We are moving Duolingo Inc. (DUOL) to the Hedgeye Technology Short Bench
-
For Updated Data: CLICK HERE
The bottom line is that while we still see a confusing misalignment between the CEO's mission versus the business, and also a product that doesn't get top grades, our linear modeling thesis of post-COVID churn gets shelved for now with DUOL reporting strong 3Q21 subscribers and preliminary data for 4Q21 showing decent enough growth into year end.
There were two key elements of our SHORT:
- Universal education goals and M/L are lofty tenets but DUOL's market mainly constitutes a land of language learning hobbyism that will sustain high levels of churn, and from which it is difficult to extrapolate a l-t retaining business (not quite a charity disguised as a business, but more like, mission driven and the mission isn’t profits or revenue).
- An acceleration from COVID wanes as language learning by DUOL has some fundamental limitations, and early signs of market saturation were evident pre COVID; modeling one-year churn into the future limits the rise of the one-year net subscribers (and resulting revenues) in the coming quarters, forcing the company to report 2022 revenue growth in the mid-20’s%, a number that would give TAM or optionality oriented Bulls pause on the durability of the current business opportunity being pursued and opened by DUOL.
That second part didn’t really play out in 3Q21 on reported numbers, and while our early data on F4Q shows slower growth (arguably seasonal), the company seems to still be sustaining high y/y growth rates which run counter to thesis point #2. The 2H results/data force us to admit that we may be wrong on the linear effects of a post-COVID deterioration in demand and increase in gross churn rates.
Why might this be happening?
It could be that a combination of the annual DuoCon event, plus a beginning of moving content from free (ads based) to behind the paywall for the first time helped lift the 2H21 net subscriber count. The latter could be considered evidence of a less natural monetization curve already appearing at 2M annual subscribers, potentially implying a more circumscribed future landing spot for ultimate subscriber count. But in the meantime, converting users from one side of the ledger to the other supports the bookings/revenue curve, and bulls can argue that the eventual landing spot will depend on the width of products launched and seeded into the market, rather than on the company’s historical (mainly) single product trajectory.
Elements of the LONG thesis we considered include:
- Duolingo English Certification is gaining momentum in part due to an acceleration of acceptance that occurred in the pandemic. The English Certification is a perfect application of DUOL’s l-t ambitions to educate, alongside a fortuitous market acceptance. The company believes the legacy certification market is $5B. It is hard to stack up the actual participants to get to a SAM anywhere in that territory as the top two providers have annual revenues in <$400M combined, and DUOL’s price is 25% that of the competition. For us this translates to a healthy $100M opportunity within reach in the coming several years from a $25-30M base today.
- DUOL is used in ~40% of US classrooms, per management. This is a very bullish, but also hard to endorse, statistic. Still, even at a fraction of this statistic it can imply that if DUOL focuses its subscription services increasingly for classrooms, that the company will have success migrating users from free (ad based) to paid. Schools are likely to be stickier customers as compared to the hobbyist or one-time use market. The combination of greater duration of retention and more legroom for moving users from one side of the ledger (ads) to the other (subscription), implies that DUOL could support growth in the coming years.
Parts of the Long case we think are more suspect.
- Machine Learning based education has demonstrated limitations, in our view
- The founder's goal of universal education seems both too far in the future to be worth consideration, and also potentially not a great market. While DUOL has aspirations to be relevant outside of language learning, our work indicates that, more than ten years into the journey, they are still receiving Incomplete marks from most language students who note the app is good for getting introduced to a language but not for completing its mastery. And, the company has launched products into the non-Language market several times, mostly without success. Lastly, are we all so certain the universal education market will be a good one? Other models that have gotten there first haven’t exactly shown stellar results. The companies we have reviewed in this category, such as COUR or PS, don't tend to perform like sustaining growth companies.
As we refresh our thesis on DUOL, we see that:
- language learning is not a great business
- potential upside risk to financials exists if DUOL is adopted more broadly into school systems (which would enhance the stickiness of subscriptions but may not become a great business)
- DUOL is still a mission driven entity in which the mission is not 1:1 in sync with being a business and unlocking revenue categories
- DUOL has not been successful outside of language, despite trying
Bulls may counter that the massive scale of DUOL shows the company has found a formula for making available content that is in demand by a mass audience, and while education happens to be how they get paid, fundamentally DUOL is a successful digital content business with enormous scale that will find ways to monetize over time.
In sum, we still have a Bearish view here on DUOL, as the mission goals and the business goals seem misaligned. We tap out because we had set $90 as a water mark for testing the thesis, and having realized that level recently, with the linear modeling side of our equation having proved itself wrong, we don’t really have a good understanding today of the timing or duration of results that will show themselves to be skinnier in growth terms.
We’ll keep the Bearish bias by moving DUOL to the Short Bench, but we’ll no longer recommend a day to day Active Short.
DUOLINGO (DUOL) RESEARCH LIBRARY
Black Book Presentation:
- DUOL | Do You Speak Churn? | New Best Idea Short (8/30/21)
- Webcast & Materials: CLICK HERE
- Topic Timestamps:
- Introduction, Duolingo Back Story, & Short Thesis (0:00 – 8:10)
- Churn Cometh, Growth Sloweth (8:10 – 19:17)
- PhD Geniuses Vs. Circumscribed Language Learning Market (19:17 – 28:46)
- Shouting "M.L.", But Product Not Improving/Little Innovation (28:46 – 36:59)
- Duolingo is Not the Only Model In the Language Learning Market (36:59 – 44:01)
- Conclusion: They Never Intended to be Here in the First Place (44:01 – 50:03)
HT3 Data Updates:
- Data Update | November 2021 | ZM, DUOL, GTLB | Key Idea Data Updates (11/1/21)
- Webcast & Materials: CLICK HERE
- Topic Timestamps:
- Hedgeye Process & Introduction (0:15 - 2:08)
- DUOL: Best Idea Short Update (36:18 - 41:53)
- HT3 | October 2021 | APPN, PATH, PLTR, DUOL | Key Idea Data Updates (10/1/21)
- Webcast & Materials: CLICK HERE
- Topic Timestamps:
- DUOL (31:26 - 45:48)
Notes:
- DUOL | EPS First Look | Good Results, Shifting Users from Ads to Subscribers (11/10/21)
- CALL INVITE | DUOL | New Short | Do You Speak Churn? (8/24/21)
- HT-Snapshot | DUOL: Two Stones, One Bird (8/7/21)
Ami Joseph
Managing Director
Twitter
LinkedIn
Yosef Vaitsblit
Director
LinkedIn
Nick Balch
Analyst
LinkedIn
