How Duolingo Reignited User Growth

Duolingo experienced a remarkable 350% growth acceleration, driven by specific product changes and an innovative growth model.
Jorge Mazal, the former Chief Product Officer of Duolingo, shared insights on this growth journey in a detailed newsletter article.

Jorge Mazal joined Duolingo as Head of Product in late 2017.
At that point, Duolingo was the most downloaded education app globally with hundreds of millions of users and aimed to provide the best education in the world universally.
However, user growth had slowed significantly; by mid-2018, the Daily Active Users (DAU) were increasing at a single-digit rate, concerning investors eager for rapid monetization.

The focus of initial strategies was more on boosting retention rather than acquiring new users due to existing organic acquisition channels.

The issue was labeled as a "leaky bucket" problem, where many users were not retained after initial engagement.
Jorge and his team suspected that gamification could improve retention as previous gamification strategies had already been implemented successfully, such as progress tracking and achievement systems.
The decision was driven by the observation that top digital games had significantly better retention rates than Duolingo's.

Forming a Gamification Team to explore new gamification features was the first step.
Inspiration was drawn from the game Gardenscapes, with the introduction of a counter mechanic simulating move limitations similar to competitive gaming.

After a couple of months, the counter mechanic was implemented but yielded no positive results:
- No change in retention or DAU observed.
- Team members experienced divisions in opinion leading to disbanding of the initial efforts.

This failure highlighted the disconnection between the mechanics borrowed from Gardenscapes and their adaptability to Duolingo, which lacked strategic depth in answering quizzes.

A new team focused on user acquisition was formed, inspired by Uber’s successful referral programs.
A free month of premium subscription (previous called Duolingo Plus, now Super Duolingo) was offered to both referrer and referee.
This led to only a 3% increase in new users, insufficient for growth objectives.
The failure to capitalize on their key user base (those already subscribing to the premium service) was noted as a key reason for failure.
Reflection led to the understanding that a different approach was necessary.

A broader comprehension of how other features operate in different contexts was determined essential.
Better judgment was needed in adapting features by assessing their suitability for Duolingo's unique context.

Utilizing Duolingo's data for actionable insights emerged as a means to find a North Star metric for growth.
Jorge drew on experiences with Zynga and MyFitnessPal to segment user retention metrics:
1. Current Users Retention Rate (CURR): Probability of returning users.
2. New Users Retention Rate (NURR): The probability of a user returning after being new to the product.
3. Reactivated User Retention Rates (RURR): The chance a reactivated user returns after a less active period.
This inspired the creation of a new retention model adapted from existing industry standards.

A model with mutually exclusive and collectively exhaustive (MECE) buckets representing user engagement was developed.
Each bucket tracks user engagement over time, representing retention rates categorized as:
1. Current Users: Users active from day one.
2. Resurrected Users: Users returning after a long absence.
3. At-risk Users: Categorized by levels of prior activity to predict churn.
Daily snapshots of user data were introduced to help better model user movement and retention rates over time, ultimately aiming to influence DAU growth.

By testing changes to retention rates, a significant impact was identified on DAU from MODIFYING CURR, which had a compounding effect on overall user retention metrics.
Recognizing the importance of CURR, changes focused on optimizing this metric helped catalyze the growth strategy.

Establishing a dedicated Retention Team with CURR as the North Star metric marked a shift from prior endeavors focused on new user acquisition.
Tracking the variable impacts of various user engagement strategies became a priority.

Implemented an enhanced leaderboard for users to compete with their friends, making it essential for engagement by instilling competition.
- The leaderboard utilizes a league system offering progression similar to competitive gaming platforms, encouraging users to engage consistently.
Preliminary results showed a 17% increase in learning time and tripling of highly engaged learners (defined as users spending 1 hour daily for 5 days).
The credibility of traditional retention metrics improved significantly as the leaderboard became an effective feature.

After analyzing previous failures with push notifications, a rule was established to prevent overwhelming users while maintaining notification effectiveness.
The retention team focused on optimizing the contents and timing of notifications, leading to positive growth through well-structured A/B testing.

Noticing that users achieving a 10-day streak had reduced drop-off rates prompted re-investment in the streak mechanic.
The introduction of streak-saver notifications helped retain users who might lose their streaks, resulting in a series of enhancements, including calendar views and rewards.
Streaks proved to be a powerful motivator for users, contributing significantly to user retention.

After numerous experiments, Duolingo succeeded in increasing CURR by 21% and reducing daily churn by more than 40%, significantly elevating DAU by 4.5x leading to a productive IPO.
Ongoing growth initiatives led to further enhancements, reflecting an adaptable approach to scaling product engagement.
The lessons learned stress the importance of context in implementation, the value of systematic experimentation, and the significance of continuously analyzing data-driven results for effective growth strategies.