Almost two years ago, I wrote a post about why I thought Reddit was a good investment. Fast forward to today, that turned out to be a good investment. However, that doesn’t necessarily mean I’m good at picking stocks and you shouldn’t trust me just because of that. Having said that, here is why I think Duolingo is a good investment at the moment.
I really liked Duolingo and bought some shares shortly after it IPO’d. I held it for years until last year, when my feelings towards the app changed. The app went through a lot of changes over the years and I had problems with the direction it was going. I specifically hated that they removed the forum where users engaged with each other, and helped each other. They replaced it with an AI ‘explain my answer’ feature that required a paid subscription. Likely due to user backlash, they made it free on the first day of the year. Another issue I had with Duolingo is that while Duolingo is useful, it is just not that effective. It prioritizes gamification over language learning. It still doesn’t have an Anki style repetition system, which is quite bizarre for such a popular language learning app. My dislike towards the app grew so much that I stopped using it last year and I also sold all my shares.
The Duolingo stock had a wild ride. It went up to 500+ last year and it has fallen more than 80% to around 106, as of February 23rd. Currently its market cap is less than 5B USD. It was puzzling to me when it reached 500+ last year as I saw it as clearly overvalued, but now at around 106, I think it became a good investment once again.
Duolingo, despite some controversy and some user complaints, is still widely popular. I went back to the app about a month ago and I found it better than when I left. I started a Duolingo Max family plan and within a few days I found five others to join my plan to share the cost — in fact, I had to reject two others. The video call feature is much better now. The units are better organized. From what I read, the free version is becoming worse because of the new energy system. The paid version, however, is pretty solid. A lot of users aren’t happy about the aggressive monetization push.
The central questions of whether to invest in Duolingo are 1) whether Duolingo is still a growth company and 2) whether the number of Duolingo Max subscribers continues to grow quickly. The worsening of the free version is somewhat worrying and might hinder its user growth. On the other hand, I think Duolingo Max will be a great revenue driver if it continues to improve. It has improved greatly since its rollout. Duolingo is set to report its earnings a few days later, on February 26th, and we’ll have a better picture then.
Some also worry about the impact of AI on Duolingo. So far anyways I think the impact is minimal. Duolingo (like Reddit) has a social element that AI currently can’t replace yet. I personally try different language learning apps frequently and I haven’t seen a serious contender yet. Some apps might be more effective but they lack the stickiness Duolingo is known for. If Duolingo continues to make good use of AI, instead of being a victim, it could greatly benefit since the cost of tokens keeps decreasing.
It’s impossible to judge whether a company is a good investment without talking about company valuation. Reddit was a good investment when it was valued at 8B, but not necessarily when it was valued at 40B. So far I have avoided calculating a value because company valuation is famously tricky. Discounted Cash Flow is one way. Then there is P/E ratio. Duolingo’s current P/E ratio is quite low because of a tax benefit last quarter. We could use forward P/E instead. Either way, the models have assumptions about revenue and profit growth. In the end, it all seems to be a “guessing game”. In my opinion, what makes it not one is this central question, “is this a useful product that’s getting better?” As a daily user of the product, we’re in a unique position to answer this question. And our experience with the product keeps us somewhat anchored. Certainly, there is limitation here as well. It’s still well worth it to play with different valuation models with different assumptions. Just ask one of the AI models to play out different scenarios.