duolyrics

Understanding the Product Vision

Before the kick-off call for this project, I dedicated some time to understanding the role that music plays in language learning. As an avid language learner and former foreign-language instructor, I already had several ideas about how songs can be used as an effective tool but I also wanted to understand the product vision. The founder of the startup, Renato Farias, told me that he had learned Spanish by listening to Hispanic music. He found it to be more fun and engaging than learning from textbooks and it helped him to memorize colloquial terms and connect to the local culture. This is the experience that he also wanted to give other language learners. 

At the time I conducted this research, duolyrics had released the MVP of their app and they had received some unsolicited feedback from beta users. Their assumptions about the typical user of their app were:

• Between 16-34 years old
• Not a complete beginner
• Interested in Hispanic music and culture
• Learning Spanish for social use

However, these assumptions had not been confirmed by formal research yet and beta testers had been chosen among friends and acquaintances. Therefore, we decided on three main objectives for the research:

• Discover the strengths and weaknesses of the current beta version of the app
• Learn more about prospective users
• Identify opportunities for users to stay engaged

Survey

As a first step, I sent out a survey to the beta testers to learn more about them and their experience with the app. Unfortunately, only 5 beta testers filled in the survey and only three were actively learning Spanish. The feedback - apart from a few comments - wasn’t very insightful and nobody was willing to take part in a user interview.

Fortunately, I had also shared a survey on various social media groups for language learners to learn more about their study habits. This survey received 95 responses and over 60% indicated that they were actively learning Spanish. The level varied from complete beginner to bilingual, with 45% indicating that they can have a basic conversation in Spanish.

Almost everyone was listening to Hispanic music at least from time to time and 58% specified that they’re using music as a study tool. One respondent remarked that they had never tried it but that “it sounds fun and interesting!” While some might only passively listen to songs in Spanish, 42% would try to actively engage with the music by reading lyrics, translating and memorizing them, singing or also by using an app.

“Ever since I first started learning Spanish, I've always used music. Same with other languages I've learnt. For Spanish, I started with music from Enrique Iglesias, and I would try to translate his songs using a dictionary.”

Understanding the Competitive Landscape

In the survey, I had also asked about the language learning apps that people were currently using. Out of the 85% who had already used a language learning app, 79 people indicated having used Duolingo. Considering that Duolingo ranks #7 on the App Store, this result is not surprising but I still found the difference with other apps quite striking. While the majority indicated using a mix of various apps, 30% have only used Duolingo.

To get a better understanding, I also asked participants about what they liked and disliked about these language learning apps. While not everyone responded to this open question, I could identify some clear themes through a thematic analysis.

From there, there were some important conclusions to be drawn with regards to the features that the users liked:

• Usability - the most important point since a language learning app needs to be easy to use.
• Bite-sized lessons - the users liked it when apps contained exercises that they could complete in 5-10 min.
• Playful & fun
• Features for progress tracking
  

With regards to the complaints and the features that users disliked:

• The material was too basic - participants complained that most apps were only targeting beginners. In some cases, more advanced exercises are hidden behind a paywall or in some other cases, users can’t test out to skip elementary lessons.
• Hard to stick to apps - some participants would forget about them
• The usage of non-conversational phrases - especially in the case of Duolyrics, many users found the content to be impractical and they were missing explanations about grammar and sentence structure to use their acquired knowledge independently.

“I found it funny and at the same time not so time efficient that Duolingo teaches you sentenses you will never have to say in your entire life like 'tu bebes agua como un caballo' and 'no soy un elefante' 😅 also, it would be nice if these apps taught you how people speak in real life and how to talk to Mexicans for example from a more cultural aspect.”

To understand the competitive landscape that duolyrics operates in, I conducted a competitor analysis. I chose to examine lirica and Lyrics Training in detail because duolyrics had pointed them out as their main competitors. Furthermore, a few survey respondents had also indicated that they were using one or the other.

During my desk research, I had noticed that the rating of an app seems to be related to user retention. As Duolingo currently ranks seventh in the category Education on the App Store, it is crucial to examine their strategies for user retention and engagement. Lastly, I included the app “Peak” in my competitive analysis. While Peak is not a language learning app but an app to train cognitive abilities, it was rated “best of the year” both on the App Store and Google Play and ranked #34 in Education on the Apple App Store at the time of writing.

In the competitive analysis, I examined the language learning experience of each app to understand their value proposition. I focused on the strategies that these competitors are using to keep their users engaged and increase customer retention. I also looked at user feedback on the App Store as well as on respective language learning blogs to understand pain points and identify design opportunities. 

Competitive Analysis

The user feedback that I found is in line with the criticism that I had received in my survey. For example, in Duolingo users criticized the lack of explanations and rules. Users also noted that many exercises were too easy because they could guess the solution. Besides, they felt discouraged when Duolingo accepted only one possible solution and didn't take variations into account. On the positive side, users loved the playful nature of the app and described it as fun, engaging and competitive.

I could identify 9 different strategies for user engagement and retention:

• Onboarding process
• Streaks
• Goal setting
• Notifications
• Levels & scores
• Awards & achievements
• Leaderboards & competition
• Encouragement
• Possibility to provide in-app feedback
  

Not every competitor was using all these strategies and obviously, each strategy has a slightly different goal:

• The onboarding process wants to provide a great experience to first-time users
• Reminders, streaks and goals aim to ensure that users come back to the app consistently
• Gamification through points, awards, leaderboards and competition has the goal to get users hooked to the respective product.
• Feedback in-app increases app retention, something I noticed prominently in Peak.

Evaluating the Current Solution

UX Design Audit

Comparing these findings to the MVP of duolyrics, I could see that they were already implementing two of these strategies. First of all, duolyrics was already showcasing progress in terms of the number of words and songs that users had learned. The app also provided an option for users to contact them. However, both functions were hidden in the “profile” tab and needed to be tested with users.

During my UX design audit, I also noticed potential usability issues. Following Jakob Nielsen's 10 heuristics for interaction design, there were a few issues related to consistency, flexibility and recognition. For example, a speech bubble icon was used to denote a tab to access a vocabulary list as well as a read-aloud function. This might confuse users because it’s not consistent but a speech bubble is also usually associated with a chat function.

In terms of accessibility, I mainly looked at color contrast. Unfortunately, the chosen color palette failed the Web Content Accessibility Guidelines (WCAG) in several cases. Sufficient color contrast is important for people with a color vision deficiency but it is also an important consideration for an app that is often used on the go in bright surroundings because there, the contrast will already be lower.

From a learning design perspective, there are a few other points to consider:

• Users might not know their level of Spanish. A quick test can help them to suggest songs appropriate to their level.
• Students should be able to revise the words that they’ve learned in the songs. This could be done e.g. through an exercise with flashcards.
• Listening to the whole song (or at least the whole verse) is more motivating than listening to a single line. This could be a reward feature for completing a set of exercises.
• Exercises for listening comprehension should use words that sound similar to each other, especially for more advanced learners.

However, from a business perspective, the last consideration is not very viable. It’s more difficult to develop an algorithm that would show the user similar sounding words than selecting other words from the song. In order to decide on possible trade-offs, it’s important to first collect user feedback. 

Interview

While a UX design audit reveals less-than-perfect areas of a product and competitive analysis allows us to understand how similar products are trying to solve user needs, usability testing helps us to collect insights about how people are actually using a product or service. Through the survey that I’ve mentioned earlier, I had recruited 5 participants for a usability test that I combined with an in-depth interview.‍

All participants had different motivations and goals for learning Spanish but their difficulties were quite similar. They struggled with grammar and speaking and one participant mentioned that she found it difficult to move beyond the basic level. While everyone was listening to Hispanic music at least occasionally they were engaging with the music in different ways. One participant was using the competitor’s app Lyrics Training, while others tried to learn songs and sing them.

Because all interviewees were using Duolingo, I was especially curious to know why they kept coming back to this app regularly. In general, people mentioned that they liked the stories feature, the structure (divided by topics) and being able to learn the basics of a language without getting too deep into grammar. However, it was the competitive nature and in particular the streak feature that made them open the app daily.

However, the streak feature also has one downside. One interviewee mentioned that she had lost her streak that she was building over several months and the streak repair feature didn’t work. Frustrated and disappointed, she said that she didn’t open the app again for over a week.  

Usability Test

After the interview part, I shifted to the remote usability test and showed the participants the prototype by asking them to explore the app as they normally would and think aloud. The feedback about the app was overall positive. Participants found it fun to use and they liked the following features:

• Glossary
• Word Pairs function
• Exercises 
• Listening to a line over and over again
• Slowing the music down
• Song broken up in verses (bite-sized)

Nevertheless, I also observed several usability issues, which I jotted down in a modified rainbow spreadsheet.

While all feedback in a usability test is valuable, it is important to prioritize usability issues, feature requests and related input. In this test, I didn't come across a critical usability issue in the sense that users were unable to complete a certain task. However, I came across several issues that left users confused or might even prevent them from using the app again.

Among others, the following issues and functions need to be addressed:

• Progress is shown in terms of words, not time
• No exercises to practice the words again
• Not possible to search and filter songs
• The word pairs/translation function is not intuitive
• Technical terms “chorus”, “bridge”, … are unclear
• Icons do not match expectations
• Not enough songs

For example, it turned out that the speech bubble icon that I had pointed out in my UX design audit, was also not intuitive for users. I also learned that users didn’t find it very motivating to see their progress in terms of the number of words and songs learned. 

“What if there was like a 100 songs like, how useful is it for me to know that I could possibly learn 3000 words and I'm only 1% of the way from learning 3000 words. [...] I feel like that would be confusing or discouraging. You're not going to make a lot of progress. Like even if you do one song a day, you're not going to make tons of progress on that.”

Recommendations

Examining my results, I proposed the following changes:

• Adopt a more accessible, high-contrast color scheme
• Implement an onboarding process emphasizing the value proposition
• Allow users to search for and filter songs as soon as more songs are being added (similar layout to Spotify)
• Ask for in-app feedback via a pop-up and allow users to request songs

I also made a few sketches and mockups to illustrate possible solutions for the usability issues related to the icons and buttons:

To increase user engagement and retention, I suggested the following changes:

• Display the number of words learned in the glossary but without a maximum
• Allow users to set a goal of how much they want to practice each day and implement a streak feature and reminder notifications.

As a second step, it might be a possibility to allow users to connect with their friends and compete with each other. However, following an ideal UX design process, it is important to continuously implement small changes and test them, instead of trying to revamp a product all at once.

Key Takeaways

This project was a fantastic learning opportunity as I could connect my knowledge of language learning and knowledge acquisition with in-depth user research and competitor analysis. I also learned a lot about different strategies for user retention and app engagement and how crucial these strategies are to reach business goals and to ensure the long-lasting success of a product. Taking into account business needs without sacrificing user satisfaction is the key to great UX design.

This project was scheduled for 40 hours but if I had more time I would have done more storyboarding to explore the user's experience with the app and refined my sketches into a new prototype and conducted a second round of testing.