Duolingo is a language learning app that is free and runs on a freemium model. 

DuoLingo: One of the most user-friendly, simplistic and beautiful interface apps with interactive learning. The best part about the app is learning by playing, I have never felt it as a learning app rather I have seen it as a very good game. The streaks help me achieve more and stay motivated. The Product Managers have made the app in a way that is goal-oriented and keeps you motivated to do more while also learning. It even shows politely that this ad helps us keep education free.

As an avid user of Duolingo myself, and as a Product Manager I thought these were some of the main pain-points that all customers face and I wanted to find a better and highly effective way to teach in addition to the amazing job that Duolingo is doing. The best part about the app is learning by playing, I have never felt it as a learning app rather I have seen it as a very good game. The streaks help me achieve more and stay motivated.

Product vision and strategy:

As a daily user myself, I feel there is more money being left on the table and we can hook learners more to the app. Hence, I chose User Engagement, Customer retention, and monetization as the main KPIs that I want to improve.

User Engagement and Retention:

English was my 4th Language and the only way I mastered it was by speaking to native speakers. I had taken tons of English language courses and knew each word perfectly and the most important area that I lacked in was language usage. I knew the words, I could totally understand what the other person is trying to convey, but I had a tough time forming my sentences and speaking.

Validation of the pain-point: I interviewed 6 multilingual customers and asked open-ended questions to seek and validate the underlying problem. These were the main point points that I sought.

  • They can understand the language thoroughly but couldn’t speak fluently because they weren’t speaking to native speakers and they thought their usage is correct.
  • Being able to read the script is not of primary importance, being able to understand the emotion and establish a conversation was the most important thing.

Redefining the Job to be done: The best way to learn a language was is to speak with native speakers and understanding the script is not a priority at this point in time.

Product Design:

I propose 4 new features that will boost user engagement rapidly and help users learn better.

  • The new feature that I propose is to have a 1:1 mock practice session with a native speaker. This is entirely crowdsourced and works on a mutual give and take policy.
  • The mock interviewers will give stars on a scale of 5 based on the performance and this can be used to rank your position on the leader board.
  • Create a new discussion forum where questions can be asked and other native speakers can answer them, and vice versa.
  • Show a detailed analysis of their progress report, people feel happy too see their progress in the data representation/visualization form.

Feature 1: Peer to peer mock practice

The best way to learn a language is to speak with native speakers, thus, conducting peer-to-peer mock practice sessions can be of great help to learners and will be a win-win for both the parties.

Implementation: As soon as a new user signs in with their Email id and enters the landing page, ask what their native language is and what language they want to learn. Along with their daily practice, they can go to the leaderboard and schedule a 1-hour peer-to-peer mock practice with other Duolingo users who speak that as their native language. Each of the parties will get a 25-minute slot and will have a 5-minute parking lot time and feedback. After the call is over each of them will get a chance to give a rating on a scale of 5 and this will help them stay motivated and get an opportunity to speak to native users.  The high-fidelity mock-ups will look like this:

Fig 1. Login/Sign-up page.

Fig 2. Native language selection after you login












Fig 4. A new feature called “practice along with peers”

Fig 3. Home page showing achievements

Fig 5. Peer mock prep schedule tab

Working: Sam’s Native Language is English and is using Duolingo to learn Chinese, Jinghan on the other hand is a native speaker of Chinese and is using Duolingo to learn English. Sam signs up for the mock practice on the Duolingo app and register’s is availability. Jinghan sees that Sam is a native English speaker and is learning Chinese in which she is proficient in. She too registers and schedules a mock round at a convenient time. Jinghan speaks to Sam in English and puts into test all her learning from Duolingo. This happens for the 1st 25 minutes and then the next 5 minutes is spent by Sam giving her feedback and gives time for asking questions. The next half an hour will be Sam’s turn to speak in Chinese and seek feedback.

Feature 2: Rating each other’s proficiency

After the peer-to-peer mock practice sessions, a peer-to-peer evaluation is carried out. After the call both will have an option to give each other a rating based on their proficiency. This rating can be combined with the leaderboard rating and a holistic view of the user can be determined. The mock-up for the user rating is shown in the mock-up below.

Fig 6. After the mock-prep, each of the users can rate the other person’s performance

Feature 3: Access to a discussion forum

Without hiring any coach for mentoring the best way to crowdsource mentorship is to build a community. Questions can be posed by learners and native speakers can answer them and thus offer a mentorship program and get their doubts get cleared. As an add-on a couple of language instructors can be used to moderate the group and the conversations. The high-fidelity mock-up for this will look like:

Fig 7. New feature: access to discussion forums

Feature 4: Offer insightful details about users learning and motivate them to engage more

We as humans we like the spotlight on us, if they see their personalized data analytics and improvement areas that will boost their confidence. Data such as their most strong and weak areas, what is their speed of learning as compared to other similar learners etc.

The high-fidelity mock-ups will look like this:

Fig 8. Peer to peer evaluation and progress analytics


Now the million-dollar question, how to generate extra revenue from this model?

The best way to monetize a freemium model like Duolingo is to have add on features that will tempt the basic plan user to shift to the premium plan. All the 4 features are extremely distinguishing and will nudge the user to take up the premium plan. To get the users a taste of all the features, all of these can be offered as a one time try and ask them to upgrade if they want more of these. They can practice with more people, get access to the discussion forum (call them Duolingoian), get access to detailed analytics, etc. if opt for the premium version else have limited access with the basic version.

Revenue estimation:

Duolingo currently has 300 million downloads on iOS and Android devices, even though there no clear picture of how many daily active users are there, I am guesstimating that 55% of the population might be using it daily

0.55(300M) = 165M, for simplicity and to deal with round numbers say there are around 170 M DAUs.

I assume that 7% of the DAU get converted to the Paid version (7% conversion in a Freemium model is really huge and this is after all of these features are implemented)

0.07*170M = 11.9M ~ 12M users get converted to Duolingo Plus.

Duolingo provides 3 flexible plans starting from $6.99 to $12.99, on average most people will choose the annual plan as it definitely takes them at least 1 year to learn a language.

The average customer pays $7.75/month* 12M users will be around $100M

The revenue generated by Ads: I assume there will be 8 M users across the globe in any given hour, and the number of impressions may be around 3M/hr.

Say the industry standards of impressions and cost per click rate is $3 per 1000 impressions:

9,000*24*30*12 = $77M

Adding these 2 sources of revenues the new revenue that could be generated by implementing these features is: $177M

The 2019 estimated revenue is around $86M

The percentage increase will be around 105%


These are the main features and that will boost user engagement, retention and in turn increase the lifetime value of a customer and overall increase the revenue.  I still have a couple of more KPIs that will improve Duolingo’s engagement and thus revenue generation. Also, I can discuss more about the trade-offs, exact monetization strategy, implementation, improve Ad-revenue generation and more. I also have different strategies which can aim at targeting different customer segments and capture value for them leveraging Duolingo’s core strengths.

The link to the high-fidelity mock-ups are given here: https://share.proto.io/G6CEOJ/

LinkedIn: https://www.linkedin.com/in/ani-bharadwaj/

About me and my Projects: http://sites.duke.edu/anirudhabharadwaj/


Thank you very much for your patient reading:

Please comment your thoughts on my analysis and please share this post.