Third Quarter 2026 Earnings Call - DUOLINGO INC

Explicit Guidance:

• Q4 2024: Expect year-over-year (YoY) bookings growth of 13%, revenue growth of 14%, and an adjusted EBITDA margin of 26.1%. Adjusted EBITDA expected to be roughly flat YoY at $84 million due to operating expenses growing faster than revenue (driven by a second portion of a $11 million team bonus accrual and continued investments).
• Full Year 2024: Expect 12% YoY bookings growth, 18% revenue growth, and a 26.5% adjusted EBITDA margin.
• 2027 Outlook: Full-year guidance for 2027 will be provided in February 2027.
• Expenses: Operating expenses will likely grow faster than revenue in Q4 due to the second portion of the performance bonus, with the full bonus payment expected in early 2027.
• Margins: Gross margin expected to decline slightly in Q4 as video calls are rolled out to all paid subscribers.

Business-Segment Outlook:

• User Growth: DAU growth is expected to remain accelerated in Q4, similar to Q3 (29% YoY). DAU growth accelerated in 8 of the top 10 countries, with double-digit subscription bookings growth in most regions.
• Subscription Strategy: Shift from "Duolingo Max" to "Super Duolingo"; Max is being sunset with individual Max users migrating to Super Family. Super is expected to capture significantly more users than Max previously did.
• Product Features: Video call (previously Max-only) will be available to all paid subscribers by year-end. Adoption of video calls is roughly doubling words spoken per active user YoY; conversion to paid subscriptions is higher when users trial video calls.
• Chess: Currently at 10 million DAUs; identified as under-monetized with significant "low-hanging fruit" expected to be addressed.
• Math: Growing at the same rate as languages, but with ambition to accelerate growth to match the trajectory of Chess.
• Monetization Levers: Introduction of longer free trials, avatar suit monetization (IAPs), and improved ad quality. A-B testing will be expanded to include thousands of monetization-focused experiments in 2027.
• Advertising: Transitioning from low-quality ad networks to a direct sales team and higher-quality ads; exited Q3 growing nicely, with expectations for growth in 2027.
• Light Tier: Currently in testing; not yet a major revenue driver. Potential for regional rollout in lower GDP per capita areas, but final rollout depends on A/B test results.

Market and Competitive Outlook:

• Competition: Management believes they are currently better than a traditional classroom for language learning, though one-on-one human tutoring remains superior.
• AI Differentiation: Anticipates closing the gap with human tutoring within a few years by combining high effectiveness with gamified retention strategies that human tutors lack.
• Gaming Partnerships: Scalability remains high; partnerships with game publishers (e.g., Brawl Stars, Eggie Party) are viewed as a multi-year strategy. These partnerships do not involve monetary exchange but drive cross-promotion and user acquisition.
• Experimentation Velocity: Number of A/B tests run per month is up 90% YoY; this rate is expected to increase by multiples in 2027, driven by AI-assisted experiment design.
• Performance Marketing: Marketing spend is profitable with improving LTV/CAC ratios; paid marketing represents ~60% of non-comp marketing spend and is expected to remain a key channel in 2027.

Macro / Industry Outlook:

• AI Costs: Inference costs per DAU are declining significantly due to a shift from frontier models to open-weight models, which are sufficient for most use cases.
• Compute Strategy: Continued reliance on inference providers with some internal testing; no immediate plan to own compute infrastructure.
• Free User AI Spend: Plan to spend "single-digit millions" on inference costs for free users in 2027 to drive engagement and growth, despite the cost.

Strategic Initiatives:

• AI Integration: Heavy investment in AI for internal productivity (reducing latency from idea to deployment) and product features (speech practice, personalized teaching).
• Expansion: Expect faster rollout of new courses/modules (e.g., Math) using AI; previously took 9 months to launch Chess, now able to do so faster.
• Monetization Focus: 2027 strategy shifts to prioritizing monetization alongside user growth, utilizing A/B testing to find new revenue levers without degrading the free user experience.
• Direct Payments: Expanding direct payment options outside app stores, particularly in performance marketing funnels, though not a primary driver for immediate volume changes.
• Capacity: "Green machine" optimization system continues to drive retention improvements via hundreds of weekly A/B tests.

Risks and Headwinds to Outlook:

• Bonus Accrual: Q4 results include $11M in bonus accruals, with a second portion impacting Q4 opex.
• Margin Pressure: Gross margin expected to decline slightly in Q4 due to the expansion of AI features (video calls) to the broader paid base.
• Bottlenecks: Scaling experimentation face new bottlenecks in idea generation and bug resolution by non-engineers.
• Light Tier Uncertainty: Rollout of the "Light" tier remains uncertain and contingent on A/B test performance.
• Adoption Risks: While one-on-one human tutoring is still superior, there is a risk that users may not find AI-tutored experiences compelling enough compared to other mobile entertainment if gamification elements fail.

Management Tone on the Future:

• Confidence: Management expresses high confidence in the re-acceleration of growth, citing the "great execution" against the strategy to prioritize teaching better and user growth.
• Strategic Shift: Confident in the transition from a pure user growth phase to a balanced approach incorporating monetization, emphasizing that monetization can expand without degrading the free experience.
• Long-term Vision: Reiterated ambition to reach one billion learners; bullish on the long-term opportunity to build a much larger business through broad user base expansion and AI capabilities.
• AI Optimism: Strongly positive on AI's ability to lower costs, increase experimentation velocity, and improve teaching quality, viewing it as a core competitive moat.
• Hedges: Management avoids guiding on 2027 specific numbers, deferring to the February announcement, and notes that specific conversion rate improvements are difficult to quantify but expected through non-friction methods.

Miscellaneous:

• Bonus Structure: Team bonus for re-accelerating user growth is being recorded; full payment expected in early 2027.
• Max Sunset: Moving Max users to Super Family plan maintains price points while improving retention; cannibalization is not seen as material.
• Video Call Cost Reduction: Inference costs for video calls dropped from ~30 cents per call to under 1 cent per call via open-weight models, enabling mass rollout.
• Experimentation Scale: Non-engineers are now able to deploy A/B tests independently, significantly increasing the volume of tests run.