lesson question screen and a button not included in my sketches.
Wireframes
The following wireframes represent a user’s path from the app home through a lesson and to the commication options screen.
App Home
Learners are given several ways to start practicing and tutorials that explain the app’s approach.
Lesson Question
This question incorporates comprehension practice while remaining concise.
Lesson Completion
The learner is given the option of to change learning
communication
“Connect Now” is an option in the bottom nav menu, and can also be tapped at the end of each lesson to practice interpersonal communication.
Results & Impact
By fostering a sense of success through interaction and greater language proficiency gains (and not only gamified rewards systems), students will maintain higher levels of motivation and be far less likely to look to other apps or methods.
I predict that learners reaching ACTFL’s intermediate-high level or beyond would at least double for an app that implements my recommendations.
