Description

I think Duolingo is worth spending time on for growth-biased investors despite being up 46% LTM (I’m expecting a “growth investor club” tag already). It’s a great product-driven, found-led, and mission-driven company in the language learning vertical with absolute market & mindshare dominance that has consistently demonstrated ability to continuously improve user engagement and grow paid subscriber penetration while exhibiting discipline on marketing spend (90% of new users are organically-acquired). This is a profitable consumer software company with incredible growth KPIs, and the stars are really getting aligned for 2023 and beyond with: 

1. Insane operating momentum: DUOL added over 18mm MAUs in 2022, more than 2020 and 2021 combined; consistent paid user conversion disproves any COVID beneficiary risk

2. Profitability inflection: 2023 is likely the 2nd inning of a long EBITDA margin ramp to steady-state goal of 30%+

3. Actively taking advantage of Generative AI: OpenAI is already being used in premium high-price subscription tier to drive further monetization (win/win for DUOL/consumers)

The stock will most likely have a pull back at some point, but it’s best to familiarize yourself with the asset ahead of that future dislocation

DUOL Overview / monetization model

Duolingo is a language learning app based on freemium model on track to generate ~$500mm+ revenues in 2023 with 10%+ EBITDA margin (and higher FCF margin given negative working capital associated with subscription)

Use case: ~55% of peoples use to learn English, while ~35% of people are English speakers who are learning another language

Scale: ~60mm MAU as of Q4 ’22 with 27% of them being DAUs

Paid subscriber base of 4.2mm, implying a paid penetration at ~8% of MAU as of Q4 ‘22 

PnL structure today: ~73% LTM Adj. Gross Margin (COGS are predominantly app store fees), heavy investment into R&D (ex SBC) at ~33% of revenue, relatively light S&M (ex-SBC) at ~17% of revenue, G&A (ex-SBC) at ~19% of revenue  DUOL has been running break-even on Adj. EBIT basis since early 2020 and LSD % EBITDA margin over the past 2 years

Three main monetization methods

• Subscription is the biggest revenue contributor: DUOL tries to upsell users to subscribe to Duolingo Plus / Super Duolingo to have an ad-free experience  and unlock certain features

• Ads: a way to monetize the free-to-pay DUOL users

• Assessment/D.E.T.: proprietary low-cost English test gaining fast adoption against legacy English proficiency exams

• Heavy use of gamification and A/B testing to drive engagement and paid conversion

From old deck (revenue mix has changed):

New DUOL users face only a handful of steps before they start learning a new language – no profile info / credit card required. This simplified experience presents a sharp contrast vs. the other major language learning apps (such as Rosetta Stone) that require more comprehensive user profile information as well as payment method to start using them

Introductory learning screen shot for Portuguese

DUOL starts teaching you words paired with cartoon depictions, and train you to complete simple phrases and sentences for you to earn points.  DUOL tracks your learned vocabulary base and gradually trains you to construct more complicated sentences as your vocab base grows over time.  The app is very religious about reminding you to continue your streak to help you form a habit of daily practice

While subscription/ads monetize language learning, Duolingo English Test (DET) has been a game-changer for language testing

• DET is an English language proficiency test that is 1) materially less expensive to take than the two legacy tests, 2) more convenient for the test-taker

• During COVID lockdown, DUOL started pushing for wider adoption for DET which can be taken at home given tech-enabled online proctoring ability

• DUOL pushed hard for testing acceptance during a period of physical test center shut-down, getting it accepted at 3k+ global universities vs. ~800 at beginning of 2020 

• Students love the value proposition and continue to “vote for” DET post pandemic given budget + convenience

• Now accepted at 3,600+ programs, at all top 25 US colleges

• Nice user acquisition channel for EM learners

• Est ~15-20% of students applying to US colleges are using DET 

• Beyond education institution: DET recently got accepted by Ireland for work visa and UK approval is pending

• Upside to penetration: ~6mm TOEFL/IELTS test-takers per year

DUOL’s relative dominance vs. competitors in the online language learning space is unquestioned, I think this is exhibits from M Science tells the story well

DUOL’s success / competitive advantage stems from its stellar gamification design, translating into heightened user stickiness

Gamification lever #1: “epic meaning & calling” – where a player believes he is doing something greater than himself or he was chosen to do something. Duolingo is notorious for constant and almost over the top notification bombardment

Lever #2: “development and accomplishment” – allowing users to feel sense of getting closer to goal. Breaking challenges into stages and showing players progress with points, gems, levels and more

Lever #3: “ownership and possession” – giving players a feeling of ownership through virtual goods/currencies and encouraging players to get more of it

Lever #4: “social influence and relatedness / envy” – allowing users to connect and compare with others

Gamification lever #5: “scarcity and impatience” – try to motivate a player by teasing him with things he may miss out on

Gamification lever #6: “loss and avoidance” – try to motivate a player through fear of losing something. For example: one core driver for withholding your streak is the fear of losing it, and DUOL even tries to CAPITALIZE on loss of streak

Duolingo Mgmt: “We came up with the idea to visually "damage" the progress learners made by making it look like the completed gold skill icons were cracked. Learners have told us how much they enjoy a screen full of perfect, shiny, completed gold skills. We wondered if we could change the design of the skill to entice learners to go back and practice older lessons and thus repair the skill to its golden glory”

DUOL emphasizes on fanatically performing A/B testing to optimize learner engagement outcome and monetization outcome

DUOL has a relatively heavy R&D budget and sees itself as a product-driven company with a mission. Engineers make up almost ~50% of total headcounts with 70% of them focused on products. R&D is the engine supporting the MAU + DAU growth. There is big focus on using AI/ML algorithm to mine the 5B+ exercises a week completed by DUOL’s users to develop tools to best personalize each user’s experience + optimize engagement / conversion

Example: algo can control the difficulty level of each exercise so the “projected mistake rate” is optimized to make the user experience both challenging and engaging

DUOL is known for running ~500+ A/B tests per quarter

DUOL flywheel for those who don’t throw up at that word 

How does DUOL fund all this R&D? By having highly-efficient customer acquisition that maximizes Marketing ROI

90% of users were acquired organically, which is quite amazing at scale. Primary channels are word of mouth (50%), with the rest from brand campaigns and social media influencers / partnerships, with only 10% acquired from paid acquisition channels. In particular, DUOL is masterful at using TikTok 

S&M spend is quite limited as % of revenue (sub-20%) and has generally been a slight source of operating leverage 

DUOL’s TikTok account exploded in 2021 once it started to aggressively featuring its signature green Owl. Much of DUOL’s earlier TikTok content was “tutorials”, simple “how to” videos teaching phrases in various languages. Then it started featuring its owl mascot Duo into these TikTok clips featuring trendy topics + funny / menacing tone – translating into instant success

DUOL then doubled down on the content type that worked (crazy owl) and follower count kept accelerating. DUOL’s TikTok strategy is much like its teaching methodology: emphasizing standout content, continuous improvement, and multiple ways to learn

How big is the TAM?

Offline language learning market shrank while digital learning market grew through COVID, they will likely both grow going forward but digital penetration should continue to move upward over time

HolonIQ (global edu mkt intelligence platform) quantifies the Global DTC Language Learning market at $61bn in 2019, with digital having ~20% of penetration ($12bn)

Caveat is that language learning market is highly fragmented, and range of estimated TAM for Online market is very wide – ranging from $6bn (Meticulous Research) to $20bn+ (Technavio). A significant piece of global language learning market comes from China – with rising uncertainty given tutor reform

Super-bulls believe TAM is potentially beyond language learning given many users perceive it as a fun and engaging gaming gap to “fill time”. This math from Evercore ISI initiation hinges on 5-10% paid sub penetration over time from 4bn language learners, which is way too aggressive (despite painting a $8-24bn TAM)

My view is more nuanced, I take a significant haircut on multiple variables (I consider these conservative estimates) and yet still get to a ~$5bn DUOL subscription revenue TAM that grows over time. The take-away is that it is very large no matter how you cut it

Engagement trends:  Q4 2022 continues to show stellar engagement acceleration – putting churn / fad risk to rest for now

Q4 was historically a “weak / flattish QoQ MAU growth” quarter due to seasonality for DUOL. but MAU growth (whether measured YoY or QoQ) was astoundingly robust in Q4 2022 – great Q for user growth/retenti. MAU YoY % growth 2 yr stack is at all-time high

The pace at which DUOL is adding MAUs YoY 2 years after COVID is way ahead of the early 2022 / COVID benefit period, this is not a fad

DUOL is excellent at pulling monetization levers 

Paid user conversion hummed along like a fine engine, Q4 Paid subs net-add of 0.5mm is extremely robust, quarterly record especially for a Q4, smashing all COVID records.  Paid subscriber penetration now at ~8%, mgmt’s LT aspiration is mid-teen % given that’s where dating apps are 

Product ramp – actively taking advantage of generative AI

DUOL announced it was introducing new subscription tier called Duolingo Max on top of Super Duolinguo. New tier uses generative AI / has improved product experience. 

Developed with the help of OpenAI, this will be priced at premium – potentially 1.5 to 3x the cost of Super Duolingo sup (quick check of Reddit confirms this pricing uplift with learners reporting annual subscription pricing around $140 per year). 

There are two new features: “Explain My Answer” and “Roleplay”. The former provides in-depth explanation to help users understand mistakes. The latter allow users to chat with DUOL characters (like the Owl) to build conversation skills by providing direction and feedback

It’s been tested in select markets and will roll out throughout 2023, potentially further allows users to ramp in-app purchase to increase monetization. DUOL is also intending to utilize generative AI to enhance automation of content creation process – can lead to greater efficiency in compression of content development timeline

Now you can argue DUOL can expand beyond the traditional language learning TAM to an extent

A well-designed app/curriculum that makes language learning more accessible + more fun + more social can technically EXPAND language learning TAM 

For example, DOUL says ~80% of its US users were not learning a language before using Duolingo

BEYOND language learning – many users view Duolingo as a fun app (hybrid of game / language learning with “language” being the icing on the cake). Most common answer for response to ‘what would you do without Duolingo’ is ‘spending more time on Instagram’

37% of DUOL users learning from English UI admits to be a former player of Candy Crush game and that Candy Crush is a waste of time

Where are we at with regard to Duo Math? This math learning app was launched in October 2022 and progress is “so far so good”

We are at early innings in terms of product adoption, no monetization until user > 1mm DAU

Question: Duo Math, though, that math product, just where is that? Are you still tweaking it? Is it far from being where you want it to be? Just talk about kind of adoption, how much interest you've really seen in that.

CEO:  we're very happy with the progress. It's growing rapidly. We're happy with that. Just to remind you of the plan, we are -- right now, the goal is to grow our math product organically for it to grow by itself. Once we get to a pretty decent number of users, call it 1 million daily active users, this is kind of -- it doesn't have to be exactly a million. Once we get there, we're going to say, okay, great. We have good product market. It is time to start monetizing. We're not there yet, but it is growing pretty nicely, and so we're happy with the results.

In terms of tweaking guests, we're working very hard on tweaking the app. We're adding a lot of the features that we know work well on the main Duolingo app to the math product. So improvements to the Streak. We're thinking of adding things like leaderboards, better notifications. So a lot of the stuff that we've learned over the last 10 years on Duolingo and that we have the code for, we're starting to add to the math app to get it -- to have higher retention. We're also adding more content to it.

One of the things that happens with the math app right now is people actually run out of content because if you are very -- it turns out we have some very heavy users that use the whole thing, and then they kind of run out of content. So we're adding more content to it. So you'll see us over the next year just make the product better and better. And because of that, it will continue growing.

2023 is expected to be a break-out year in margin expansion

Going fwd, DUOL anticipates it will get lvg from both S&M and G&A starting in ’23, guiding to 30% incremental EBITDA margin. I will use the following charts to lay out DUOL’s margin progression vs. its goal (everything below is ex-SBC)

Historically DUOL was hyper focused on maximizing growth while keeping EBITDA margin around break-even or slightly positive. For example, 2022 wasn’t a huge “margin leverage” year (but a stellar growth year), but it looks like the company has enough visibility to start pulling margin levers harder in 2023

2020-2021- year of breakeven with occasional positive EBITDA quarters

2022: every quarter was EBITDA positive 

2023: real inflection – guiding to 10-12% (starting from Q1 at ~10%)

To the SaaS bears, yes the SBC is still plentiful, but given DUOL’s market cap the annual dilution should be ~1% / year at current valuation (~75mm SBC over a $6.5bn mkt cap). I like to look at assess a SaaS company’s organic revenue growth on a “revenue growth per diluted share” basis to assess these long duration growth stories, and I can live with this

One thing to note is that DUOL drives consistent outperformance vs. sandbagged EBITDA guidance 

For context, FY 2022 revenues came out at 27% higher than the initial guidance mid-point

Revenues have consistently been beating by ~11% and quarterly EBITDA usually comes out ~6mm ahead on average

Return expectation / valuation

Given the long duration nature of the business, I think about “what do I need to believe” to make ~20% IRR on DUOL for 5 years. Assuming exit multiple slightly higher than INTU/ADSK/ADBE, 30% steady-state EBIT margin, DUOL needs to hit ~$2bn revenue by 2028 to make this math work, implying mid to high 20% revenue CAGR, which I think is do-able given past history, current growth cadence, and TAM size

What do we need to believe to make 15% IRR over the next 10 yrs assuming similar exit multiple and end-point margin?

Catalyst

Durable growth paired with margin expansion