Duolingo Inc (DUOL) 2021 Q3 法說會逐字稿

Good day, and welcome to the Duolingo Third Quarter 2021 Earnings Call. (Operator Instructions)

Please note, this event is being recorded.

I would now like to turn the conference over to Debbie Belevan, Head of Investor Relations. Please go ahead.

Thank you, operator, and welcome to Duolingo's third quarter earnings call. Today, we released our third quarter results after the market close. You can find our press release and the accompanying presentation on our IR website at investors.duolingo.com.

With me on the call today are Luis von Ahn, Co-Founder and Chief Executive Officer; and Matt Skaruppa, Chief Financial Officer. After our prepared remarks, we will open the call to Q&A.

First, I'd just like to remind everyone that during this call, we'll make forward-looking statements regarding our future events and financial performance, which are subject to material risks and uncertainties that could cause actual results to differ materially. We caution you to review the risk factors contained in our SEC filings. These forward-looking statements are based on assumptions that we believe to be reasonable as of today, and we undertake no obligation to update these statements as a result of new information or future events. Additionally, we will 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 with that, I'd now like to turn the call over to Luis.

Thank you, Debbie, and welcome, everyone. I want to start this call with the same topic that I start with in many of my presentations about the company, our mission, which is to develop the best education in the world and make it universally available. When we launched Duolingo 10 years ago, we decided to start with teaching languages because we believe we can impact an enormous number of people's lives by helping them learn a new language to get a better job or getting to school or even just communicate better on a trip.

But you'll notice that we don't say in our mission that we want to develop the best language education in the world. We say we want to develop the best education in the world. So we don't plan on stopping with just languages. We made strong progress on our mission this past quarter, and we have exciting plans for the future. As I said in our S-1, I plan to devote my life to this mission.

Now with that, let's get into the results of our third quarter. This was an exceptional quarter for us. We not only went public, but we also delivered record results. As you can see on our highlights page, we continued to grow our users, our paid subscribers, our bookings and our revenue, hitting record numbers in all these metrics. Our strong MAU and DAU growth was driven by continued organic word of mouth, some strategic investment in performance marketing and brand awareness campaigns and a relentless focus on making our product more fun and engaging.

Our monthly and daily active users grew 13% and 16%, respectively, over the prior year, which was a period when many parts of the world were still locked down due to COVID. We also grew bookings 57% year-over-year as more learners purchased Duolingo Plus, our subscription product. Because of that, we ended the quarter with 2.2 million paid subscribers, a 49% increase year-over-year. Our subscribers now represent 5.5% of our monthly active users. I'm also pleased to report we exceeded our revenue and bookings estimates, and as a result, have increased our financial year 2021 guidance. Matt will cover the financials in further detail as well as our updated guidance.

But now I'd like to share with you some highlights for the quarter, specifically the progress we're making towards our strategic initiatives. I'll start by discussing the progress we've made in growing users. About 95% of our monthly active users, or MAUs, are learning completely for free, while just over 5% of them pay for a Plus subscription, which removes ads and offers additional features to help people learn faster. Our monthly active users have been growing steadily since we started the company. And this growth is overwhelmingly organic. Our very high-quality mobile product that is free, fun and effective, leads people to telling their friends about it, has made our apps synonymous with language learning, and has driven our consistent growth. By now, we have nearly 42 million monthly active users.

This chart also shows the impact of the COVID lockdowns last year, which drove a spike in new users. This quarter, we achieved the highest ever MAUs in our history, even higher than Q2 of 2020 when we saw this surge in activity. As we mentioned on the last earnings call, we believe we are back on trend with our pre-COVID growth rates as the world has been opening up, and we think that this quarter's user growth further confirms that.

I'd also like to highlight the seasonality of our user growth. You can see from the highlighted sections of the chart, the areas with green circles on them, that the past 4 years have the same trend. In November and December, our MAUs declined during the busy holiday season. And then because of New Year's resolutions and people's outlook for the new year, we see a large spike in users in January. Those are the parts of the graph that have a steep slope and [N] in the center of each dark green circle. During this time, we also run our New Year's promotion, which is primarily advertised in the app and offers a discounted price for an annual subscription. The timing of this promotion aligns well with the season of resolutions as the New Year inspires people to set personal aspirational goals for the coming year, like, for example, learning a new language.

It is a great fit between how users are thinking about themselves, their lives and our brand, which is one of enriching your life through language learning. The New Year is the only time of the year we offer a discount. Annual subscribers can also renew each year at the discounted price. So this New Year's cohort is particularly sticky in terms of renewals. After the January spike, we see a general downward trend in MAUs until the summer months where usage flattens out and then rebounds in the fall.

Our daily active users, or DAUs, follow a similar seasonal path, and we achieved a record high this quarter. Internally, we are much more focused on DAUs because learning a language requires consistency and daily practice, and users that come back to our app every day are more likely to achieve their language goals. This is why we constantly focus on driving daily engagement, and DAUs is the key metric we use to measure the effectiveness of that focus.

On the topic of seasonality, we thought it would be helpful to provide an overview of the various marketing campaigns we have throughout the year that contribute to the variability in user and bookings trends. Our biggest and most impactful is our New Year's campaign, which I just went through. But we run other nondiscount campaigns in other parts of the year.

Our IPO in July, while not an annual event, was very positive given its high visibility, and we believe it drove some increase in users. Another one of my favorite initiatives, which drives awareness and user growth is our annual Duocon event, the largest gathering of language lovers from all over the world. We held our third annual Duocon conference in August, which was a huge success, drawing a record 350,000 viewers and driving valuable impressions across multiple media channels. The goal of Duocon is to provide informative talks by feature language experts, highlight exciting new developments on our products and provide a behind-the-scenes look at the talent and technology that makes Duolingo successful.

Next year, we hope to have a hybrid format and hold Duocon in person once again, this time in Brooklyn. Stay tuned for updates. And rounding out the year, we will send out our year-in review in December, which is a personalized summary of our learners' accomplishments during 2021. Because it drives a lot of buzz and social media sharing, it helps create word-of-mouth growth. Last year, duo365, the hashtag we set for the year-in review, became a trending topic on Twitter and generated a nice increase in DAUs. This year, we look to build on this success by expanding our reach through influencers and other targeted PR.

Another important strategic initiative for us is to teach better, and we are making important strides in this area. As I've said before, a surface-level look at Duolingo can be misleading because the app looks so cute and gamified that many people don't realize the amount of sophistication that is in the background and how much personalization there is in the learning experience.

We know every exercise you've done on Duolingo, whether you got it right or wrong. And if you got it wrong, we know the most likely piece of practice you need to learn from your mistake. And we use that data to make a model for each user. So when you start a lesson on Duolingo, we use artificial intelligence to pick exercises that are just right for you. And this leads to very positive learning outcomes.

In the previous study, we compared Duolingo's effectiveness versus traditional university language courses. We ran the study for people learning French and Spanish, and we found that Duolingo learners who complete unit 5, about halfway through the French or Spanish course, are as proficient in reading and writing as a person who has taken 4 semesters of university classes. In a new study we just released this quarter, we found that 7 units of Duolingo is equivalent to 5 university semesters. And like our previous study, the Duolingo units can be accomplished in half the time than the university semesters, with an app, entirely for free. This is something we're very proud of.

In addition to improving efficacy, we're constantly working to make the product more engaging. If you use the app, you're already familiar with our diverse cast of characters who guide learners throughout their lessons and cheer them on. Initially, these characters all had the same computerized voice. But we've been making great progress using text-to-speech technology and giving these characters unique voices and personalities in various languages, and that way, bringing them to life. We think of this as an investment in a brand asset as well. Since our cast of characters can make appearances in other apps we develop from Duolingo ABC, where they are already part of the experience, to our upcoming math app.

As an example, we just finished production of our first in-app animated short with 2 of our characters explaining the custom of giving (foreign language) or kisses in Spanish-speaking countries. We rolled this out on our social channels last month. And now I'd like to play it for you to give you a sense of how we're bringing our characters to life by not only teaching language but also culture.

(presentation)

I hope you enjoyed that. We plan to roll out more of these animated shorts within the app in coming months. The idea is that this is how we are going to explain culture the Duolingo way. Instead of making learners read through boring text, our characters will explain the culture of each language through these fun animated videos. Another initiative we've been focused on is making Duolingo more social, because learning with friends is more fun, especially when friends can cheer each other on.

Increasing social interactions on our app has also proven to be very effective in driving engagement. Today, roughly 40% of our daily active users are following at least 1 other user on Duolingo, and we're working to drive this percentage higher. We have a feature, internally called kudos, which allows users to congratulate their friends for certain achievements, including strict milestones, completing 10 lessons in a day and getting promoted to the next league in our leaderboards, just to name a few.

This quarter, we saw a 160% increase in the number of kudos sent by users, highlighting the growing popularity of this feature. Today, 8% of our DAUs are sending or receiving kudos on a given day. We also launched a phone contact sync feature, which makes it easy for our learners to add new friends. Early results are very promising. In the first few days, we already have over 500,000 learners that have opted to sync their contacts with us. Even better, many of them are following each other and congratulating each other's learning progress through our kudos feature. We're excited to see our app become even more social in the coming months.

Growing paid subscribers is another strategic initiative as this drives the majority of our bookings and revenue, which supports the platform and our mission to develop the best education in the world and make it universally available. Today, over 5% of our monthly active users subscribe to our Plus offering. In addition to having an ad-free experience, we offer other benefits that help users learn faster.

We recently launched 2 new features for Duolingo Plus. One, we call Legendary where users who reach a certain level in the app, get their skills to change to a different color. And the other called Mistakes Inbox, where we capture in 1 place mistakes you've made while using the app and give you a chance to focus on practicing those skills. We believe that each of these has had a positive impact on our user conversion from free to paid as well as subscriber retention.

And speaking of that, we also spent considerable time improving the retention of our subscribers. We found that annual subscribers are not only more likely to stick with their language goals, but they have 2x the lifetime value compared to monthly subscribers because their retention rate is much higher. These learners on our annual plan now make up roughly 80% of our subscribers.

In July, we quietly rolled out our family plan on iOS for Duolingo Plus, which allows up to 6 users to sign up under the same plan. If our experience is similar to other companies who have rolled out a family plan, then this plan will have even higher retention rates than our individual annual plan, which would further increase the lifetime value of our subscribers. At this point, we are not broadly advertising the family plan just yet, so we believe we're just getting started.

Our next strategic initiative is to become the language proficiency standard. Our Duolingo English Test, or DET, falls under this initiative. And in Q3, we achieved some key milestones, including DET's 5-year anniversary. I'm very proud to report that our DET delivered strong growth this quarter and year-to-date. Given that we are lapping quarters where the DET business expanded extremely rapidly due to COVID lockdowns, it's all the more impressive. We believe we've maintained our momentum, and are continuing to build on the strong trends from last year.

As of the end of the third quarter, all of the top 25 U.S. universities in terms of international student volume now accept the DET as proof of English proficiency for international undergraduate admissions. Today, we have over 3,600 programs worldwide that accept the DET, and we are excited about the growth we are seeing.

Before I turn it over to Matt, I'd like to quickly touch on the recent news around app store fees. Last month, Google reported that on January 1, 2022, they would be lowering their app store fees on subscription revenues collected from first-tier subscribers from 30% to 15%. This is a very positive development for us, as today, around 20% of our revenue comes from the Google Play Store. As we've said before, we view the app stores as great partners. They enable us to distribute our product globally and they help us monetize via subscriptions with relatively little friction for the user. This latest development by Google allows us to continue that relationship but at a lower cost so we can reinvest these savings back into our products and learners' experience.

And now I'll turn it over to Matt.

Thanks, Luis. I'm excited to talk to you all about our results this quarter. We not only exceeded expectations on several key metrics, but we also achieved all-time highs. As will become common on these calls, before I get to the detailed financials, I'm going to start with a review of our growth framework to put the numbers in a clearer context. As to our growth framework, first, we have our language learning app, which generates revenue through subscriptions, advertising and in-app purchases, or IAP. Second, we have our new initiatives, which include the Duolingo English Test, or DET, as well as other new products that we may monetize in the future.

Subscriptions, which make up a majority of our bookings, is driven by growing the number of free users, converting more of them into Plus subscribers and increasing the lifetime value of those subscribers. Growth in advertising and in-app purchases is also driven by growing users as well as through optimizing our ad and IAP revenue per user. The Duolingo English Test grows by increasing adoption of the test at institutions and with students taking the test. Finally, new product growth will be driven by expanding into new areas beyond language learning, like literacy.

Now I'll present our Q3 results. As you can see, we delivered strong performance compared to the prior year and achieved record results in bookings, revenue, subscribers, MAU, DAU and gross profit. Total bookings increased 57% year-over-year to $73.1 million. Subscription bookings increased 64% year-over-year to $55.4 million. And revenue increased 40% year-over-year to $63.6 million.

Now I'll take you through the key drivers of our results. As I mentioned when I was going through our growth framework, our language learning business grows as users grow and as we monetize them better. This quarter, our performance was driven by strong user and subscriber trends. MAU increased to a record 41.7 million users, and DAU increased to a record 9.8 million users. Paid subscribers grew to 2.2 million subscribers, an increase of 49% over the prior year. Our subscribers as a percentage of MAU now stands at 5.5% compared to 4.2% a year ago.

As for bookings and revenue, this quarter saw strong growth in those financial metrics as well. Subscription bookings were driven by a combination of user growth and higher bookings per MAU, which was driven by an increase in subscribers on the annual plan in addition to driving more conversions from enhanced Plus features. As Luis mentioned earlier, annual subscribers have about 2x the lifetime value as compared to monthly subscribers. We are also encouraged to see annual plan subscriber retention increase in Q3 year-over-year. We ended the quarter with the highest September retention we've seen since 2018.

Our revenue grew 40% year-over-year. It grew slower than bookings because as I mentioned, our bookings increased in part because of a mix shift to more annual plan subscriptions. The revenue for the annual plan is recognized over 12 months. And so the revenue for a large portion of our new annual plan bookings hits in later quarters. Our gross margin remains strong at 71.6%, which is 50 basis points higher than during the third quarter a year ago. We did see a slight quarter-over-quarter decline, which was driven by our DET business, where we added additional test proctoring resources.

Moving on to our non-GAAP operating expenses. We report 3 major categories of costs: R&D, or research and development; sales and marketing; and G&A. R&D increased from 34% to 40% of revenue this quarter and S&M decreased from 24% to 23%. G&A rose from 15% to 17% of revenues this quarter due to additional headcount, professional fees and other costs, a significant amount of which were related to becoming a public company.

As we have discussed, we expect that over time, we will continue to invest in R&D as we build our amazing product through hiring more engineers, product managers, designers, data scientists and technical experts. We expect to see operating leverage in the other 2 cost categories, sales and marketing and G&A.

As a reminder, non-GAAP operating expenses represent GAAP expenses adjusted for depreciation, amortization, stock-based compensation and onetime expenses. This quarter, we recognized approximately $21 million in stock-based compensation. A majority of this was the acceleration of shares and expenses driven by the IPO and from the achievement of certain performance hurdles from our founder share grants. You can find a reconciliation of these items in the appendix.

Our growth in revenue and stable gross margin was offset by a faster increase in costs, which led to a decline in adjusted EBITDA to negative $6 million this quarter. For the remainder of the year, we expect OpEx to follow a similar seasonal pattern as last year, when Q4 OpEx as a percentage of revenue was a few percentage points below that of Q3. We added over $430 million in cash to the balance sheet this quarter, driven by the net IPO proceeds and strong bookings growth. Free cash flow increased from $0.6 million in the year ago quarter to $5.2 million this quarter.

Before I get into Q4 guidance, I would like to remind you all of what we said during the IPO process about how we think about balancing growth and profitability. We think that we have a long-term sustainable growth path given the underlying growth of our free users, and the room we have to increase subscriber penetration and monetization. We will grow users and subscribers, primarily by investing in R&D to create an ever more engaging and effective product that will lead to higher engagement and efficacy. We will run the business with cash flow discipline as we have historically, but we will continue to reinvest significantly in talent, especially in engineering, product and design.

Now turning to guidance. For Q4 2021, we are guiding to $79 million to $82 million of bookings or 38% to 44% year-over-year growth, $66.5 million to $69.5 million of revenue or 38% to 44% year-over-year growth, and an adjusted EBITDA loss of $3.5 million to $6.5 million. For the full year 2021, we are raising our guidance on bookings to $282.5 million to $285.5 million of bookings or 49% to 50% year-over-year growth. For revenue and adjusted EBITDA, we are guiding to $244 million to $247 million of revenue or 51% to 53% year-over-year growth, and an adjusted EBITDA loss of $5 million to $8 million.

Before I wrap up, I just want to reiterate what Luis mentioned earlier that there is a considerable seasonality to our business. We expect to see our user metrics fluctuate throughout the year, especially around holidays and associated marketing campaigns, and our guidance reflects this.

As we get into the busy holiday season, we anticipate seeing relatively flat user activity in November and early December, followed by a spike coming from our year-in review, and especially our New Year's resolutions of our users. This leads to an increase in usage at the end of December that goes through January. We look forward to providing our 2022 guidance on our fourth quarter earnings call in early March.

And with that, I'll turn it back over to Luis for closing comments.

Thanks, Matt. To wrap it up, I would like to thank our incredibly talented team that drives our mission forward every day. I started Duolingo with Severin because we believe everybody in the world should have access to high-quality education. Every employee at Duolingo, whether they're making people want to learn more or helping the rest of the organizations succeed, they contribute to a world where education is available to millions of people. Of course, we're not done yet, but with continued hard work, we will make a huge dent in this goal. I look forward to reporting on our progress in the future.

(Operator Instructions) And our first question comes from Mario Lu of Barclays.

First one is on marketing. You guys mentioned, you guys increased strategic marketing in the third quarter. I believe there's channels such as TikTok or social media channels that you guys tapped into. So just curious what you guys saw entering these new channels. And then there's also other reports that cultural phenomenons like Squid Game on Netflix drove increased users to learning that language. So anything to talk about there in terms of the overall strategy in terms of marketing?

Thank you, Mario, for the question. So the first thing I -- just to remind everybody, our main focus is on product, we're a product-driven company. We're going to continue spending most of our efforts on making our product better so that people tell each other. And so that we grow through word of mouth.

We have been doing all kinds of different marketing. As you saw, and as you mentioned, we've had a lot of success with TikTok. With no expenses whatsoever, we grew our TikTok account basically from 0 users to about 1 million followers. And we're getting a lot of media coverage about that -- about just saying, Duolingo is the only brand that understands TikTok. So we're very happy with that. We are seeing an increase in the number of users that sign up to Duolingo that say that they came from TikTok or they heard about Duolingo from TikTok. So we're pretty happy with that.

The other thing -- you mentioned the Squid Game. That is also something else we saw -- just an increase in the number of people learning Korean through the Squid Game. So this is a pretty cool thing. So we're -- that's basically -- the other thing to say about our marketing is -- there's kind of 2 buckets of our marketing. One is this bucket that we just talked about is kind of like the TikTok stuff, everything -- things where we don't spend any money.

There's the other bucket where we spend money. And generally, what we're doing there, there's kind of 2 buckets within that bucket. One is we are doing some marketing in certain geographies that were underpenetrated. And that is -- we do some influencer marketing and some small amounts of performance marketing. And then the other bucket within this bucket is just experiments. So we've been experimenting a lot. Some of the experiments work, some of the experiments don't work so well. But basically, we're still -- with marketing, we're still in the early days. I mean we really only started paying for marketing in the last couple of years. So we're still in the kind of test-and-learn setting.

Great. That's helpful. And then just one follow-up on social. You guys talked about kudos during your presentation. So just curious over time -- I know a la carte revenue is still a small percentage of total. But as you kind of increased user engagement, is there a greater opportunity to grow the a la carte revenue? Or is this social -- these initiatives mainly to increase retention?

That's a great question. So for now, the majority of our social initiatives are to increase retention. That said, we have recently formed a team to do a la carte in app purchases. And so we do think that, that is going to be increasing a little faster than half so far. But I would like to remind everybody that we are mainly a subscriptions business, and that will continue being the case for the foreseeable future.

The next question comes from Justin Patterson of KeyBanc.

Great. Luis, I appreciate you starting with your mission and reinforcing you are a product-driven company. As you look to reinvest around app store changes, how do you think about allocating capital toward the product side versus potentially passing some of those savings on to the users to bring pricing down and just further grow that subscriber base?

Thank you, Justin. Thank you for the question. So we are -- I think, as Matt said, we are going to continue investing a lot on our product. That's our focus. We really believe there's a lot of runway to make our product better, and we're happy with where we're at, but there's just a lot -- we can teach a lot better. We can make it significantly more engaging. We can also convert a lot more of our users into paying subscribers by adding features to Plus. So we think there's a lot of runway there.

In terms of kind of passing savings to our users, the one thing that reminds me of -- one thing that -- one initiative that we're going to be working on and we have -- we're going to continue working on, we have already been working on, is regional pricing. Right now, the price of Duolingo, the majority of countries have the same price, and we understand that is -- that just has to do with the fact that we're relatively early in our monetization. It really shouldn't be the case that users in a country like Colombia pay the same amount as users in the United States. So you're going to see that over the next few months, we are going to continue having a kind of different pricing per country. We already have 4 countries where we have different prices, and you're going to be seeing that.

And sometimes people ask us, why can't we do that just much faster? It turns out this takes a little time. Because whenever we do our regional pricing experiments, we need to figure out what it does to our revenue. And part of seeing that is we have to look at what it does to resubscriptions or just re-ups and renewals. And so it takes a little bit of time. But that's one place where we do think that we're going to, I guess, pass on savings to the user.

Great. And if I could squeeze in one quick follow-up. You added a lot of interesting new features roll out from the Mistakes Inbox, the Legendary. Could you talk about just any learnings around engagement and conversion from those product updates?

Yes. I mean, generally -- we do have a bunch of new features. We actually have significantly more features in the pipeline. I mean we run hundreds of AB tests per quarter. And so you're going to see our product continually change over time, just get better and better because we run so many AB tests. I mean the general learnings that we have are that we can increase the fraction of people who subscribe. I mean if you see 1 year ago, the percentage of our MAUs that were subscribers was something like 4-ish percent, whereas we are now at about 5.5% of our MAUs are subscribers, and that is all through basically adding features like Mistakes Inbox and Legendary or other kind of changes like that.

We also -- another big learning is that we can move our user retention. And this is not just subscriber retention, but also free user retention. We look at this number that we call current user retention rate, which is basically the chance that -- a current user is somebody who used Duolingo today and also another day in the last 7 days. Current user retention is the chance that they use it today and also come back tomorrow if they used it in the last 7 days. We know we can increase that, and we've been increasing that over the last couple of years. It's gone from something like 65% to close to 80% now. And so that keeps going up and up. And that's generally just good for us, not only because it's good for our free users, but also increases the retention of our paying subscribers. So the big learnings is we can move those numbers.

Our next question comes from Nat Schindler of Bank of America.

I was actually just wondering if you could talk a little bit about the learnings you might have seen as kids went back to school. This was probably the first quarter where we saw a really substantial change from the year prior in schooling. And I wanted to know if that had any impact on usage.

Thank you, Nat. Good question. So every year, around back-to-school time, August-ish, September if you look at our monthly active users graph or daily active users graphs, every year we do see an increase in usage, in number of users and also increase in the amount of time that they use Duolingo as people go back to school. We saw that increase again this year. So it was no different than last year. I don't know if it was 0 difference, but it was pretty similar to last year -- basically, just like every year, we see an increase. So we're doing that.

I should say, our belief is that a significant -- at least in the United States, a significant number of language classrooms use Duolingo. Our belief is that roughly 40% of language classrooms in the United States are using Duolingo in one way or another. So we see that in our numbers compared to the summer -- from the summer to going back to school. So we saw that this year again.

But to be clear, no change with COVID, without COVID, with kids actually physically in school versus doing school from home.

I don't -- no change that popped up to -- for us to really notice. There may be a small change there, but nothing -- basically, it looks -- the graph looks pretty similar to previous years.

The next question comes from Ralph Schackart with William Blair.

On the paid subscriber number in the quarter, it was much better than we had modeled. During the prepared remarks, you talked about increases in annual plans and retention. Any way you could provide a little bit more color there? Were these marketing-driven programs to market the annual plans or win back campaigns and retention? Any color there, and I have a follow-up.

No, happy to provide some additional color. So I think the primary drivers were what we already talked about, the product changes. The product changes, as Luis mentioned, like the Legendary rollout, the Mistakes Inbox, making the app more social. I think these things, as he mentioned, do increase conversion. And so that's what we saw over the quarter. The shift to annual plan certainly impacts conversion because those experiments impact the conversion to the annual plan as well, but that actually had a more profound impact on the bookings per subscriber, bookings per MAU number. So it was kind of a twofer in that regard.

Great. And Luis, you talked about, I think you said 4 countries, you had different pricing programs. Maybe just any color on what you're observing there, and sort of does that lead you to more optimism as you rolled out new pricing strategies to other countries in terms of driving conversion?

Yes. I mean it -- so with these countries that we saw, we saw an increase in total bookings from these countries whenever we change the price. So that's good. I should mention that in many of these countries, there's -- a lot of the countries where we already monetize well, we think that the price is pretty good, like the U.S. and the U.K., et cetera. For a lot of times in developing countries where truthfully, our price is just too high right now because it's the same as the prices in the United States.

What we're going to see -- what I think we're going to see is we're going to see possibly increased bookings, but I think -- it's not going to be as immediate as you might think. Because I think in a lot of cases, this is for many of these countries, lowering the prices isn't necessary, but not sufficient condition for a lot of people -- a lot more people to describe. I think a lot of times just things like -- for example, in many of these countries, people are just not used to paying for digital subscriptions. So the way we see it is we have to have the right price, and we're going to do that. But over time, I think we're going to see the penetration in a lot of these countries increase a lot as, for example, they started getting Netflix. We just don't think that for many of these countries, we're going to be the first subscription that a lot of people get. But once they're used to getting something like Netflix, we think we're going to come soon after.

The next question comes from Mark Mahaney of Evercore ISI.

Okay. And Luis, I think you began this whole conference call talking about how you wanted the company to be a learning application, a learning platform, not just a language learning platform. So given that, can you just talk about the cadence or the time line you have in terms of getting into learning -- being a platform where people can learn subjects other than language?

Yes. That's a great question, Mark, because it's something that I think about a lot. And like you said, I mean, our goal has always been to be -- to develop the best education in the world, and not just language education, and then make it universally available. That's our mission. That's our goal.

So we've started -- and we've been working on languages for this whole time. One of the key things that we know is that a lot of the technology that we have can be applied to teach other subjects. For example, a lot of our gamification where we have technology built for that or also a lot of our personalization systems that try to really figure out what is the best thing to teach to each user at each time, all of that can be reused to each many other subjects. Maybe not every single subject in the world, but many other subjects. So we are going to be reusing that and making other products.

We do have one other project -- product out there called Duolingo ABC, which we launched last year. We are also working on a product to teach elementary school math. It's going to be a mobile app. My best guess is this is going to be launched next year. Although I will tell you, I've been working on developing products, digital products for the last 20 years. And it's very hard to give a precise time line on when this is going to happen. I think we're going to be able to launch that next year.

And then just in -- the other thing that I will say is as you've seen for us as a company, we are not a company that puts out a product and then puts $50 million of marketing behind this product to make it grow from 1 day to the next. Our belief is that really good products should grow by themselves mostly. So we are going to put this out, but I would not expect that next year, for example, our math app suddenly has 50 million users and makes $50 million of revenue. That's unlikely. So I would say, at least for next year, it's not likely that we're going to see any of this affect our revenue, but we're going to be putting these products out and working on them for a while, and having -- growing them through product improvements.

Yes. I got it, Luis. And the emphasis on leading with product makes a ton of sense. I get that. The other question I wanted to ask is, I know this is a small part of your revenue stream now, advertising. Could you talk a little bit of -- double click a little bit on products -- product development for advertisers? And I don't think that's a top priority for you. I think it's really about users and consumers, et cetera. But just talk about where you are in terms of having marketing solutions that are good for marketers and advertisers on your platform.

Yes. It's a great question. I mean as you know, so I don't know the exact numbers, about 10% of our revenue comes from advertising. And a few things to say. The first thing to say is, as you correctly said, we are mainly focused on subscriptions. We really believe that is the right thing to do. So we're mainly focused on subscriptions. And right now, our ads -- basically, our free users, which is 95% of our users, give or take, see ads. And most of the ads they see are -- all the ads that they see are programmatic ads that come from ad networks like Google or Facebook. That's kind of what we have.

A couple of things to say. Our ads business has been growing pretty nicely. I mean it's been growing about as fast as the rest of our businesses. So that's good. And we believe that it will continue growing. It will also grow just naturally -- as our users grow, our ads business will grow. We also do have some optimizations to do, for example, trying to figure out which users may just never subscribe to Duolingo Plus, and then we just would monetize entirely through ads or also trying to figure out if there are users that are more okay with a higher ad load. So there's some optimizations to do.

The one thing that I will say is that you should not expect that we will suddenly triple our ad revenue or anything like that because I just don't want to make our free user experience crappier. Right now, we feel pretty good with our ad load. It's not that heavy. And our fear -- and not just our fear, we see it in our numbers. If we suddenly start tripling the ad load or something like that, engagement goes down. So we're -- we like the free user experience, and we're going to keep it that way.

The next question comes from Aaron Kessler of Raymond James.

Great. A couple of questions. Maybe just on Google Play. Matt, you mentioned that and how should we think about that starting to impact the model? Obviously, we believe it kind of goes from 30% to 15%. But should we start to think about that impacting the model on 2022? And just, do you happen to have the share count for the quarter as well? It's probably in the Q, but I didn't see it in the release.

Yes. No. Glad you asked the question. As Luis mentioned just real quick in the opening remarks, the app store fees are obviously undergoing a lot of change, and we're paying very close attention to that. But we view them as good partners historically that will hopefully continue to be good partners. And so we view this as just a nice tailwind for us.

And then the Google Play fee decrease for subscribers that are less than 1 year old, that will impact our margins next year, starting in January 1, 2022, on bookings that occur after that date. And so yes, we expect to see a nice tailwind there on the margin side. And after that, it's anyone's guess what happens outside of Google.

For the share count, the fully diluted share count is -- depending on the stock price, if it's around 60, it's around 46 million fully diluted. The weighted average share count is in the Q. So I'm not sure which one you're looking for.

Got it. Yes, that's helpful. Maybe just quickly on the Google Play, again. So should we expect the impact to start to impact kind of current revenues and COGS? Or is it only when you sign up new subscribers, and then that would flow through?

Yes. It's only going to be on new flow through. Because yes, the folks that got bookings this year, for example, like right now, if you signed up for the annual plan, you paid us $80 through the Google Play store, that's going to be at a 30%. On January 1, that drops to the 15%. So it's really only on go-forward revenue.

This concludes our question-and-answer session. I would like to turn the conference back over to Luis von Ahn for any closing remarks.

Thank you. I just want to thank everybody for the great questions, and we look forward to talking to you in future quarters.

The conference has now concluded. Thank you for attending today's presentation, and you may now disconnect.