Q4 2022 Duolingo Inc Earnings Call

We hope you enjoyed that fund tictoc, it's one of the many examples of organic learner generated content that goes viral every day without any involvement from us and we love to see the creativity and enthusiasm of creators like this which continues to drive our brand awareness and word of mouth growth and now let's get on with today's call.

Good afternoon, and welcome to dual <unk> fourth quarter and full year 2022 earnings webcast. All attendees are in listen only mode. Today. After market close we released our year end shareholder letter with our Q4 and 2022 results and commentary, which you can find on our IR website at investors Dot Duolingo Dot com on today's call.

We have Luis <unk>, our co founder and CEO and Matt <unk>, our CFO . They will begin with some brief remarks before opening the call to questions analysts will be able to ask the question by using the raise hand feature and please note. This event is being recorded.

Just a reminder, that 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 our risk factors of 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 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 now I will turn it over to Luis.

Thank you Debbie and welcome everyone.

Pleased to report that we finished 2022 with a record quarter, achieving our highest ever number of via use of subscribers.

We ended Q4 with $16 3 million via use which is 62% year over year growth and we ended 2022 with $4 2 million paying subscribers, which is 67% year over year growth.

These metrics lead 2022 to be our best year ever in terms of bookings and revenue, which grew 46% 47% respectively.

Was a fantastic year.

Our <unk> ratio, which is an indicator of user engagement improved to 27% this past quarter up from about 24% a year before we.

We have also seen the number of learners with long streaks grow immensely to remind you our streak as the number of consecutive days that are user learns with this.

This mechanic has proven to be an efficient tool for encouraging engagement and helping us grow deal use 63% of our da use now have a streak of seven days or greater which is up from 53% last year.

What's even more impressive is that we now have over $3 million <unk> with a streak of over one year that means that more than 3 million people use dual every single day for the last year or longer.

It is also impressive.

But it took us eight years to reach 1 million via use with a year long streak and only 21 months to get from $1 million to $3 million. This is one example of the compounding effect of our product improvements.

Our user growth was strong in every region of the world.

We also have users nearly all agents are also see economic statuses, and we have a variety of motivations to learn including work school in travel.

The result is a diversified business that has been largely on correlated with macro and geographic trends.

Our investments in monetization have also shown the power of compounding over the past three years, we've seen <unk> increase across nearly all cohorts users who are new to the app have converted more readily to our paid subscription and users who have been on the up for longer periods like multiple years have also seen an increase in their conversion rates from free to paid.

And now I'd like to discuss what you should expect from US this year.

And a lot of waste, we expect 2023 to look similar to 2022, we expect users and bookings to grow nicely.

We still have a massive 60 billion dollar addressable market, but we have only barely penetrated and we have strong organic growth on our side.

And I want to remind you that currently our subscribers make up about 8% of our monthly active users. So we have plenty of room to further monetize our current user base.

We also have an attractive opportunity to increase bookings from la Carte purchases like we did in 2022.

But I do want to point out again that you should not expect any bookings from our math or literacy apps this year.

As to where we'll invest the majority of our engineers designers and product managers will continue to build on the success. We've had in growth efficacy and monetization that means that we'll continue to run experiments to increase user engagement improve how we teach and optimize how we convert free users debate I.

And in our shareholder letter I discussed the two additional areas we're focused on in 2023.

First degenerative, AI, which will power a higher tier duolingo, Max description as well as help us reduce content creation cost.

And the second is improving how we decent monetize English learners.

I'd like to emphasize that we are excited about these initiatives, but they are still in the early stages.

And I also wanted to touch on profitability.

Coming off a strong 2022, we.

We are well positioned to grow users and bookings rapidly while also delivering higher profitability.

We plan to continue to compound the returns we've seen from our past R&D investments. So we don't plan to hire as many people as we did last year just to remind you we've never gone nuts with hiring and in fact, we've grown revenue at about twice the rate of head count in the past four years.

This year, we plan to show operating leverage across all cost R&D sales and marketing and G&A.

Given those strengths, we're guiding towards 10% to 12% adjusted EBITDA margin for the year up from about 4% in 2022.

And with that I'll turn it over to Matt to talk more about our financial outlook.

Thanks, Luis to quickly recap the highlights of our impressive Q4 and full year 2022 results in the fourth quarter, we delivered 39% bookings growth year over year, which was about 46% on a constant currency basis, we saw 42% revenue growth year over year, which was 47% on a constant currency base.

We had a net loss of $13 9 million compared to a net loss of $17 5 million in the year ago quarter, we posted our highest quarterly adjusted EBITDA of $5 million, which was a 5% adjusted EBITDA margin and we had a free cash flow margin of about 11%.

For the full year, we delivered 46% bookings growth year over year, which was about 52% on a constant currency basis, we saw 47% revenue growth, which was about 51% in constant currency. We had a net loss of $59 6 million compared to a net loss of $60 1 million last year, we saw adjusted EBITDA positive.

$2 5 million compared to an adjusted EBITDA loss of $1 million last year, and we had a free cash flow margin of about 12, 5%.

In 2023, we expect to continue delivering strong bookings growth and we will do so while making progress towards our long term profitability goal of an adjusted EBITDA margin of 30% to 35%.

For the full year 2023, we are guiding to $530 to $542 million in total bookings.

$486 million to $498 million in revenue and an adjusted EBITDA margin of 10% to 12%, which translates to about a 32% incremental margin at the midpoint.

If our Q1 2023, we are guiding to 127 $5 million to $135 million in total bookings of 111% to $114 million in revenue and an adjusted EBITDA margin of 9% to 10%.

Our guidance assumes current prevailing foreign exchange rates as a reminder, roughly half of our revenue comes from outside the U S. So every 1% increase or decrease in the value of the dollar versus our basket of currencies has about a $2 million headwind or tailwind respectively on full year total bookings.

Prevailing exchange rates, we've used in our guidance foreign exchange is expected to be a three percentage point headwind to Q1 2023 year over year bookings growth.

In terms of quarterly cadence of our bookings for the remainder of the year, we expect our year over year growth for total bookings in Q2 to be about the same as Q1, we expect Q3 year over year growth rate to be slightly higher and we expect Q4 to be our biggest quarter in terms of dollar bookings given our outperformance in Q4 of this year it'll be a slightly lower growth rate.

Has the pricing in 2022, we lowered pricing prices in several countries to get our prices more in line with each country's GDP per capita this.

This initiative and our intentional mix shift from monthly to annual plans was.

It was to increase the overall LTV of our platform and that drove down subscription revenue per paid sub but we're done for the most part with lowering prices globally and in fact based on strong conversion trends. We believe we have the opportunity to raise prices in some places in our experiments with pricing as we speak.

We have not included any material amount of bookings or revenue from doing a max in our guidance. Since we've only just started testing it but we will provide you with an update when we report Q1 results.

I would like to emphasize that we are focused on managing the business. So that we achieve our full year adjusted EBITDA margin of 10%, 12%, we will be making spending decisions throughout the year based on our revenue and gross margin trends to ensure that we are in this range for the full year.

So to remind you given the seasonality of our revenue our Q2 and Q3 margins will be slightly lower than Q1, and Q4 will be higher.

We will continue to run the business with discipline and prudently manage our expenses starting in Q1, we expect to see meaningful leverage in total non-GAAP opex compared to the fourth quarter of last year and in 2023, we expect to achieve about 7% to nine percentage points of operating leverage and total non-GAAP opex for both periods. The improvement is mostly in sales in March.

And G&A and to a lesser extent in R&D.

We ended the year with approximately 48 million fully diluted shares outstanding during the year end closing price in 2020 in 2023, we expect to end the year with about 2% dilution from equity issued to employees, which is less than the roughly 3% dilution we had in 'twenty two.

With that I'll turn it back to Luis.

Thank you Matt.

Before we get into Q&A I'd like to thank our very talented team of dose Ocwen, who continue to find new ways to innovate underwrite our learners and now we would be happy to take your questions I'm going to turn it back to Debbie to manage the queue.

And.

Here we go.

Okay. Thanks, Louise and as I mentioned earlier, if you have a question you can just use the right hand feature.

And the first question comes from Ralph Shakur of William Blair.

Hi, good afternoon. Thanks for taking the question first question.

Shareholder letter I think you called it a quote Super end of 2020 to think that may incorporate some of the beginning of the year campaign that you run to attract new users just curious what drove the really strong performance into year end and then how does how does your new year's campaign compared to expectations and I have a follow up.

Thank you great question. So, yes, we had a great Q4, great year end.

Main reason, there's two main reasons.

The first one is just a compounding product improvements I mean, we've just been getting better and better at making our products stickier and attracting more users. We have a very high performing growth team that does that and so that's one thing and then another one is some of our marketing campaigns were just did really well a good example of that was our year in review campaign, where.

Towards around December 14, or so or December to December 10th December 14th are all the users get a get a summary of what the what they did throughout the whole year and they are expected to go up while they are encouraged to shared we've got a lot of shares millions of millions of shares and we hit the top trending topics on Twitter and stuff like that by by just having people share.

Theyre duolingo stats over several days so stuff like that really helped us do the outperformance in terms of the.

New year's campaign, we started just for reference we start a new year's campaign on December 28th usually every year and it goes all the way through January so for this for this quarter. That's about three three days of campaign and they performed better than we expected.

And the reason for that outperformance just has to do with the fact that every year, we take the learnings from the previous year of the campaign and we just get a little better. So every year, we run a few AB tests to try to make the campaign more effective and we take those learnings for the next year and so that that worked out pretty well for us.

Great. Thanks, Luis maybe just a quick follow up.

And just in terms of dollars I think you reaccelerate for six consecutive quarters, which is pretty tough to do but just maybe some color whats driving the strong acceleration in.

And growth and thinking.

Remarks, you talked about sort of broad base, but were there any sort of one or two regions saw stronger growth than others. Thank you.

Yes, it's a great question too so in terms of regions. It really is growing in every single region in the world.

If you want to know stronger regions. The U S was particularly strong Asia Pacific, particularly strong in Western Europe is particularly strong.

In terms of growing via use.

And in terms of why it is I mean, it's again the same two reasons. It is a continuous product continued product improvements that just compound over time. The product is just better in stickier and then we just get better and better at figuring out what straw.

Strikes a chord with marketing so things like our social media campaigns like uptick of our campaigns on our Twitter campaigns and also some of the influence of work that we do and.

Some of the stuff is stuff we don't even know for example, you just saw this video at the beginning of this of this call.

These are entirely organic videos that other people make is almost every single day I don't know if it is a fact that every single day in the world, but almost every single day various on organic video that goes viral.

Matt mentioned duolingo, so all of that just keeps getting keeps growing our users.

Thanks, guys I appreciate it.

Alright, Thanks, Ralph and the next question comes from Justin Patterson of Keybanc.

Great. Thank you very much.

Recently, there was a really interesting posts on the duo Lingo blog.

<unk> is using data science to optimize da use I think current user retention rate was that metric has been optimizing the past few years, we'd love to hear more about just how you are using a lot of these top down statistical models to really drive growth linked the product team together and then now that you've gone pretty far with the current metric.

How are you thinking about tailoring that to.

Individual cohorts similar of a bottoms up approach to take advantage of things like say Gen Z I think much shorter attention spans and making sure. They are still having healthy engagement. Thank you.

Well I'm glad you read that blog posts, we're very proud of our growth model.

For the year. So we just gotten significantly more sophisticated at analyzing our user base and making changes that that improve certain cohorts.

And in this in this post we mentioned.

Ben.

I would say the single most important metric for this company over the last several years, which is something we call <unk> or current user retention rate and what that is it is it is there.

The rate of people that if they were here today and they were here over the last one other time at least over the last seven days what is the probability that they're going to come back tomorrow, that's current user retention rate.

That that number is around 80% right now and we've increased that over the last two three years, we've increased that number from about 65% to about 80% and every 1% increase both Kirk and increase our views, but by a lot.

And so.

That model has been.

<unk> really helped us and.

And we keep getting more and more sophisticated you are right. We're starting to this was a model at first when you're when you're just.

Before this growth model, we would just treat all our users in one cohort then we started getting more sophisticated on which we've kind of new user current users et cetera, now we're getting even more sophisticated than that I don't think we have a specific things.

<unk> for Gen Z that we do but we are.

Relatedly, we are making our sessions shorter and shorter because.

Like you said, they have a shorter and shorter attention spans.

Okay.

Great and the next question comes from Andrew Boone JMP.

Great. Thanks for taking my questions.

Can we start off with just any learnings from an app purchases I saw the double of this last quarter can you talk about what successes are there and then how we should think about that for 2023, and then secondly on profitability.

You know, Matt talked about real contribution margin kind of especially in the back half of this year is there any reason to think that that steps down as we think about kind of a.

Three year plan or something beyond this upcoming year. Thanks, so much.

So I can take the IP question. Madam then you can take the next one so for in App purchases. Yeah, we have a very high performing team behind us that if that is just making purchases you know more and more prominent and more and more successful in the up after you saw they've been growing very rapidly we expect them to continue growing rapidly this year.

In terms of learnings.

The main way, we do in our purchases by the way is by selling jumps. So there's this in app currency called jumps that you can either earned by doing stuff on the App or you can also buy them and then you can use <unk> to buy other stuff kind of power ups in the up one of the main power ups that we used in the timer booths.

That helped for certain time lessons, so we have to be.

Lessons from dwelling whatnot timed, but there are some lessons that you can do some practice lessons that are timed.

And the idea is that if youre not very fast you run out of time, where you can buy these timer boosts with gyms and theirs.

<unk> got to get a lot of people to spend on may not purchases. This is the type of stuff youre going to get to see us do so in general for us for in App purchases were not going to be selling content within our purchases, but mainly we are going to be selling gamification things that people really like and and so far that strategy has been working pretty well.

Yes.

Great and then on profitability.

So Andrew the profitability increased.

In 2022 by about 450 basis points from 2021 on an adjusted EBITDA basis, which was good progress towards our long term profitability goal of about 30% to 35%.

This year, we're guiding to 600 to 800 basis point improvement over that and so as we look out kind of in the total into the longer term. Our goal is unchanged, we want to be at a 30% to 35%.

Adjusted EBITDA margin in between last year and lower guidance for this year, we think we're showing the right progress towards that goal.

Okay. Thank you.

Thanks, and then the next question comes from Ryan Macdonald of Needham.

Hi, Thanks for taking my questions and congrats on a strong close to the year Lewis I'll start with you on our on the Duolingo English test there was recently a a ban in China.

Government on online degrees that if a Chinese citizen essentially is taking a fully online degree from an international University. They are no longer recognizing that <expletive> Ryan. It seems like this is going to start creating sort of a shift in that inflection point and more Chinese learners looking to go in person oversees I'm just curious.

If you've seen any impact of this thus far on.

<unk>, our sign ups for the dwelling English testing and how you think.

What are your expectations for that are into 'twenty here.

Well thank you for the question.

We have not seen any impact based on this.

I think theres, many reasons I mean for one year.

China as you know around.

A relatively large chunk of the b the dwelling languished tastemakers, but we'd really have to speakers are all over the world.

The other thing is the reasons for taking the tests are very varied.

Some of them as people who are applying for international.

Undergraduate admissions, there's also graduate admissions Theres transnational education, there's there's all kinds of different pockets and maybe one pocket.

<unk> has some some sort of policy implications to it but in general the pockets of pretty healthy.

Kind of one thing the other thing is we are still you know we're not the leader in this market we are.

We're coming into this market with an online test than it is a market that usually have had all these offline test.

So given that we are still in the <unk>.

Kind of growth face of this we don't it doesn't affect us too much when there's something that changes the full volume of it because we are coming in with with whatever.

Percentage of tests that we have which is <unk>.

Inefficiently under 25%.

We're still growing quite a bit and pretty fast. So this is not something that where we're seeing yet.

That's helpful. Appreciate it and then.

Follow up you talked a lot about generative AI and the use of that I'm very curious around the content creation and you mentioned an increased focus on on sort of more advanced English content for non English speakers are basic English speakers.

How quickly do you think that the generative AI can compress your content creation timeline.

And how do you manage for quality in that and then for Matt are you assuming any within your the leverage that you're showing this year.

<unk> margin side are you, assuming any leverage from that content creation from Janet Nebraska. Thanks.

Yeah. Thank you for the question. So yeah, we're very excited about generative AI.

And just to put things into context.

<unk>, we've always lean into AI from the beginning I mean, a lot of our products are built based on AI et cetera over the last couple of years generative AI has become significantly better on works definitely going to lean into this as well.

You mentioned content creation Thats, one of the places, where we think that it can help us a lot.

And speed things up quite a bit it's a bit early for me to give you a precise estimate of how much it can speed it up but we are certainly I mean, it's already happening we already speeding things up and also lowering costs.

One of the places where we're using that is to be able to create a much more English learning content. So youll see us create a lot more of that in terms of quality assurance.

We're still gonna have humans go over it but the timeframe you know it used to be the case that all of this was almost entirely generated by humans by now.

Has to be reviewed by humans and it's a lot quicker to review them to generate.

So in terms of timeframe, we're going to forget I get a massive increase in that.

I assume it's the same for cost.

So yes. This is something we're very excited about.

The other thing I should mention of our generative AI. So content creation is one of them and we're gonna definitely use that but it also is going to help us just create better experiences to our users kind of more online conversational practice that we're gonna be able to do and that's the type of stuff that we're putting into our new higher tier subscription dwelling lomax. So so this is something we're very excited about.

As a company.

We've partnered with open AI and we're gonna be working very hard on this.

Yes, and Ryan Luis Luis answered the question for the guide I mean, we only included in the guide things we have direct line of sight into it's not generative is so new we don't have a ton of <unk>.

Your line of sight into direct savings, yet, but as we said.

Overtime.

Probably the way to bet.

Okay.

Thanks, Ryan and the next question comes from Aaron Kessler of Raymond James.

Okay. Thank you a couple of questions maybe for Matt.

On the paid conversion rate was very strong in Q4 can you just talk about that was ups family plans are part of that and then how should we think about bank paid conversion in 2023.

And then maybe any update I think you've been testing, maybe perhaps human tutoring as well as any.

Thoughts on that direction.

Yes, as well, especially as maybe as it relates to doing a max how you're talking about thinking about integrating that as well for.

Sure Yeah, I'll, let I'll, let Luis.

To that last part for the.

Payer penetration, we had mentioned I think throughout most of last year. So we expect that a 1% to 2% increase each quarter year over year in that metric and that's coming from all of the things that Luis has described before that all the product improvements that compound over time.

To make our free to paid conversion go up in general there.

It wasn't any one particular item like family plan or this or that it was all of them compounding over time I think going forward as we mentioned on I think it would either Q2 or Q3 call last year.

Given the growth in Niu.

So that denominator of that number is just growing really really nicely.

And the retention trends, we would expect it to actually instead of adding one or two points every quarter year over year be closer to one this year.

Just because again that denominators growing really nicely. So it will ultimately turn into subscribers over time.

But in the short term it'll be about one.

One point.

And then you want to talk about Maxim and tutors.

And human tutoring I mean, Joe the first question was about human tutoring or the second question about humans.

Visits.

Over the years, we've waffled on human tutoring at first we launched duolingo with no humans involved then we try to we tried putting it in we try to.

And over the last year, we've made the.

The decision that I think is the right decision that we are not going to get into human tutoring at all we've tried it we are out there.

There's a number of reasons for that are you. There is nothing that isn't like we have anything I can see when tutors is just and on our end we are a technology company and that's what we excel at and in addition to that if you look at it.

Theaters are great, but they're not getting any better there does there.

Whereas technology in particular AI in particular, a lot of stuff coming from generative AI is just getting better and better every year. So we're going to bet on that much in the same way you know like 10 years ago. We were told we could not teach a language with synthesize computerized voices.

We bet on back to even though 10 years ago synthesize computerized voices still sounded kind of robotic today. They sound excellent we feel the same with with AI.

And five years. So I don't know 510 years, we will be very happy that we will have that on on technology, rather than rather than humans.

Thank you.

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

Yes. Thank you I actually wanted to just go a little deeper on some of the questions you had before really about how the users of change you saw a dramatic acceleration in both.

Users and paid subscribers over this year versus the prior year has there been a real change in the demographics of those so anything substantially different about users. This year than there were last year.

But that's it's a great question.

If there has been it's small.

Generally our user base is very wide. We have uses from every single country in the world every single country of the World is represented we have users from every single associate economic status from all.

All ages as long as they know how to read so it's kind of from six years to about 100.

We just really have users from them from all over the place.

It's very wide user base.

So it hasnt changed too much.

The one thing we've noticed is it has gotten slightly ever so slightly younger.

And that probably has to do with a lot of our prominence on on a social media like to talk.

But but that's it so we haven't really noticed too much of a change.

And as those changes towards going slightly younger as that happened in both.

Free users and paid.

B.

Honest answers I don't know I know that it's a free users getting a slightly younger I don't know for paid.

Maybe.

And then on the wheel yeah, I'd have to go back and double check not on the page I think what we've seen on paid is that the conversion.

Trends haven't really changed some of the accelerated for example, just overall conversion of.

Kind of cohorts across the board is has increased over the course of last year. So folks who are new to the platform and the first couple of months they converted more rapidly than in 2021 folks who've been on the platform for long periods of time, they converted a little bit faster than.

And then they had in the past.

We've always seen that people, who use the app more convert more.

That.

That stayed the same or gotten a little bit better.

So yes, it was kind of a broad based just like our users grew around the world.

And a pretty broad based way so the same trend on the free to paid conversion.

Thank you very much and once again very impressive.

Thank you.

Thanks, Nat and next question comes from Mark Mahaney of Evercore.

Two questions you had this line in the release about how the number of returning users is roughly the same as the number of brand new users I would just assume that that was always been the case or it's been the case for a while but is that something new that's happened because I get your point about how.

People will use it and then they'll either come back to it when they realize the importance of learning a new language or there'll be another opportunity maybe they're traveling to a different country. So is that a new trend or not what's the so what out of that and just real quickly dual masso that math product just where is that are you still tweaking is it far from being where you want it to be.

Just talk about kind of adoption how much interest you've really seen in that thanks.

As for your first question. This is not a new thing its been a while since returning users are the same as new users I'm going to assume at the very early of dwelling that was not the case, but it's been a while but it's been like that.

It's because.

It is rare for us to see a user that uses duolingo goes away and never comes back.

That is a rare usually people may even go away for two years and then come back its just because it because there's such a little friction. So we see that the reason we put that there is mainly due to to emphasize that it.

It is not right to think of our use is coming having churn and been gone forever that it's rare it's more that they come.

Stay for a few months or something then they may take a break for a couple of months and they come back or something so so something like that.

And different math, we're very happy with the progress is growing rapidly we're happy with that just to remind you of the plan.

We are right now the goal is to grow our MAF product organically.

For it to grow by itself once we get to a pretty decent number of users call. It 1 million daily active users.

This is kind of.

It doesn't have to be exactly 1 million once we get there we're going to say okay. Great. We have good product market fit it is time to start monetizing.

We're not there yet, but it is growing pretty nicely and so we're happy with the results in terms of tweaking guests were working very hard on tweaking. The app, we're adding a lot of the features that we know work well on the main duolingo app to the to the map product so improvements to the street.

We're thinking of adding things like leaderboards better notifications. So a lot of the stuff that we've learned over the last 10 years on duolingo and and we have the code for we're starting to add to the map app to get it to have higher returns. We're also adding more content to it.

You know one of the one of the things that happens with the math up right now is that people actually run out of content. Because if you are very heavy turns out we have some very heavy users that use the whole thing and then they kind of run out of content.

So we're adding more content to it so you'll see us over the next year I'd, just make the product better and better.

So that it will continue growing.

Okay. Thank you Louise.

Thanks Mark.

Next question comes from Eric Sheridan of Goldman Sachs.

Great. Thank you maybe first question would be following up on something that was implied in March 1st question travel has obviously been a big tailwind and the shift to services over goods has been a big tailwind in the macroeconomic environment can you talk a little bit about what <unk> seen with travel potentially as a catalyst for us.

Page both top of the funnel the bottom of the funnel as <unk> continued to build momentum and how maybe as a result of cross border travel continuing to return that to be a nice tailwind for the business going forward over 2023 and beyond.

And then the second dynamic would be you talked a lot about morality and tech talk can you talk a little bit about where you are excited to experiment and think about multilingual buildings morality on platforms on the content side or the distribution platform side, maybe even away from <unk> and how we should be thinking about that building in the years ahead. Thanks, so much.

Great Great question.

Let's see in terms of travel we don't see in our metrics that more and more people are using duolingo for travel we did see a very temporary dip in the height of Covid a couple of years ago that it was like a pure a slightly fewer since then it's been about the same.

And we just you know travel it's interesting for a lot of people like us that are you know.

To be honest from the world wide perspective pretty wealthy I think travel is like a big motivation for using duolingo, but for a lot of our users.

We're using it to learn English or as a hobby or something so we don't see these things I was like Oh, My God travel has opened up and therefore dwelling was exploding that just we just haven't seen much of that we do ask when people sign up why they're using dwelling when some fraction tell us they're using it for travel I don't know the number off the top of my head I'm going to assume it's something like 15% to Dallas.

That is that they're using it for travel, but it's just not something that we.

We look too much too.

In terms of your second question about virality so.

<unk>.

So far we have experimented with a few social media platforms to get our word out by the way and this is mainly organic as you know we don't we don't pay for this stuff. We just make really good content Tech talk has been really good for us.

Not just picked up in the U S. We've actually local lifestyle ticked off too.

Portuguese and Spanish.

Other languages to one of the things that we're excited about is a tick tock accounts in the other languages that are actually growing really fast and it's actually amazing how by the way. This is this amazing thing about Brazil anything you do with social in Brazil, thus well and so as soon as we put a tick tock in Portuguese it just kind of went nuts and if you. If you go see those videos. They are amazing we just had.

Dressed up.

Wednesday from the Netflix show Wednesday.

And it it got I don't know, how many 10 million views or something.

So we're excited about platforms like <unk> and actually also Twitter, where we're doing pretty well on Twitter.

Now when you ask about virality and general another thing that we're doing is we're just adding a lot more things inside the product. This is not a social media or anything inside the product to get more people to share things and.

And we now have a team behind it it's a really good team too and they are able to get a lot more people to share things one of the main things that people share.

So now way to share sentences much more easily and much more easily one of the main thing people share our weird sentences that they see on dual angles. So some of the courses for one reason or another have a very weird sentences.

I don't know if I can think of one of them like the German courts has something like the sentence or something like that.

Right next to the bed lay the body of the husband or something like it really weird stuff that people share and that you know that.

It just gets a lot of people. So we're working on getting people to share like that.

Working out pretty well.

Alright, thank you.

Hey, Thanks, I, probably wrote that question about the.

That sentence about the husband Okay.

Alright next question comes from Mario Lu of Barclays.

Great. Thanks for taking the question. So the first one is on App store fees I will.

Just curious your thoughts on the recent trend.

Asked bypassing the App store for example games like Fortnite and apps like Spotify Netflix have done this for a while but more recently more traditional mobile games are implementing.

A direct to consumer payment method I was just curious your thoughts on any initiatives that you guys have on your end.

Were not currently working on that.

For us.

We love the App source obviously.

It's hard to say that we wouldn't love it if they decrease their fees.

Of course, we would but we do really well with the <unk>, they're an amazing distribution mechanism.

And the thing when you go off App stores is that it's harder to distribute a product and also it adds quite a bit of friction friction one of the amazing thing about the App stores is that people already have the credit card in there and they just have to tap once and it gets it there. So you may be able to lower your fees, but you add so much friction through your users, but it's not clear to us that we would make that much more money.

It's something that we're always thinking about and May eventually do it do it but it's not something that we're currently working on too much.

And remember Mario we did get a quite significant bump from Google lowering their apps appstore fees last year or so.

Got it helpful. And then second one just on kind of changes to the App itself I believe past few quarters, you talked about the change to a single bearing track being at Terry.

Any updates to that.

And just in general how do you balance between kind of optimizing for learning and versus having a more casual AD tech and it attracts the masses. So I think just user feedbacks can be like.

Loud minority, but the removal of like it seems like oil prompts or tech answers I believe got some negative feedback. So just wanted to hear your thoughts there. Thanks, Mario you regret it.

[laughter] yeah. So it's a great question about the changes to that so we've made we made some pretty significant changes, we're always making that better I mean this is what we work on this is why we spend so much on R&D, where literally always making that better over the last year. We made this large change to our homes.

Screen going from giving a lot of choice about what to learn next to this linear path.

The main reason, we did that linear path by the way was to teach better we noticed than we were when we were giving people. So much choice what was going on was a significant fraction of them. We're just coming into extend the streak with really easy lessons that we're not actually getting them to learn whereas now with the linear path. We've encouraged our significantly larger fraction of users to whenever they.

Come to increase their streak. They actually also learn something so we the main reason we did that was actually the teach better and also to simplify our product which has really helped us.

That was actually to teach better.

The way, we think of it as.

When we're making changes we look at we have we have three things that were improving we're improving how we teach we're improving how engaging the opex. So basically you know how much time people spend on it how retention et cetera, and we're improving how we monetize and we have large teams behind each one of these.

I'll kind of group of metrics and what happens is that each team for example, there's a monetization and what they're trying to do is get more people to subscribe.

Whenever they run a test to try to get more people to subscribe.

What we what we enforced is that they cannot screw up the other metrics. So yes, you can get more people to describe what you cannot decrease.

Our number of da use you cannot decrease the time people spend learning or you cannot decrease the learning outcomes. Similarly, when we get people to the group of people that if that is dedicated to teach better. We tell them. Okay. You can you can you can make changes that pizza better, but you cannot screw up the monetization or you cannot screw up the.

The user engagement, that's how we balance that and we found that that works pretty well in most every change in the App you can actually trade down that change was made by the group onboard by the team of people that are trying to make us grow faster or make us make more money et cetera, and so.

That's how we do it hopefully that gives you some color.

That's very helpful. Thank you.

And our next question comes from Arvind <unk> from Piper Sandler.

The muted arvin.

Yeah.

Yes, okay.

Can you hear me now.

Yeah.

Oh.

Perfect.

I guess my first question is you know.

Among the management team who has the longest.

And it comes out.

Right.

Let me now asset.

But.

Good morning.

But I'm willing to bet. It's me I'm willing to bet at me, but I don't know Dan.

[laughter] Louie, yes, we can.

Can be pretty sure about that.

[laughter] alright.

Makes sense.

Yeah Yeah.

Yes.

Really good metrics.

The user engagement and you.

You know kind of the revenue growth and bookings, but but I know, there's a whole lot of other metrics you track that.

<unk>.

Beyond what you'll share with us. So can you just conceptually if you're not able to hit hard numbers, even conceptually share a bit more underneath on some of the metrics, that's giving you confidence in kind of.

The shape of the growth for this year.

Yeah, I mean, I don't know if we can share precise numbers of but there are certain things that we share the stuff that we don't share, but we do track.

A lot of metrics.

In general.

I was saying before there are three groups of metrics. There is what we call the growth metrics, which is just usually are.

Our entire user base.

Are they are they using the product more do we have more da use et cetera. There is the monetization metrics, which as you know.

How much money, we make basically and then there's the metrics about how well we teach in each one of them we have a.

Ton of metrics.

In terms of growth some of the things that at least I look at I look at time spent learning.

I look at a <unk> ratio.

I look at rest of erection right. So kind of number of people that are being retroactive I look at the retention of resurrected users I look at the retention of current users I look at the retention of new users.

That's the type of stuff and monetization of the types of stuff. We look at are things like <unk>.

Conversion from free to paid LTE.

LTV retention of paying users daily bookings is something we look at a lot.

Number of users or fraction of users that have.

I have an IAP or do I purchase an IAP and learning we look at we look at a number of things, but one of the main things we look at its difficulty of the content that we're able to give out without scaring users off away.

So.

There's a lot of metrics that we track.

I don't even know all of them off the top of my head.

Yeah Yeah.

That's helpful and then.

Kind of bookings growth you know how can I say.

Externally in some sense is a bit of a black box, but.

Are you able to share like when you project for bookings growth what are the inputs that go into into kind of coming up to take our projections for bookings bookings through it I mean.

We don't need another specific numbers, but just conceptually what what goes into.

Predicting user growth yeah.

Something like that.

We just take a ruler on go up like that and it works pretty well I'm kidding, Matt has a much better answer.

Maybe.

Maybe not.

I mean, the main thing to think about with bookings growth growth is that.

There's just the two dimensions theres, new bookings growth, new subscribers and those renewals and so.

On the subscription side and then you can add in as obviously an IAP.

On the new users Luis basically talked about how we have a really complicated.

Model and Markov model, if you want to read our blog post on our growth model how.

How we predict users and input into that is new users and thats a combination of organic growth.

In marketing expense, so we predict new users that flows into.

Some element of new bookings.

And then we're just basically doing cohort maps.

For our.

Cohorts of new users over time to see what we think they will convert into.

And then we're adding into that for the subscription business the renewal cohorts, which we have really good data on at this point.

I feel pretty confident about so you add that together your subscription bookings.

Got an ads in IEP, which again are.

<unk>.

Less detailed in the cohorts make a little less sense for that.

And then you've got bookings so it's really that simple.

Perfect.

That's super helpful. Thank you.

Okay, well, that's all we have for questions. So I'm going to turn it back over to Luis.

Thank you for all the excellent questions and please please I beg you to your dwelling a lesson.

Yes.