HBS 9-323-016 — Case reading aid

Duolingo: Teaching Languages
to the Masses

Pittsburgh (est. 2012) · CEO Luis von Ahn · 500+ employees · IPO 2021
40.5M
Monthly active users
9.6M
Daily active users
$250.8M
Revenue (2021)
500M+
Total downloads
40+
Languages offered
6%
Paid subscribers
01 — Product mechanics

User journey

📲
Step 1
Download app
Free on iOS & Android
🌐
Step 2
Pick language
40+ languages, placement test
📚
Step 3
Complete lessons
3–5 min bite-sized sessions
⚡
Step 4
Earn XP & streaks
Daily habit loop reinforcement
🏆
Step 5
Compete on leaderboards
Weekly leagues, Diamond tier
A/B testing culture Hundreds of A/B tests running at any given time. Every feature data-validated before rollout.
ML personalization Machine learning adapts difficulty and content to individual learner patterns and weaknesses.
In-house experts Dedicated team of language scientists and curriculum designers. No outsourced content.
02 — Gamification

Engagement mechanics

Streaks

Daily flame counter tracks consecutive days of practice. Users post 1,000-day streaks on social media as badges of honor. Missing one day resets to zero — creating powerful loss aversion.
"Streak is the most powerful retention feature we have." — von Ahn

Leaderboards

Weekly competitions in tiered leagues (Bronze to Diamond). Users who joined leaderboards spent +20% more time learning. Creates social pressure without requiring friends.
"People learn more when they feel like they're competing." — von Ahn

Hearts

Limited lives system — make too many mistakes and you must wait or practice to earn hearts back. Forces focused attention on each question. Also creates friction that drives subscription upgrades (unlimited hearts in Plus).

Notifications

Persistent, personality-driven push notifications from the Duo owl mascot. Users created memes about the aggressive reminders — free viral marketing. "Turn it into a habit" philosophy drives the notification design.
03 — Revenue

Revenue breakdown (2021)

Subscriptions

$180.7M
72% of revenue
Duolingo Plus / Super at $7/mo. Unlimited hearts, no ads, offline access, progress quizzes.

Advertising

$38.5M
15% of revenue
Native in-app ads between lessons. Kept non-intrusive to protect retention and DAU metrics.

Duolingo English Test

$24.7M
10% of revenue
$49 proctored test accepted at 3,000+ institutions. COVID accelerated adoption vs. TOEFL ($200+).

Other

$6.9M
3% of revenue
Virtual goods, in-app purchases (streak freezes, heart refills), and Duolingo for Schools.
04 — Competition

Competitive landscape

Company Founded Price Threat Key difference
Rosetta Stone 1992 $8–12/mo Medium Formal 30-min lessons, ~500K subscribers. Legacy brand but lost mobile-first generation.
Babbel 2007 $7–13/mo Medium Conversation-focused curriculum. Forbes "best overall" language app. No free tier.
Busuu 2008 Freemium Low Community-based feedback from native speakers. Smaller user base.
Memrise 2010 Freemium Low Video-based learning with real native speakers. Spaced repetition focus.
Pimsleur 1963 $15–20/mo Low Audio-based spaced repetition method. Premium positioning, niche audience.
05 — Dilemmas

Strategic tensions

Free vs Monetize
Option A
Keep core product free — maximize reach, user growth, and organic virality
vs
Option B
Push paid conversion beyond 6% with harder paywalls and premium-only features
Engagement vs Learning
Option A
Optimize for gamification metrics (DAUs, streaks, time spent)
vs
Option B
Optimize for actual language proficiency and real-world fluency outcomes
Ads vs Experience
Option A
Increase ad load to grow the $38.5M advertising revenue stream
vs
Option B
Protect user experience, retention, and daily active user metrics
Organic vs Paid Growth
Option A
Rely on word-of-mouth model that built 500M downloads with minimal spend
vs
Option B
Invest in performance marketing to accelerate growth for public market expectations
06 — Key insight

The engagement-learning paradox

Gamification drives metrics, not fluency

Duolingo's gamification drives engagement metrics but course completion equals advanced beginner at best. NYT: "Language apps are not other humans." The company optimizes for what it can measure (DAUs, streaks, time spent) but what users actually want — fluency — may require human interaction that no app provides.

This creates a ceiling: great at getting people started, uncertain at getting them fluent. The product excels at habit formation and vocabulary acquisition, but struggles with conversation, nuance, and cultural context.

Gamify to engage → Users stay for the game → Learning plateaus → Need more gamification ↻
07 — Timeline

Key milestones

2011
Founded by Luis von Ahn & Severin Hacker at Carnegie Mellon University
2012
App launches, starts with English-only courses for Spanish speakers
2013
Named Apple iPhone App of the Year
2016
Monetization begins — ads, Duolingo English Test, subscription tier
2019
First marketing spend. MAUs hit 27.3M
2021
IPO on NASDAQ. 40.5M MAUs, $250.8M revenue
2022
Case written. 500+ employees, 94/6 free/paid split
2023+
GPT-4 integration, Duolingo Max tier launched, surpasses 100M+ MAUs
Duolingo Case Overview · HBS 9-323-016 · Moon (2022) · ESADE Marketing for Entrepreneurs