Duolingo Inc (DUOL) 2023 Q3 法說會逐字稿

Good evening, everyone. You haven't accidently tuned into a replay of Duocon. This is actually Duolingo's third quarter earnings webcast. Today after market closed, we released this quarter's shareholder letter, a copy of which you can find on our IR website at investor.duolingo.com.

各位晚上好。您並非無意中收聽了《Duocon》的重播。這其實是 Duolingo 第三季財報網路廣播。今天收盤後，我們發布了本季度的股東信函，您可以在我們的投資者關係網站 Investor.duolingo.com 上找到該信函的副本。

On today's call, we have Luis von Ahn, our Co-Founder and CEO; and Matt Skaruppa, our CFO. They will begin with some brief remarks before opening the call to questions. (Operator Instructions). Please note that this event is being recorded. (Operator Instructions)

出席今天的電話會議的有我們的共同創辦人兼執行長路易斯‧馮‧安 (Luis von Ahn)；和我們的財務長 Matt Skaruppa。在開始提問之前，他們將首先進行一些簡短的評論。 （操作員說明）。請注意，該事件正在被記錄。 （操作員說明）

And just a reminder, we'll make forward-looking statements regarding future events and financial performance, which are subject to material risks and uncertainties. Some of these risks have been set forth in the risk factors in our filings with the SEC. These forward-looking statements are based on assumptions that we believe to be reasonable as of today, and we have no obligation to update these statements as a result of new information or future events.

請注意，我們將對未來事件和財務表現做出前瞻性陳述，這些陳述將受到重大風險和不確定性的影響。其中一些風險已在我們向 SEC 提交的文件中的風險因素中列出。這些前瞻性陳述是基於我們認為截至目前合理的假設，我們沒有義務因新資訊或未來事件而更新這些陳述。

Additionally, we'll present both GAAP and non-GAAP financial measures on today's call. These non-GAAP measures are not intended to be considered in isolation from, a substitute for or superior to our GAAP results, and we encourage you to consider all measures when analyzing our performance. And now I'll turn it over to Luis.

此外，我們將在今天的電話會議上介紹公認會計原則和非公認會計原則財務指標。這些非 GAAP 衡量標準並非孤立地考慮、替代或優於我們的 GAAP 結果，我們鼓勵您在分析我們的業績時考慮所有衡量標準。現在我將把它交給路易斯。

Thank you, Debbie, and welcome, everyone. We had another fantastic quarter. We surpassed our expectations and had impressive user, bookings and revenue growth. And it was a fun quarter 2 with the Barbie campaign and also preparing for our best Duocon yet. We also feel very good about the coming quarter, which is why we're raising our full year guidance to now reflect 40% year-over-year bookings growth. And we're also materially raising our full year adjusted EBITDA margin. Matt will share more details on our results and outlook shortly.

謝謝你，黛比，歡迎大家。我們又度過了一個美好的季度。我們超出了我們的預期，並實現了令人印象深刻的用戶、預訂和收入成長。第二季的芭比活動非常有趣，同時也為我們迄今為止最好的 Duocon 做準備。我們對下一季也感覺非常好，這就是為什麼我們將全年指引提高到目前反映的預訂量年增 40%。我們也大幅提高了全年調整後 EBITDA 利潤率。馬特很快就會分享有關我們的結果和前景的更多詳細資訊。

Every year, we run thousands of experiments to make our products more fun, engaging and effective. A lot of these experiments result in small wins that add up over time. If you look back at the past 2 years, you can see evidence of this in action. In Q3 2021, we had about 10 million DAUs. And this past quarter, we had over 24 million. In Q3 of 2021, we had 2.2 million subscribers. In this past quarter, we had nearly 6 million.

每年，我們都會進行數千次實驗，以使我們的產品更有趣、更有吸引力、更有效。許多這樣的實驗都會取得一些小小的成果，但隨著時間的推移，這些成果會逐漸累積起來。如果你回顧過去兩年，你可以看到這方面的實際證據。 2021 年第三季度，我們的 DAU 約為 1000 萬。上個季度，我們的用戶數量超過 2,400 萬。 2021 年第三季度，我們有 220 萬訂閱者。在上個季度，我們有近 600 萬。

The way we have done this has been through our process of making small changes, running a test to see how users react and then doubling down where we see gains. We've also hit some exceptional home runs. For example, our marketing has been a source of extraordinary wins that were fantastic for the business even though they are hard to predict.

我們做到這一點的方法是透過進行一些小的改變，執行測試來了解用戶的反應，然後在我們看到的收益上加倍。我們還打出了一些出色的本壘打。例如，我們的行銷一直是非凡勝利的源泉，儘管這些勝利很難預測，但對企業來說卻是極好的。

Think of our partnership with HBO's House of the Dragon, and most recently, our Barbie social campaign built around the inclusion of our trademark "ding" in the Barbie movie. The campaign generated 140 million organic social impressions, which is a record for us.

想想我們與 HBO 的《龍之家》的合作關係，以及最近我們圍繞在芭比電影中加入我們的商標“ding”而開展的芭比社交活動。該活動產生了 1.4 億自然社交印象，這是我們的記錄。

To create the opportunity for both incremental improvement and home runs, every so often, we make bolder moves that open up entirely new areas for experimentation. For example, last year, we redesigned the home screen of the app to give us more room to experiment with engagement, efficacy and monetization. And earlier this year, we set out to create a multi-subject app experience by integrating our existing math course and a new music course into our flagship app.

為了創造漸進式改善和全壘打的機會，我們經常採取更大膽的舉措，開闢全新的實驗領域。例如，去年，我們重新設計了應用程式的主螢幕，為我們提供了更多的空間來試驗參與、功效和貨幣化。今年早些時候，我們開始將現有的數學課程和新的音樂課程整合到我們的旗艦應用程式中，打造多學科應用程式體驗。

This change, adding math and music to the main app is a good example to highlight. We did this to give our users an even more engaging experience. By adding new subjects into the main app, we believe we can more rapidly scale these new subjects with our gamification mechanics like streaks, leaderboards and quests that have been so effective in the language learning app.

這一變化（將數學和音樂添加到主應用程式）是一個值得強調的好例子。我們這樣做是為了給我們的用戶帶來更具吸引力的體驗。透過在主應用程式中添加新科目，我們相信我們可以利用我們的遊戲化機制（例如在語言學習應用程式中非常有效的連勝、排行榜和任務）來更快地擴展這些新科目。

As more users access these new courses, we believe it will increase their commitment to our platform and that they'll be more likely to recommend us, which will further drive organic user growth. It's clear that our strategy is working as demonstrated by our exceptional growth. As I've explore, we can't expect to accelerate forever, but it's very gratifying to see that our product improvements and creative marketing efforts are resonating with our learners. And I'm excited about our ongoing innovation and look forward to all the energy that our new year brings.

隨著越來越多的用戶訪問這些新課程，我們相信這將增加他們對我們平台的承諾，並且他們更有可能推薦我們，這將進一步推動用戶的自然成長。很明顯，我們的策略正在發揮作用，我們的卓越成長證明了這一點。正如我所探索的，我們不能指望永遠加速，但很高興看到我們的產品改進和創意行銷努力與我們的學習者產生共鳴。我對我們持續的創新感到興奮，並期待新的一年帶來的所有能量。

And with that, I'll turn it over to Matt.

有了這個，我會把它交給馬特。

Thanks, Luis. I'll walk through how we did this quarter in more detail, and then I'll provide our Q4 and updated full year guidance. As Luis shared, we are extremely pleased with this quarter's performance, which exceeded our expectations, thanks to our team's continued strong execution.

謝謝，路易斯。我將更詳細地介紹我們本季的表現，然後我將提供我們的第四季和更新的全年指導。正如路易斯所說，我們對本季的業績非常滿意，這超出了我們的預期，這要歸功於我們團隊持續強勁的執行力。

Our DAU increased 63% year-over-year to 24.2 million, and MAU increased 47% to 83.1 million. This growth continues to be, not only rapid but also high quality and broad-based with strong growth from around the world with improving free-to-pay conversion as well. Our total paid subscribers increased by 60% to 5.8 million. This continued strength in user and subscriber growth drove bookings and revenue growth of 49% and 43% year-over-year, respectively, or 48% and 42% on a constant currency basis.

我們的 DAU 年成長 63% 至 2,420 萬，MAU 成長 47% 至 8,310 萬。這種成長不僅持續快速，而且品質高、基礎廣泛，來自世界各地的強勁成長以及免費付費轉換率的提高。我們的付費訂戶總數增加了 60%，達到 580 萬人。用戶和訂戶成長的持續強勁推動預訂量和營收年比分別成長 49% 和 43%，以固定匯率計算分別成長 48% 和 42%。

Turning to profitability. We've made tremendous progress in expanding the bottom line as we continue to see very strong top-line growth, coupled with cost discipline. Our net income totaled $2.8 million compared to a net loss of $18.4 million in the year-ago quarter. We also posted an adjusted EBITDA of $22.5 million or a 16.3% adjusted EBITDA margin. This is a roughly 14-point expansion year-over-year. This quarter, we capitalized additional R&D expenditures compared to Q3 last year as we continue to innovate on Max and math and music. Excluding this would have led to an adjusted EBITDA margin expansion of about 12 points year-over-year.

轉向盈利能力。我們在擴大利潤方面取得了巨大進展，我們繼續看到非常強勁的收入成長，再加上成本控制。我們的淨利潤總計 280 萬美元，而去年同期淨虧損為 1,840 萬美元。我們也發布了 2,250 萬美元的調整後 EBITDA 或 16.3% 的調整後 EBITDA 利潤率。年比成長約 14 個百分點。與去年第三季相比，本季我們資本化了額外的研發支出，因為我們繼續在 Max、數學和音樂方面進行創新。如果排除這一點，調整後的 EBITDA 利潤率將年增約 12 個百分點。

Based on our strong results and the trends we're seeing so far, we feel confident about raising our full-year outlook, issuing the following Q4 guidance. For Q4 2023, we are guiding to $167 million to $170 million in total bookings, $145 million to $148 million in revenue and an adjusted EBITDA margin of 19.8% to 20.8%. For the full year 2023, we are raising our guidance to $598 million to $601 million in total bookings, $525 million to $528 million in revenue, and we are updating our adjusted EBITDA margin range to 16.6% to 16.9%.

基於我們強勁的業績和迄今為止所看到的趨勢，我們有信心提高全年展望，並發布以下第四季度指引。對於 2023 年第四季度，我們預計總預訂量為 1.67 億至 1.7 億美元，營收為 1.45 億至 1.48 億美元，調整後 EBITDA 利潤率為 19.8% 至 20.8%。對於 2023 年全年，我們將總預訂量指引提高至 5.98 億美元至 6.01 億美元，營收指引至 5.25 億美元至 5.28 億美元，並將調整後的 EBITDA 利潤率範圍更新至 16.6% 至 16.9%。

Our full year guidance calls for 40% and 42% year-over-year bookings and revenue growth, respectively, at the midpoint. As a reminder, at the end of every December, we start a promotion that discounts our annual subscription. Since we offer this promotion only once a year, Q4 bookings performance has more seasonal variance than other quarters. Note that our guidance assumes current prevailing foreign exchange rates as well. And because we've achieved significant operating leverage this year across the business, we feel good at raising our full year adjusted EBITDA margin guidance by about 225 basis points at the midpoint versus our last call.

我們的全年指引要求中點預訂量和營收年增 40% 和 42%。提醒一下，每年 12 月底，我們都會啟動年度訂閱折扣促銷活動。由於我們每年只提供一次此促銷活動，因此第四季度的預訂表現比其他季度具有更大的季節性差異。請注意，我們的指導也假設當前的現行外匯匯率。由於我們今年在整個業務中實現了顯著的營運槓桿，因此我們很樂意將全年調整後 EBITDA 利潤率指引比我們上次的預測中位數提高約 225 個基點。

Now I'll provide some color on average revenue per subscriber, OpEx and share count. Our average revenue per subscriber has declined by about 7% to 8% for each of the past 3 quarters, driven by a combination of foreign exchange impacts and regional mix shifts. Q3 saw higher-than-expected conversions in non-U.S. countries, which kept the year-over-year change in that same range. We expect the year-over-year change in ARPU to improve in the coming quarters.

現在我將提供有關每個訂閱者的平均收入、營運支出和份額數量的一些資訊。在外匯影響和區域結構變化的共同推動下，過去 3 個季度我們每個訂閱者的平均收入下降了約 7% 至 8%。第三季非美國國家的轉換率高於預期，這使得年比變化保持在同一範圍內。我們預計 ARPU 的年比變化將在未來幾季有所改善。

As to OpEx compared to Q3 of this year, in Q4, we expect non-GAAP R&D as a percentage of revenue to decrease by 2 points and non-GAAP sales and marketing to decrease by almost 2.5 points. We expect non-GAAP G&A to be roughly flat as a percentage of revenue in Q4. We ended the quarter with approximately 48.8 million fully diluted shares outstanding using the quarter end closing price. And we expect to end the year with about 1% to 1.5% dilution from equity issued to employees.

至於營運支出，與今年第三季相比，我們預計第四季非 GAAP 研發佔收入的百分比將下降 2 個百分點，非 GAAP 銷售和行銷將下降近 2.5 個百分點。我們預計第四季非 GAAP G&A 佔營收的百分比將大致持平。使用季末收盤價計算，本季末我們已發行約 4,880 萬股完全稀釋後的流通股。我們預計到年底，向員工發行的股權將被稀釋約 1% 至 1.5%。

Finally, as Luis said a few moments ago, the last years have been extraordinary for our business. Our user growth has benefited from compounding continuous product improvements, including home run improvements like we've seen with our streak mechanic. We have unlocked social-first marketing and have had some big brand moments like the Barbie movie, and we've also seen large increases in conversion from free to paid version.

最後，正如路易斯剛才所說，過去幾年對我們的業務來說是非凡的。我們的用戶成長得益於持續的產品改進，包括我們在連勝機制中看到的本壘打改進。我們已經開啟了社交優先行銷，並擁有了一些像芭比娃娃電影這樣的大品牌時刻，我們也看到從免費版本到付費版本的轉換率大幅增加。

Looking ahead, we feel very good about next year. But as Luis has already reminded you, our user growth is unlikely to accelerate forever, and it may be hard to repeat some of the onetime events that have happened this year. Even so, we feel good about our ability to continue our strong top-line growth and make progress towards our long-term margin.

展望未來，我們對明年感覺非常好。但正如路易斯已經提醒您的那樣，我們的用戶成長不太可能永遠加速，並且可能很難重複今年發生的一些一次性事件。即便如此，我們對繼續強勁的營收成長並在長期利潤方面取得進展的能力感到滿意。

And with that, I'll turn it back to Luis.

說到這裡，我會把它轉回給路易斯。

Thanks, Matt. I just want to thank the team for all their hard work and dedication. We know we've only made a tiny dent in the massive market opportunity ahead of us, and we're just getting started. And now we would be happy to take your questions. I'll turn it back to Debbie to manage the queue.

謝謝，馬特。我只想感謝團隊的辛勤工作和奉獻精神。我們知道，我們只是對擺在我們面前的巨大市場機會做出了很小的貢獻，而我們才剛開始。現在我們很樂意回答您的問題。我會將其交還給黛比來管理隊列。

All right. Thanks, Luis. (Operator Instructions). And your first question comes from Ralph Schackart at William Blair.

好的。謝謝，路易斯。 （操作員說明）。你的第一個問題來自威廉布萊爾的拉爾夫沙卡特。

Maybe, Luis, I'll start with you. You've been public for a couple of years now, and you've seen some big extraordinary growth. And I'm sure some of this has been based on conservatism and the outperformance and then the business is obviously doing really well. Maybe just kind of taking a step back, if you were to isolate maybe the top 1, 2 or 3 things that have really driven this significant outperformance, that would be really helpful, I think, for investors. And then I have a follow-up after that.

也許，路易斯，我會從你開始。你們已經上市幾年了，你們已經看到了一些巨大的非凡成長。我確信其中一些是基於保守主義和出色的表現，然後業務顯然表現得非常好。也許只是退後一步，如果你要找出真正推動這種顯著優異表現的前 1、2 或 3 個因素，我認為這對投資者來說真的很有幫助。之後我會跟進。

Well, thank you for the great question. Yes, we've had a really good growth. And honestly, it's exceeded our expectations. The main 2 things that have made us grow so fast. The first one is just product improvements. I mean, we run hundreds of A/B tests every quarter, and they compound. And so our product is just significantly better than it was 2 years ago. And so that's the biggest thing.

嗯，謝謝你提出這個好問題。是的，我們的成長非常好。老實說，它超出了我們的預期。讓我們快速成長的主要有兩件事。第一個只是產品改良。我的意思是，我們每個季度都會運行數百次 A/B 測試，並且它們會復合。所以我們的產品比兩年前好得多。這是最重要的事。

Then the second biggest thing is we just got a lot better with our marketing. And the combination of getting much better with marketing and the product getting better has made it so that Duolingo has really struck a chord with this -- it's like now part of the zeitgeist and part of the reason that we've exceeded our expectations is because things have happened that we didn't -- there was no way for us to expect. We could not expect that the Barbie movie was going to add Duolingo in there. We could not expect that there would be SNL skits and stuff like that. So that -- I think that's the combination of what has happened. I mean, we're very happy, but it's been hard to also forecast.

第二件大事是我們的行銷做得更好了。行銷的進步和產品的進步相結合，讓 Duolingo 真正引起了共鳴——這就像現在時代精神的一部分，而我們超出預期的部分原因是因為我們沒有發生的事情已經發生了——我們無法預料。我們沒想到芭比電影會在其中增加多鄰國。我們沒想到會有 SNL 短劇之類的東西。所以——我認為這就是所發生的事情的結合。我的意思是，我們非常高興，但也很難預測。

Great. And then maybe, Matt, in the prepared remarks, you talked about the potential of ARPU to improve. And I think you'll be coming up on 2 years since you did some of the price adjustments in Europe, just maybe give some color on how ARPU may improve going forward.

偉大的。馬特，也許您在準備好的演講中談到了 ARPU 提高的潛力。我認為自從您在歐洲進行了一些價格調整以來，您將迎來兩年的時間，也許您可以就 ARPU 未來如何提高提供一些資訊。

Yes. Thanks, Ralph. So ARPU has been impacted by foreign currency. That's hard to predict. But the other thing was the regional pricing that you mentioned. So we did that about a year ago. That was Q2 of 2022. So that impacted bookings in that quarter and then going forward and has flowed through to revenue. So that was expected, and that's what we mentioned on the last call.

是的。謝謝，拉爾夫。所以ARPU受到了外匯的影響。這很難預測。但另一件事是你提到的區域定價。所以我們大約一年前就這麼做了。那是 2022 年第二季。因此，這影響了該季度的預訂量，然後影響了未來，並影響了收入。所以這是預料之中的，這也是我們在上次電話會議中提到的。

In terms of this quarter, there was some experiments we ran that were really effective at driving free-to-paid conversion outside the U.S. So that kept it range-bound this quarter. I think the reason we continue to think we'll see improvements is the -- starting to lap the pricing changes that we did in Q2 of last year. And also, our pricing philosophy is one that we continue to experiment with pricing, and we've added a lot of value to the product over time. So we'll experiment with raising prices over them as well.

就本季而言，我們進行了一些實驗，這些實驗確實有效地推動了美國以外地區的免費到付費轉換，因此本季的情況保持在區間內。我認為我們繼續認為我們會看到改進的原因是——開始接受我們去年第二季所做的定價變化。而且，我們的定價理念是我們不斷嘗試定價，隨著時間的推移，我們為產品增加了許多價值。因此，我們也會嘗試提高它們的價格。

Next question comes from Justin Patterson of KeyBanc.

下一個問題來自 KeyBanc 的 Justin Patterson。

I appreciate the call out on experiments in the letter. Luis, as you've expanded the range of products in the core app to music and now math, how should we think about just the pace of experiments and A/B test really ramping to really kind of optimize these new elements within the app? And then for Matt, the incrementals on the business remain very attractive, north of 50% margins within the Q4 guide. So as you're sitting here today thinking about 2024 budgets, what are really the key investment areas that you're thinking about to keep driving this healthy growth going forward?

我很欣賞信中對實驗的呼籲。 Luis，當您將核心應用程式中的產品範圍擴展到音樂和現在的數學時，我們應該如何考慮實驗和 A/B 測試的步伐，以真正優化應用程式中的這些新元素？對於馬特來說，業務增量仍然非常有吸引力，利潤率超出了第四季度指導值的 50%。因此，當您今天坐在這裡思考 2024 年預算時，您正在考慮的真正關鍵投資領域是什麼，以繼續推動這種健康的成長？

Thank you, Justin, and excellent shirt. So you asked about math and music and what that's going to do to our experiments. So I think it's worth just saying what we did, what -- we had a -- well, we still have a math app and we've been developing a music course. We decided the best strategy was to put both the math and music course inside our main app. There's many reasons for this, but the biggest reason is that we can immediately take advantage of any positive change that we do in the main app for engagement or for monetization or for anything immediately.

謝謝你，賈斯汀，還有很棒的襯衫。所以你問了數學和音樂以及它們對我們的實驗有何影響。所以我認為值得說一下我們做了什麼，我們有一個——嗯，我們仍然有一個數學應用程序，我們一直在開發音樂課程。我們認為最好的策略是將數學和音樂課程都放在我們的主應用程式中。造成這種情況的原因有很多，但最大的原因是我們可以立即利用我們在主應用程式中所做的任何積極的改變來提高參與度或貨幣化或立即進行任何其他事情。

You take advantage of those because math and music are just going to be other courses in the app, just like French or Spanish, just -- it's as if we were adding another language. So that will -- what that'll do is it'll allow us to grow these courses or these subjects a lot faster than if they were independent apps. We're very happy with that.

你可以利用這些，因為數學和音樂將成為應用程式中的其他課程，就像法語或西班牙語一樣，就好像我們添加了另一種語言一樣。因此，這將使我們能夠比獨立應用程式更快地發展這些課程或這些科目。我們對此感到非常高興。

Now I should mention, and I should caution, the main thing that we are concentrated on for the short term is really making both the math and music courses more comprehensive and really work on them a lot. We haven't even launched them. And technically, they are going to be launched to all our users on iOS starting tomorrow. So once we launch them, we're going to be working, specifically on the courses. So I would not expect too much increase in revenue in the short term.

現在我應該提一下，我應該警告一下，我們短期內關注的主要事情實際上是讓數學和音樂課程更加全面，並且真正在這方面做很多工作。我們甚至還沒有啟動它們。從技術上講，它們將從明天開始向 iOS 上的所有用戶推出。因此，一旦我們啟動它們，我們就會開始工作，特別是在課程上。所以我預計短期內收入不會有太大成長。

But over the long term, the hope is that the -- Duolingo gets known, not just as a language learning app but so as an app where you can learn math and music, so we'll start attracting users for that. And a lot of the monetization experiments that we run will apply to math and music. So that's the idea. My sense is that it'll probably slightly increase the pace of experimentation. And so that basically answers your question.

但從長遠來看，我們希望 Duolingo 不僅作為語言學習應用程序，而且作為一款可以學習數學和音樂的應用程式而廣為人知，因此我們將開始為此吸引用戶。我們進行的許多貨幣化實驗將適用於數學和音樂。這就是我們的想法。我的感覺是，它可能會稍微加快實驗的速度。這基本上回答了你的問題。

Great. Yes. And to follow up on budgeting in 2024. The amazing thing about the business right now is that the core business, we're still early in the opportunity. So we have tons of room to run. So the vast majority of our investment next year will be in the core business.

偉大的。是的。並跟進 2024 年的預算。目前該業務的令人驚奇的事情是，在核心業務上，我們仍處於機會的早期階段。所以我們有很大的運行空間。所以我們明年的絕大部分投資都會投在核心業務上。

Just to remind everyone how we invest in that. There's really 3 areas we invest in. We invest in growth. A bunch of engineers and product managers and designers who help design experiments to drive things like innovations in the streak mechanic or make the app more social.

只是提醒大家我們如何對此進行投資。我們確實投資了 3 個領域。我們投資於成長。一群工程師、產品經理和設計師幫助設計實驗，以推動連勝機制的創新或使應用程式更具社交性。

We invest in monetization. So driving free-to-paid conversion. And then we invest in learning to make our app more effective. That's going to make up the vast majority of our investment next year, just like it did this year. We have incremental layers on top of that, that we add.

我們投資於貨幣化。從而推動免費到付費的轉換。然後我們投資學習以使我們的應用程式更加有效。就像今年一樣，這將構成我們明年投資的絕大多數。我們在此之上新增了增量層。

So we're investing Max. We're investing in math and music like we just mentioned. But I don't want folks to lose sight of the fact that the investment we're going to make next year is going to be like this year, which is in that core app, which has great growth momentum and a huge market that we're still growing into.

所以我們投資Max。正如我們剛才提到的，我們正在投資數學和音樂。但我不希望人們忽視這樣一個事實，即我們明年要進行的投資將會像今年一樣，即核心應用程序，它具有巨大的增長動力和我們所開發的巨大市場。還在成長中。

Thanks, Justin. And next question comes from Ryan MacDonald of Needham.

謝謝，賈斯汀。下一個問題來自尼達姆的瑞安·麥克唐納。

Congrats on the nice quarter. Luis, I'm curious, with the integration of math and music into the core app, how do you start to think about moving forward about the experience for the learner based on whether you're a paid subscriber versus a free user on math and music in the core app? And how do you think about moving forward, incentivizing more consistent engagement of moving that MAU to DAU over time?

恭喜這個美好的季度。路易斯，我很好奇，隨著數學和音樂整合到核心應用程式中，您如何開始考慮根據您是數學和音樂的付費訂閱者還是免費用戶來改善學習者的體驗核心應用程式中的音樂？您如何看待繼續前進，激勵更一致的參與，隨著時間的推移將 MAU 轉變為 DAU？

Yes. So thank you for the question, Ryan. So in terms of monetization for math and music. Really, the way to think about them for now is it's as if they were another language course. All that applies to language courses applies to math and music. So for example, if you are a free user and you do the Spanish course, at the end of the lesson, you have to see an ad.

是的。謝謝你的提問，瑞安。所以就數學和音樂的貨幣化而言。確實，現在考慮它們的方式就好像它們是另一門語言課程。所有適用於語言課程的內容也適用於數學和音樂。例如，如果您是免費用戶並且正在學習西班牙語課程，那麼在課程結束時，您必須看到廣告。

Similarly, if you're a free user and you're doing the math course, at the end of the lesson, you have to see an ad. There's also a notion of also hearts where whenever you make a mistake, you lose the heart that happens in math, that happens in music, that happens in the language course. They're very similar in that respect. And we're going to continue working on trying to monetize our users better by basically offering extra things in the premium subscription.

同樣，如果您是免費用戶並且正在學習數學課程，那麼在課程結束時，您必須看到廣告。還有一個「心」的概念，每當你犯錯時，你就會失去心，就像數學、音樂、語言課程中發生的那樣。他們在這方面非常相似。我們將繼續努力透過在高級訂閱中提供額外的東西來更好地從用戶身上獲利。

So there may be -- just like with languages, we have things like the practice hub where certain extra things -- we're going to have extra things for both math and music. Now I should remind -- there are a lot of questions about math and music, which is great. We're very excited about math and music, but I should remind you, for the foreseeable future, the vast majority of our business is in language learning. There's still a lot of room to growth -- to grow in language learning. So the majority of the experimentation is still going to be there. So yes, I think that's that.

所以可能——就像語言一樣，我們有練習中心之類的東西，其中有一些額外的東西——我們將為數學和音樂提供額外的東西。現在我要提醒一下——有很多關於數學和音樂的問題，這很好。我們對數學和音樂感到非常興奮，但我應該提醒您，在可預見的未來，我們的絕大多數業務都是語言學習。語言學習方面還有很大的發展空間。因此，大部分實驗仍然存在。是的，我想就是這樣。

That's helpful. And maybe I'll follow-up, and this could be for Matt or Luis. That was interesting in the shareholder letter about the experimentation you did on paid advertising this year and how that drove sort of a 50% increase in users on the platform. Well, I'd just love to hear a little bit more color on maybe what you did differently this year that drove such a notable increase.

這很有幫助。也許我會跟進，這可能是針對馬特或路易斯的。股東信中提到了你們今年在付費廣告所做的實驗，以及這如何推動平台上的用戶增加了 50%，這一點很有趣。好吧，我只是想聽聽更多關於你們今年做了什麼不同的事情，從而推動瞭如此顯著的成長。

So there are a number of things. I mean, we've just gotten a lot better at figuring out what to say in our ads. And also, we are in a really interesting position where we are a company that has users in essentially every single country in the world. And we don't have to deal with things like licensing music or anything like that. We don't have to deal with stuff. We just can serve the app to any country in the world.

所以有很多事情。我的意思是，我們在弄清楚廣告中要說什麼方面已經做得更好了。而且，我們處於一個非常有趣的位置，我們是一家在世界上幾乎每個國家都有用戶的公司。我們不必處理諸如音樂許可之類的事情。我們不必處理一些事情。我們可以向世界上任何國家提供該應用程式。

So one of the things that we've gotten really good at is being able to shift budget between countries. If prices get higher in one country and lower than another, we just shift budget, because for us, it doesn't matter all that much whether a user comes from Vietnam or from Thailand or something. I mean, it matters in the conversion rate, but we have the math to figure out when it makes sense to shift budget from one country to another.

因此，我們真正擅長的事情之一就是能夠在國家之間轉移預算。如果一個國家的價格上漲而低於另一個國家/地區，我們只需調整預算，因為對我們來說，使用者是否來自越南、泰國或其他國家並不重要。我的意思是，這對轉換率很重要，但我們有數學來計算何時將預算從一個國家轉移到另一個國家是有意義的。

And so we've gotten very effective at that. And that's something that not many apps can do because, for us, it's just really -- we just have users in every single country in the world. So it's stuff like that, that we just, over time, have gotten smarter at. I should mention, though, we did that, but it is still -- it is the case that the overwhelming majority of our users come in organically. I mean, this is -- we have relatively small marketing budgets where we're doing that. But yes, we've gotten a lot better at it.

所以我們在這方面非常有效。這是很多應用程式都做不到的，因為對我們來說，我們在世界上每個國家都有用戶。所以，隨著時間的推移，我們在這類事情上變得越來越聰明。不過，我應該要提到的是，我們確實這樣做了，但情況仍然如此——我們的絕大多數用戶都是有機進入的。我的意思是，我們這樣做的行銷預算相對較小。但是，是的，我們在這方面已經做得更好了。

Your next question comes from Zach Morrissey of Wolfe Research.

您的下一個問題來自沃爾夫研究中心的紮克莫里西。

So I just wanted to, I guess, double-click on the user growth side of things. You've talked about the improvement in retention in prior years and quarters, specifically the current user retention rate. As you kind of think about the growth we've seen this year, it's obviously been very impressive.

所以我想，我只是想雙擊用戶成長方面的事情。您談到了前幾年和前幾季的用戶保留率的提高，特別是目前的用戶保留率。當你想到我們今年所看到的成長時，你會發現這顯然非常令人印象深刻。

Can you kind of parse through how much of that has come from kind of improvement in the retention? And as we kind of look ahead next year, how do we think about kind of room for further improvement in retention kind of supporting user growth?

您能否分析一下其中有多少是來自於保留率的提高？當我們展望明年時，我們如何考慮在支援用戶成長的保留方面進一步改進的空間？

And then second, just on competition. Obviously, the growth you guys are posting kind of speaks to your ability to kind of execute well. But obviously, Google, last month, kind of grabbed some headlines kind of entering the space.

第二，關於競爭。顯然，你們所取得的進步在某種程度上說明了你們良好執行的能力。但顯然，谷歌上個月進入這個領域佔據了一些頭條新聞。

So if you could just kind of talk about your views on kind of the evolving competitive landscape? And more specifically, how Duolingo's kind of competitive moats can kind of last relative to a larger platform such as Google, which has scale, data, technology, et cetera.

那麼您能否談談您對不斷變化的競爭格局的看法？更具體地說，相對於谷歌這樣擁有規模、數據、技術等的更大平台，Duolingo 的競爭護城河如何能持續下去。

Some great questions. For user growth, we believe that the main thing that has affected user growth is improvements in free user retention. That's it. I mean, it's not 100%. Like I said, a lot of our marketing has helped too, but I would say more than 50% is just improvements in user retention. Current user retention rate is probably the biggest lever that we've had. It's not the only one, but it's the biggest lever that we have to move.

一些很好的問題。對於用戶成長，我們認為影響用戶成長的主要因素是免費用戶保留率的提高。就是這樣。我的意思是，這不是100%。就像我說的，我們的許多行銷也有幫助，但我想說超過 50% 的只是用戶保留率的提高。目前的用戶保留率可能是我們擁有的最大槓桿。這不是唯一的一個，但它是我們必須移動的最大槓桿。

We expect there's still a lot of room there for us to improve. I mean, obviously, again, just like with user growth, you cannot improve user retention forever. You also cannot improve user growth forever. But we still think there's room there, so we're happy with that.

我們預計我們還有很大的進步空間。我的意思是，顯然，就像用戶成長一樣，你不可能永遠提高用戶保留率。您也無法永遠提高用戶成長。但我們仍然認為那裡還有空間，所以我們對此感到滿意。

In terms of competition, there was the headline about Google. Google, of course, is a company that we all admire. I spent a couple of years at Google. A lot of us have spent time at Google. The particular feature that made headlines is we actually -- the team that has -- that's working on this feature, which is a small research team has been in contact with us. They are trying to do something to just make the search experience better for people who are in the language learning ecosystem. And one of the things that they want to do is start sending those users to apps like ours. So we don't see them really as competitive. And that's kind of one thing.

在競爭方面，頭條新聞是關於谷歌的。當然，谷歌是一家我們都欽佩的公司。我在谷歌工作了幾年。我們很多人都曾在谷歌工作過。成為頭條新聞的特定功能實際上是我們 - 團隊 - 正在研究此功能，這是一個小型研究團隊一直在與我們聯繫。他們正在嘗試做一些事情，讓語言學習生態系統中的人們獲得更好的搜尋體驗。他們想做的事情之一就是開始將這些用戶發送到像我們這樣的應用程式。所以我們並不認為它們具有真正的競爭力。這是一回事。

The other thing is -- and just for -- in general, for competition, we have a strong belief that the hardest thing about learning a language is staying motivated, and that is something that we really excel at. And we just don't see our competition spending much effort, and that's certainly not Google, but kind of the rest of our competition. And I think that's really probably the biggest thing that has made us grow a lot.

另一件事是——而且只是為了——總的來說，為了競爭，我們堅信學習語言最困難的事情就是保持動力，而這是我們真正擅長的事情。我們只是沒有看到我們的競爭對手付出太多努力，這當然不是谷歌，而是我們的其他競爭對手。我認為這可能是讓我們成長很多的最重要的事情。

It's because a lot of our users, we give them motivation to continue going. And that has done things, for example, we've quoted this stat before, in the U.S., 80% of our users were not learning a language before Duolingo. So the reason we're getting so many of these users is because we have made learning a language so easy and so engaging that they just come.

這是因為我們的許多用戶，我們給了他們繼續前進的動力。這已經起到了作用，例如，我們之前引用過這個統計數據，在美國，我們 80% 的使用者在 Duolingo 之前沒有學習過語言。因此，我們之所以能吸引如此多的用戶，是因為我們讓學習語言變得如此簡單、如此吸引人，以至於他們就來了。

And so to me, that's the biggest moat. I mean, we have other things. The streak is a big mode, for example, people don't want to lose their streak when they go to another app. That's a big moat. Our brand has essentially become synonymous with language learning, particularly with younger users.

對我來說，這是最大的護城河。我的意思是，我們還有其他事情。連勝是一種大模式，例如，人們不想在轉到另一個應用程式時失去連勝。那是一條大護城河。我們的品牌本質上已成為語言學習的代名詞，尤其是對年輕用戶而言。

So there are a lot of moats like that. But I would say the best one is just that we have a product that is more engaging, and it's getting more and more engaging the more users we have because we run more experiments. So that's kind of our view on it.

所以有很多這樣的護城河。但我想說，最好的就是我們擁有一個更具吸引力的產品，而且我們擁有的用戶越多，它就越有吸引力，因為我們進行了更多的實驗。這就是我們對此的看法。

Next question comes from Alex Sklar, Raymond James.

下一個問題來自 Alex Sklar、Raymond James。

Luis, first for you, this is a bit of a follow-up to Justin's question, and it's a bit early given you just launched music. But how do you think about the ideal content course footprint within the main app? Is more always better? Is there an optimal level of [choice] that you found? And as you think about kind of future organic or inorganic content, do you think it will always make sense now to keep it in the one primary app?

路易斯，首先對你來說，這是賈斯汀問題的後續，考慮到你剛剛推出了音樂，現在還為時過早。但是您如何看待主應用程式中理想的內容課程足跡？更多總是更好嗎？您是否找到了最佳的[選擇]水平？當您考慮未來的有機或無機內容時，您認為現在將其保留在一個主要應用程式中是否總是有意義？

I'm not sure I fully understand your question. I mean, in terms of more, you mean more subjects or more general kind of hours of content? I'm not sure I fully understand.

我不確定我是否完全理解你的問題。我的意思是，就「更多」而言，您是指更多的主題或更一般的內容時間？我不確定我是否完全理解。

More subjects, more languages or more expansions into things like math or music or et cetera.

更多科目、更多語言或更多擴展到數學或音樂等領域。

Yes. I mean, you're not going to see us expand in every single subject, at least not for the time being. I mean, in fact, you're not going to see us be adding more subjects other than math and music. Maybe we'll add a few other languages, yes, like foreign languages, but you're not going to see us add other subjects in the short to medium term.

是的。我的意思是，你不會看到我們在每個主題上都進行擴展，至少暫時不會。我的意思是，事實上，除了數學和音樂之外，你不會看到我們增加更多的科目。也許我們會添加一些其他語言，是的，例如外語，但你不會看到我們在中短期內添加其他科目。

We're very committed to making math and music succeed. I mean, we haven't even launched them. So you cannot possibly say they are successful yet in the main app because we haven't even launched them. So we're very committed to making them succeed.

我們非常致力於讓數學和音樂成功。我的意思是，我們甚至還沒有啟動它們。所以你不可能說它們在主應用程式中已經成功，因為我們甚至還沒有啟動它們。因此，我們非常致力於讓他們取得成功。

Our belief is that there are things that make a lot of sense to learn with the Duolingo app. They are usually things that you can learn on your own, that take a long time to learn where the thing that you're learning requires a lot of repetition. I mean, certainly a lot of the arithmetic and a lot of the parts of math, the way you learn them is through a lot of repetition.

我們相信，使用 Duolingo 應用程式學習一些東西非常有意義。它們通常是您可以自己學習的東西，需要很長時間才能學習，並且您正在學習的東西需要大量重複。我的意思是，當然很多算術和數學的很多部分，你學習它們的方式都是透過大量的重複。

That's true for music and that's also true for languages. So things that require a lot of repetition, that take a long time to learn and also that we think will have a very large audience. These are the types of things that we will be adding to the app. Although for the time being, we're going to stay just with math and music. Yes. I think that was the only part of your question, right?

對於音樂來說是如此，對於語言也是如此。因此，那些需要大量重複、需要很長時間才能學習的東西，而且我們認為會有大量的受眾。這些是我們將添加到應用程式中的內容類型。儘管目前我們將只專注於數學和音樂。是的。我認為這是你問題的唯一部分，對吧？

Yes. No, perfect. Great color there. And then maybe just as a follow-up. In terms of kind of the tracking of at-risk paid subscribers. What did you see in kind of the month after you announced math and music in the main app in terms of kind of the usage of that cohort?

是的。沒有十全十美的。那裡的顏色很棒。然後也許只是作為後續行動。在追蹤有風險的付費用戶方面。在您在主應用程式中宣布數學和音樂之後的一個月裡，您對該群體的使用有何看法？

Of people who are at risk?

有危險的人嗎？

Yes. Perhaps they're paid subscribers. They're using it less and kind of did that bring them back...

是的。也許他們是付費訂閱者。他們使用它的次數減少了，這確實讓他們回來了...

Yes. After we have -- I don't -- I actually don't know the answer to your question. I don't know if we've kept track. My sense is that if there had been a big change, I would have heard about it. So there probably wasn't a very big change. Although I am answering here -- I don't actually know the answer to your question.

是的。在我們——我不會——之後，我其實不知道你問題的答案。我不知道我們有沒有跟蹤。我的感覺是，如果發生重大變化，我會聽說的。所以可能沒有太大的變化。雖然我在這裡回答——但我實際上不知道你問題的答案。

Next question comes from Andrew Boone of JMP.

下一個問題來自 JMP 的 Andrew Boone。

Luis, can you provide us an update on Max? And what I really want to also dig in on for this is just anyone that watched OpenAI's Dev Day and understood your partnership there. But text-to-speech is just -- it feels like it's coming. And so talk to us about, a, what prevents that from coming to Duolingo? And how does Max start to evolve as just more of these tools become available?

Luis，您能為我們提供 Max 的最新情況嗎？我真正想深入探討的是任何觀看過 OpenAI 開發日並了解你們在那裡的合作夥伴關係的人。但文字轉語音只是——感覺它即將到來。那麼請和我們談談，a，是什麼阻止了它出現在 Duolingo 上？隨著越來越多的這類工具的出現，Max 是如何開始發展的呢？

Well, thank you for asking about Max. Okay. So just to provide color about Max. So we announced Max earlier this year. When we announced it, it is a higher-tier subscription. So we had free Duolingo, Super Duolingo and then Max. What we put into Max in this higher tier were 2 features based on generative AI and in particular, OpenAI generative AI. One is Roleplay and the other one was Explain My Answer.

嗯，謝謝你詢問關於麥克斯的事情。好的。所以只是提供有關麥克斯的顏色。所以我們在今年早些時候宣布了 Max。當我們宣布它時，這是一個更高級別的訂閱。所以我們有免費的 Duolingo、Super Duolingo，然後是 Max。我們在這個更高層中為 Max 添加了 2 個基於生成式 AI 的功能，特別是 OpenAI 生成式 AI。一個是角色扮演，一個是解釋我的答案。

What we've been doing -- and when we announced it, we said, look, this is a large feature. Usually, for large features, it takes us about a year to roll it out to all our users because in that time, what we're doing is we're giving it to a fraction of our users, typically at first, a very small fraction of our users. We improve the feature, and then we give it to a larger and larger fraction, et cetera.

我們一直在做的事情——當我們宣布它時，我們說，看，這是一個很大的功能。通常，對於大型功能，我們需要大約一年的時間才能將其推廣給所有用戶，因為在那段時間，我們正在做的是將其提供給一小部分用戶，通常是一開始的一小部分用戶。我們的用戶的一小部分。我們改進該特徵，然後將其賦予越來越大的部分，等等。

By now, we're kind of halfway-ish from this. We have -- we've increased the rollout by now. A significant fraction of our users on iOS that are learning either French or Spanish have access to Max, as in at least can buy it, so we present it to them. And the reason we've increased the rollout is because we feel good about the features themselves. We've improved both Roleplay and Explain My Answer to the point where people are using them more, et cetera. So that's kind of one thing.

到目前為止，我們已經完成了一半。我們現在已經增加了部署。我們的 iOS 用戶中有很大一部分正在學習法語或西班牙語，可以使用 Max，因為至少可以購買它，所以我們將其提供給他們。我們增加推出的原因是因為我們對這些功能本身感覺良好。我們改進了角色扮演和解釋我的答案，以便人們可以更多地使用它們，等等。所以這是一回事。

Another thing that I think is important to mention is, when we announced Max, we said, well, we're putting these AI features in a higher-tier subscription, in part because we have to pay for them. But we also said, we believe that the cost of using this -- or putting these features on our app is actually going to go down because accessing large language models like OpenAI's, the price is going to go down, and we have already seen that. I mean, this is -- the price is actually going down.

我認為值得一提的另一件事是，當我們宣布 Max 時，我們說，我們將這些人工智慧功能放在更高層級的訂閱中，部分原因是我們必須為它們付費。但我們也說過，我們相信使用此功能的成本- 或者將這些功能放在我們的應用程式上實際上會下降，因為訪問像OpenAI 這樣的大型語言模型，價格將會下降，而且我們已經看到了這一點。我的意思是，價格實際上正在下降。

So what you're going to see us do over the next few months is -- I think it's important to start talking separately about the fact that we have a 3-tier subscription and AI features. Because what you're going to see us do is we're probably going to be shuffling features to see -- and experimenting each time, of course, to see what is the best place to put our features.

因此，您將在接下來的幾個月中看到我們所做的是——我認為開始單獨討論我們擁有三層訂閱和人工智慧功能這一事實很重要。因為你將看到我們所做的是，我們可能會重新排列要查看的功能，當然，每次都會進行實驗，看看什麼是放置我們的功能的最佳位置。

So it may be the case that Explain My Answer goes into Super as opposed to Max. And then we take some features from Super and put them into Max. So you're going to see us do that. And the idea is really to maximize the revenue that we can make and also to maximize the amount of features that we give to our users or value that we give to our users in a 3-tier strategy without necessarily saying, "Well, the highest tier is AI."

因此，“解釋我的答案”可能屬於“超級”，而不是“麥克斯”。然後我們從 Super 中獲取一些功能並將它們放入 Max 中。所以你會看到我們這樣做。我們的想法實際上是最大化我們可以賺取的收入，並最大化我們為用戶提供的功能數量或我們在三層策略中為用戶提供的價值，而不必說，「嗯，最高的第一層是人工智慧。”

On the other side, for AI, we're going to continue developing these features because we're very happy with them. And so that's kind of what we're going to see. We don't have anything to say specifically yet about how much contribution that the 3-tier system is going to have, for example, for 2024 bookings, because we're -- there's just -- there's a lot of stuff -- there are a lot of moving parts. But we knew this, that when we introduced the 3-tier system, it was going to take us a while to figure this out.

另一方面，對於人工智慧，我們將繼續開發這些功能，因為我們對它們非常滿意。這就是我們將要看到的情況。我們還沒有具體說明 3 層系統將有多少貢獻，例如，對於 2024 年的預訂，因為我們 - 只是 - 有很多東西 - 那裡有很多活動部件。但我們知道，當我們引入三層系統時，我們需要一段時間才能弄清楚這一點。

Your next question was about text-to-speech, which is basically speech synthesis from OpenAI. We have our own speech synthesizers. We're very happy because we made our voices for our own characters. So for now, we're probably going to continue with that, because each one of our characters has its own voice, et cetera. Over time, if we see that OpenAI speech synthesis it's very good. Of course, we may evaluate it.

您的下一個問題是關於文字轉語音，這基本上是 OpenAI 的語音合成。我們有自己的語音合成器。我們很高興，因為我們為自己的角色配音。所以現在，我們可能會繼續這樣做，因為我們的每個角色都有自己的聲音，等等。隨著時間的推移，如果我們看到 OpenAI 語音合成，那就非常好。當然，我們可以對其進行評估。

I'll use my follow-up to follow up on that. In terms of just their capabilities though, right, like you clearly see somebody else that is now become a conversational in terms of language. And so how do we port that progress back to Duolingo?

我將用我的後續行動來跟進這一點。就他們的能力而言，是的，就像你清楚地看到其他人現在在語言方面已經成為對話者一樣。那我們該如何將這項進展移植回 Duolingo 呢？

Sorry, I'm not quite sure I understand.

抱歉，我不太確定我是否理解。

Within ChatGPT, it feels like it's becoming more conversational as you use it, sort of within like the premium tiers, right? And so the question is, how do we think about language learning starting to become conversational and really more advanced in terms of one-on-one tutoring?

在 ChatGPT 中，當您使用它時，感覺它變得更加會話化，有點像高級級別，對吧？所以問題是，我們如何看待語言學習開始變得對話式，並且在一對一輔導方面真正變得更加先進？

Oh, yes. We are definitely working on features that are related to that. You will see us -- you will see -- I mean, I don't have anything specific to announce, but we are definitely working on features that will feel a lot more like that, and you'll see us in the next few months start experimenting with them.

哦是的。我們肯定正在開發與此相關的功能。你會看到我們——你會看到——我的意思是，我沒有什麼具體要宣布的，但我們肯定正在開發感覺更像這樣的功能，你會在接下來的幾年中看到我們幾個月開始嘗試它們。

Thanks, Andrew. And the next question comes from Arvind Ramnani at Piper with an awesome Zoom background.

謝謝，安德魯。下一個問題來自 Piper 的 Arvind Ramnani，他擁有出色的 Zoom 背景。

I had to kind of compete with Justin, right? But -- so I guess my first question is you have this very creative intros when you open up your prepared remarks. At what point do you take the group of analysts out here and you convert this into a whole video? We've been showing up for your conference calls. you had Meghan Markle last quarter, when do we get our time?

我必須與賈斯汀競爭，對嗎？但是——所以我想我的第一個問題是，當你打開準備好的演講時，你有這個非常有創意的介紹。您什麼時候將分析師小組帶到這裡並將其轉換為完整的影片？我們一直在參加您的電話會議。上個季度你有梅根馬克爾，我們什麼時候有時間？

We're working on it, Arvind.

我們正在努力，阿爾溫德。

Wait, wait, wait. Arvind, you want to be in one of our TikToks? Is that what you're saying?

等等，等等，等等。 Arvind，你想加入我們的 TikTok 之一嗎？你是這麼說的嗎？

Yes. Yes. There's no shame in that ask.

是的。是的。這個問題並沒有什麼可恥的。

We'll throw that in the hopper, Arvind. We'll throw that in there.

我們會把它丟進料斗裡，阿爾文德。我們會把它丟在那裡。

Perfect. So I wanted to ask about your product development roadmap, right? Like, there are many ways you can go. You have -- you take your core language and you come up with Duolingo Max, DuoRadio, and you have that kind of enhancing the core sort of language. And then, of course, like math and music, it's another very interesting vector. How are you sort of, like, making the choices between, like, product enhancements on your core versus kind of expansion?

完美的。所以我想問一下你們的產品開發路線圖，對吧？就像，你可以走很多路。你採用了你的核心語言，並提出了 Duolingo Max、DuoRadio，並且你可以對核心語言進行這種增強。當然，就像數學和音樂一樣，這是另一個非常有趣的向量。您如何在核心產品增強與擴充之間做出選擇？

Yes, an excellent question. The majority of the effort is going into the core. We believe -- language learning is a very big business that we are only scratching the surface on, and that is where we are, the category leaders. We are not yet the category leaders for math or music education. We are the category leader for language education. So we're spending the majority of effort on that.

是的，這是一個很好的問題。大部分工作都集中在核心方面。我們相信，語言學習是一項非常大的業務，我們只是觸及了表面，而這就是我們所處的類別領導者的位置。我們還不是數學或音樂教育領域的領導者。我們是語言教育領域的領導者。所以我們在這方面投入了大部分精力。

That said, what's nice about putting math and music into the main app is that a lot of the changes that we make to improve the core immediately helped math and music. So for example, we may make a change to make the leaderboards more exciting. Well, math and music are already in the app, so it'll immediately go into math and music.

也就是說，將數學和音樂放入主應用程式的好處在於，我們為改進核心所做的許多更改立即對數學和音樂有所幫助。例如，我們可能會做出一些改變，讓排行榜更加令人興奮。嗯，數學和音樂已經在應用程式中，所以它會立即進入數學和音樂。

So most of the effort goes into the core, but some of that goes into math and music. And then just to give you an idea, math and music, I mean the total number of people working on math and music is, like, 50, which is a small fraction of our whole product team. So that's just to give you an idea.

因此，大部分精力都投入到核心方面，但其中一些投入數學和音樂。然後只是為了給您一個想法，數學和音樂，我的意思是從事數學和音樂工作的人員總數約為 50 人，這只是我們整個產品團隊的一小部分。所以這只是給你一個想法。

Great. And I guess, like, are there any gating factors in terms of kind of investing? Is it like you're trying to balance, like, profitability that -- kind of Matt's kind of keeping the purse strings a bit tight to kind of say, like, hey let's get probability. And is it -- or is it more like the people -- like, what's the gating factor that will allow you to kind of push on both?

偉大的。我想，在投資方面是否有任何限制因素？是不是你在試圖平衡獲利能力——馬特有點勒緊錢包，比如，嘿，讓我們看看機率。是——或者更像是人們——比如，讓你推動兩者的控制因素是什麼？

That's a great question, Arvind. The truth of the matter is that the gating factor is, I don't know why. I cannot explain this, but the gating factor for any kind of product development ends up being time, is this funny thing where it takes like 2 years to make a good product. I've never been able to see something goes faster.

這是一個很好的問題，阿爾溫德。事情的真相是，門控因素是，我不知道為什麼。我無法解釋這一點，但任何類型的產品開發的控制因素最終都是時間，這是一件有趣的事情，需要大約 2 年的時間才能製造出一個好的產品。我從來沒見過有什麼事情進展得更快。

And it doesn't help if you throw a ton of people into it because they just kind of step on each other's toes. The right way, I think, to grow a product is to start with a small, and when the product is gaining traction, you put more and more people. That is the right way to grow a product.

如果你把很多人投入其中，那也無濟於事，因為他們只是互相踩到對方的腳趾。我認為，發展產品的正確方法是從小規模開始，當產品越來越受歡迎時，你就會投入越來越多的人。這才是發展產品的正確方法。

And so we're not -- it's interesting, I mean we really do we have increased profitability quite a bit. But in the day-to-day conversations, we don't find ourselves being like, "Oh, my God, I wish I had 30 more people to throw at this problem." I mean it's more just getting our act together on, really, what it means to scale the content for math, what exactly do we want to teach for math next, and usually throwing more people at this just has not helped us in the past. So that -- the gating factor is just -- it's kind of like it has to be a certain time in the oven. It's like you can't cook a cake much faster or a bigger cake much faster. That's it.

所以我們不是——這很有趣，我的意思是我們確實提高了盈利能力。但在日常對話中，我們不會發現自己會說：“天哪，我希望能再有 30 個人來解決這個問題。”我的意思是，更多的是讓我們共同行動，真正地，擴展數學內容意味著什麼，下一步我們到底想教什麼數學，通常在這方面投入更多的人在過去對我們沒有幫助。所以——門控因素就是——有點像它必須在烤箱裡待一段時間。就像你無法更快地煮蛋糕或更快地煮更大的蛋糕一樣。就是這樣。

Great. Next question comes from Mark Mahaney at Evercore.

偉大的。下一個問題來自 Evercore 的 Mark Mahaney。

Two questions, please. First, I know you've had a lot of questions on math and music. So maybe something a little bit more niche. There was some news article about chess, and it just made us think about there's -- what about skilled -- games of skill, skill categories like that? Long term, is that -- is there an interesting appeal to including those somehow in the app?

請教兩個問題。首先，我知道您對數學和音樂有很多疑問。所以也許有些更小眾的東西。有一些關於國際象棋的新聞文章，它讓我們思考——技能遊戲怎麼樣？諸如此類的技能類別？從長遠來看，將這些內容以某種方式納入應用程式是否有有趣的吸引力？

And then secondly, this bookings growth and the MAU and the sub growth is super impressive. But could you give us some color on where these new MAUs, these kind of almost record levels of MAUs and max record levels of new subs, where they're coming from? Is this -- are you just getting a flood of new people coming in from China or from India? Just help us realize what the source of these new subs and MAUs is particularly from a geographic perspective?

其次，預訂量成長、每月活躍用戶數和次成長非常令人印象深刻。但是您能否告訴我們這些新的每月活躍用戶數、這些幾乎創紀錄的月活躍用戶數和新用戶的最高創紀錄水平是從哪裡來的？這是──你們剛剛迎來大量來自中國或印度的新人嗎？請幫助我們了解這些新訂閱者和每月活躍用戶的來源是什麼，特別是從地理角度來看？

Yes. Okay. Well, in terms of adding content to keep things like chess, things like games, you're unlikely -- for one, you're unlikely to see us do that. For one, in the next -- again, in the short to medium term, we're not going to be adding more subjects other than math and music because we want to make those really succeed.

是的。好的。嗯，就添加內容來保留象棋、遊戲等內容而言，你不太可能——首先，你不太可能看到我們這樣做。首先，在接下來的中短期內，我們不會增加數學和音樂以外的更多科目，因為我們希望這些科目能真正成功。

But in the long term, just me personally, and I think that's true for a lot of people that work at Duolingo, we're just a lot more interested in pure education. And so that's -- there's nothing wrong with teaching how to play chess or games or anything like that. I love that. I love chess, but it's just -- I think we're a lot more interested in just kind of pure education. That's probably what you're going to see. For the second question, I'll let Matt take that one.

但從長遠來看，就我個人而言，我認為對於在 Duolingo 工作的許多人來說都是如此，我們只是對純教育更感興趣。因此，教如何下棋或遊戲或類似的事情並沒有錯。我喜歡那個。我喜歡國際象棋，但我認為我們對純粹的教育更感興趣。這可能就是您將會看到的。對於第二個問題，我會讓馬特回答。

Yes. So as I mentioned in my prepared remarks, Mark, the beauty of our growth has been broad-based. So when you look at kind of the countries with the biggest MAU or DAU around the world, and you try to see user growth has been -- DAU growth has been around 60% for the past several quarters, what's the variation? Like, is the U.S. growing fast? It's growing close to average. And then there are other companies that are growing above average, but it's broad-based. It's around the world. So the user growth is really -- it's super geographically diverse, and we're glad about that.

是的。正如我在準備好的發言中提到的，馬克，我們的成長之美是基礎廣泛的。因此，當您觀察全球 MAU 或 DAU 最大的國家時，您會嘗試查看用戶成長情況——在過去幾個季度中，DAU 成長一直在 60% 左右，其中的變化是什麼？比如，美國成長快嗎？它的增長接近平均。還有其他公司的成長高於平均水平，但基礎廣泛。它遍布世界各地。因此，用戶的成長確實在地域上非常多樣化，我們對此感到很高興。

In terms of subscribers, we look at where our subscriber growth is coming from, and it hasn't changed. I mean, every quarter, the stack rank of countries kind of fluctuates every so little bit. But it's not really fundamentally changed over the past couple of quarters. So that's why we mentioned we felt like the user growth has been broad-based and high quality, so we feel good about it.

就訂閱者而言，我們關注訂閱者成長的來源，這一點沒有改變。我的意思是，每個季度，國家的排名都會略有波動。但在過去的幾個季度裡，這種情況並沒有真正發生根本性的改變。這就是為什麼我們提到我們覺得用戶成長基礎廣泛且高品質，所以我們對此感覺良好。

And then maybe one question just on the level of investments going forward. You're obviously ramping up margins very aggressively. That sounds -- that is fundamentally a good thing. Just kind of addressed like the -- is there any risk that you just have a lack of investment opportunities? I know that doesn't sound right, but…

然後可能還有一個關於未來投資水準的問題。顯然，你們正在非常積極地提高利潤率。聽起來——這從根本上來說是一件好事。就像這樣解決──是否存在缺乏投資機會的風險？我知道這聽起來不對，但…

No, no. I…

不，不。我...

How do you -- how are you balancing -- like, you're ramping up profitability. How are you balancing that with kind of the need for whatever to invest in Duolingo for the next 3 to 5 years?

你如何——你如何平衡——比如，你正在提高獲利能力。您如何平衡這一點與未來 3 到 5 年內投資 Duolingo 的需求？

Yes. I think it's great question. I'm sure Luis has some thoughts as well. But just from the raw mechanics of the math, we mentioned on the last call that I think it's -- if you look at other tech companies that -- certainly ones that we respect and that you guys follow, in their first year of real profitability, they typically see a pretty big jump in profit margin. And then they see steady progress towards their long-term margin. That's kind of what we expect. As we said on the last call, it's what we expect on this call.

是的。我認為這是一個很好的問題。我相信路易斯也有一些想法。但僅從數學的原始機制來看，我們在上次電話會議中提到，我認為——如果你看看其他科技公司——在真正盈利的第一年，我們肯定會尊重並且你們會追隨的公司，他們通常會看到利潤率大幅上漲。然後他們就會看到長期利潤穩定進步。這正是我們所期望的。正如我們在上次電話會議中所說，這也是我們對這次電話會議的期望。

And while we've been doing this path towards increased profitability this year, we've been hiring, and we've been investing in the product and the business. So I think we feel good about our balancing long-term growth with long-term profitability over time. And so I think Luis mentioned on the last -- one of the last questions, the gating factor so far has not been investment in our product cycle, our product journey. I don't know, Luis, if you have anything else to add to that?

雖然我們今年一直在採取這條提高盈利能力的道路，但我們一直在招聘，並且一直在投資產品和業務。因此，我認為隨著時間的推移，我們對平衡長期成長和長期獲利能力感到滿意。因此，我認為路易斯在最後一個問題中提到了最後一個問題，到目前為止，限制因素並不是對我們的產品週期、我們的產品旅程的投資。路易斯，我不知道你還有什麼要補充的嗎？

Matt will not let us eat lobster every day anymore, and we're upset about that. I'm kidding.

馬特不再讓我們每天吃龍蝦，我們對此感到不安。我在開玩笑。

We've always adhered, in the business, with the cost discipline it turns out.

在業務中，我們始終堅持成本紀律。

I'm kidding. No, I mean just generally, I don't think that we've been really -- there has never been a product initiative that we haven't done because we can't hire the people. There's been -- we have not done it because we don't think it's a good idea or because we don't think we're ready to do it, et cetera, so I think we keep investing really well, at least in R&D.

我在開玩笑。不，我的意思是一般來說，我認為我們從來沒有真正做過任何產品計劃，因為我們無法僱用人員。我們還沒有這樣做，因為我們認為這不是一個好主意，或者因為我們認為我們還沒有準備好這樣做，等等，所以我認為我們繼續進行很好的投資，至少在研發方面。

Great. Next question comes from Curtis Nagle at BofA.

偉大的。下一個問題來自美國銀行的柯蒂斯·內格爾。

So Luis, I thought you made kind of an interesting point in terms of the free-to-paid conversion in some of the regions. I think it was last call, you talked about how it can be very difficult to drive that, whether you're lower prices or whatever you're trying to do. So obviously, something changed. So it'd be great just to hear a little bit more about, I guess, how you cracked that egg and just what led to some success in this quarter.

路易斯，我認為您在某些地區的免費到付費轉換方面提出了一個有趣的觀點。我認為這是最後一次通話，您談到如何推動這一目標非常困難，無論您是降低價格還是無論您想要做什麼。很明顯，有些事情發生了變化。因此，我想，如果能更多地了解您是如何打破這個雞蛋以及是什麼導致了本季度的一些成功，那就太好了。

Yes. So Curtis, I think I'm the one who said that in my prepared remarks. And the answer is going to be one that you're going to get used to us hearing, which is -- it's actually not one thing. It's a bunch of experiments that we've been running around the world to understand how that free-to-pay conversion can increase.

是的。柯蒂斯，我想我就是在準備好的發言中說了這句話的人。答案將是一個你會習慣於我們聽到的答案，那就是——這實際上不是一回事。我們在世界各地進行了一系列實驗，以了解如何提高免費到付費的轉換率。

In this particular quarter, one of the things we did was we optimized how we're showing what we call the hook to go from free to paid, so the kind of internal super ad that says, "Would you like to try Duolingo?" We optimize that in a bunch of markets around the world, changed the language a little bit, made it more local, more colloquial, and then that was a big win. But again, that's one of many wins across the quarter. So that's really what drove it over the past kind of 3, 4 months. Luis, I don't know if you have anything else.

在這個特定的季度，我們所做的一件事是我們優化瞭如何展示我們所說的從免費到付費的鉤子，所以那種內部超級廣告說，“你想嘗試 Duolingo 嗎？”我們在世界各地的許多市場中對其進行了優化，稍微改變了語言，使其更加本地化、更加口語化，然後這是一個巨大的勝利。但這又是本季眾多勝利之一。這就是過去三、四個月的推動因素。路易斯，我不知道你還有什麼事嗎？

Yes, that's it. I mean, it's just a generally a lot of -- we have teams exactly dedicated to getting more people to subscribe.

對，就是那樣。我的意思是，我們有專門致力於讓更多人訂閱的團隊。

Okay. Makes sense. And then, Matt, just a quick follow-up. So if I heard correctly, dilution, I think 1 to 1.5.

好的。說得通。然後，馬特，快速跟進一下。所以如果我沒聽錯的話，稀釋的話，我認為是 1 到 1.5。

Right.

正確的。

I think that was brought down by 2. So just, I guess, what's the delta?

我認為這個數字下降了 2。那麼，我想，增量是多少？

Yes, it's come down, in part because we -- when we made our plans for the full year and guided to that 2% range. We'll probably hire -- end up hiring a couple of fewer people this year. But then the other thing is that's on a treasury stock method. So as the share price goes up, the amount of shares we grant over the course of the year goes down. So those are the 2 factors, nothing crazy.

是的，它有所下降，部分原因是我們——當我們制定全年計劃並指導到 2% 的範圍時。我們可能會招募——最終今年會少招募幾個人。但另一件事是庫存股方法。因此，隨著股價上漲，我們在一年中授予的股票數量就會減少。這就是兩個因素，沒什麼瘋狂的。

Your next question comes from Chris Kuntarich, UBS.

您的下一個問題來自瑞銀集團 (UBS) 的 Chris Kuntarich。

Great. Maybe just a bit of a follow-up on what was just asked there on the regional payer penetration. Just curious how we should be thinking about, like, really how much of this is kind of rinse and repeat on a market-by-market basis versus kind of what needs to be done is really kind of the experimentation and it's unknown kind of really what the payer penetration is going to look like in some of these more nascent markets. And just curious how we should be thinking about across that 8% payer penetration. Like, where are some of these earlier markets that we're at on a payer penetration versus some of our more advanced? And then maybe just one follow-up after.

偉大的。也許只是對剛剛詢問的地區支付者滲透率問題的一些後續行動。只是好奇我們應該如何思考，例如，到底有多少是在逐個市場的基礎上進行沖洗和重複，而不是需要做的事情實際上是一種實驗，而且真的是未知的在一些較新興的市場中，付款人的滲透率會是什麼樣子。只是好奇我們應該如何考慮 8% 的付款人滲透率。例如，與一些更先進的市場相比，我們在付款人滲透率方面處於哪些早期市場？然後也許只是一次後續行動。

I'm happy to -- I think it's a little bit of both. I think some of it is -- look, there are some experiments that we're running that help payer penetration in every single country. It's just sometimes, talking about -- the exact language that we use is better, and we find better language to use that help the whole world. That said, there's a big difference in payer penetration in a country -- in a wealthy country like the U.S. versus a country like India. And that's not just true for Duolingo, this is true for essentially every product out there.

我很高興——我認為兩者兼而有之。我認為其中一些是——看，我們正在進行一些實驗來幫助付款人滲透到每個國家。只是有時，我們使用的確切語言更好，並且我們找到更好的語言來幫助整個世界。也就是說，一個國家的支付滲透率有很大差異──美國這樣的富裕國家與印度這樣的國家。這不僅適用於 Duolingo，幾乎適用於所有產品。

So there are some things that you're probably going to see us try there. It's not just decreasing the price, which we've already done, but there are things like, well, maybe we sell more through the family plan. That's just an example, or maybe we sell more through in-app purchases or something, because we do know that for some of these markets, some slight differences are just needed, so I think it's a bit of both.

因此，您可能會看到我們在那裡嘗試一些事情。這不僅僅是降低價格（我們已經做到了），而且還有一些事情，例如，也許我們可以透過家庭計劃銷售更多產品。這只是一個例子，或者也許我們透過應用程式內購買或其他方式銷售更多，因為我們確實知道對於其中一些市場來說，只需要一些細微的差異，所以我認為兩者兼而有之。

My -- if I had to guess, my sense is that we're going to do more of the global stuff than the specific -- the market specific. It's kind of -- it takes a lot of effort to do a lot of market-specific stuff. We may do market-specific stuff for very large countries like China and India. You won't see us do market-specific stuff for countries kind of smaller, like, I don't know, the Czech Republic or something like that, which is a significantly smaller country.

我的——如果我必須猜測的話，我的感覺是，我們將做更多的全球性的事情，而不是特定的——特定市場的事情。這有點——需要付出很大的努力才能做很多針對特定市場的事情。我們可能會為像中國和印度這樣的大國做針對特定市場的事情。你不會看到我們為較小的國家做針對特定市場的事情，例如，我不知道，捷克共和國或類似的國家，這是一個小得多的國家。

Got it. Very helpful. And just the follow-up would be on -- I think you had talked last quarter that revenue per sub would be effectively flat. And I think you're now talking about improving, should we be taking that as assuming that we're down kind of low single digits as far as revenue per subscriber? Or yes, just any more color you could provide around that would be helpful.

知道了。很有幫助。接下來就是——我想你上個季度曾說過，每個子項目的收入實際上將持平。我認為您現在正在談論改進，我們是否應該假設我們的每位訂戶收入下降了較低的個位數？或者是的，您可以提供更多的顏色，這會有所幫助。

Yes. We -- it's the same general trend. So this quarter, we thought it would be slightly less of a decline on ARPU year-over-year. And it would have been, but for the fact that we found this other lever around the world that drove mix shift outside the U.S. And that was a trade-off we were happy to make because it drove LTV positive subscription, so that was the right decision. It probably delayed the lapping or the kind of 0% year-over-year growth by a quarter or so, but it's still the general trend. The same trend as was on the last call.

是的。我們——這是相同的整體趨勢。因此，本季度，我們認為 ARPU 年比下降幅度會略小。本來會是這樣，但事實上我們在世界各地發現了另一個推動美國以外地區混合轉型的槓桿，這是我們很樂意做出的權衡，因為它推動了 LTV 的積極訂閱，所以這是正確的決定。它可能將研磨或0％的同比增長推遲了四分之一左右，但這仍然是總體趨勢。與上次通話的趨勢相同。

Great. I'm not showing any further questions. So I'm going to turn it back to you, Luis, to wrap up.

偉大的。我不會提出任何進一步的問題。所以我要把它轉回給你，路易斯，來結束。

Thank you, everyone, and thank you for the great questions, and we will see you next quarter.

謝謝大家，也謝謝你們提出的好問題，我們下季再見。