Duolingo (DUOL 2.01%)
This autumn 2022 Earnings Name
Feb 28, 2023, 5:30 p.m. ET

Contents:

• Ready Remarks
• Questions and Solutions
• Name Members

Ready Remarks:

Debbie Belevan

[Commercial break] We hope you loved that enjoyable TikTok. It is one of many many examples of natural learner-generated content material that goes viral day-after-day with none involvement from us. And we like to see the creativity and enthusiasm of creators like this, which continues to drive our model consciousness and word-of-mouth development. And now, let’s get on with at present’s name.

Good afternoon, and welcome to Duolingo’s fourth-quarter and full-year 2022 earnings webcast. All attendees are in listen-only mode. At the moment, after market shut, we launched our year-end shareholder letter with our This autumn and 2022 outcomes and commentary, which you will discover on our IR web site at buyers.duolingo.com. On at present’s name, we now have Luis von Ahn, our co-founder and CEO; and Matt Skaruppa, our CFO.

They may start with some transient remarks earlier than opening the decision to questions. Analysts will have the ability to ask a query utilizing the raise-hand characteristic. And please word this occasion is being recorded. Only a reminder that we’ll make forward-looking statements concerning future occasions and monetary efficiency, that are topic to materials dangers and uncertainties.

A few of these dangers have been set forth in our threat components of our filings with the SEC. These forward-looking statements are primarily based on assumptions that we consider to be cheap as of at present, and we now have no obligation to replace these statements on account of new data or future occasions. Moreover, we’ll current each GAAP and non-GAAP monetary measures on at present’s name. These non-GAAP measures should not meant to be thought-about in isolation from, or substitute for or superior to, our GAAP outcomes.

And we encourage you to think about all measures when analyzing our efficiency. And now, I’ll flip it over to Luis.

Luis von Ahn — Co-Founder and Chief Government Officer

Thanks, Debbie, and welcome, everybody. I am happy to report that we completed 2022 with a file quarter, attaining our highest-ever variety of DAUs and subscribers. We ended This autumn with 16.3 million DAUs, which is 62% year-over-year development, and we ended 2022 with 4.2 million paying subscribers, which is 67% year-over-year development. These metrics led 2022 to be our greatest 12 months ever when it comes to bookings and income, which grew 46% and 47%, respectively.

It was a incredible 12 months. Our DAU to MAU ratio, which is an indicator of consumer engagement, improved to 27% this previous quarter, up from about 24% a 12 months earlier than. We’ve additionally seen the variety of learners with lengthy streaks develop immensely. To remind you, a streak is the variety of consecutive days {that a} consumer learns with our app.

This mechanic has confirmed to be an environment friendly device for encouraging engagement and serving to us develop DAUs. 63% of our DAUs now have a streak of seven days or higher, which is up from 53% final 12 months. What’s much more spectacular is that we now have over 3 million DAUs with a streak of over one 12 months. That signifies that greater than 3 million individuals use Duolingo each single day for the final 12 months or longer.

It is also spectacular that it took us eight years to succeed in 1 million DAUs with a year-long streak and solely 21 months to get from 1 million to three million. That is one instance of the compounding impact of our product enhancements. Our consumer development was sturdy in each area of the world. We even have customers of almost all ages, of all socioeconomic statuses, and who’ve quite a lot of motivations to study, together with work, faculty, and journey.

The result’s a diversified enterprise that has been largely uncorrelated with macro and geographic developments. Our investments in monetization have additionally proven the ability of compounding. Over the previous three years, we have seen conversion enhance throughout almost all cohorts. Customers who’re new to the app have transformed extra readily to our paid subscription and customers who’ve been on the app for longer durations, like a number of years, have additionally seen a rise of their conversion charges from free to pay.

And now I might like to debate what it’s best to count on from us this 12 months. In a whole lot of methods, we count on 2023 to look just like 2022. We count on customers and bookings to develop properly. We nonetheless have an enormous $60 billion addressable market that we now have solely barely penetrated, and we now have sturdy natural development on our aspect.

And I need to remind you that at present, our subscribers make up about 8% of our month-to-month energetic customers. So, we now have loads of room to additional monetize our present consumer base. We even have a lovely alternative to extend bookings from ala carte in-app purchases like we did in 2022. However I do need to level out once more that you shouldn’t count on any bookings from our math or literacy apps this 12 months.

As to the place we’ll make investments, nearly all of our engineers, designers, and product managers will proceed to construct on the success we have had in development, efficacy, and monetization. That signifies that we’ll proceed to run experiments to extend DAUs and engagement, enhance how we train and optimize how we convert free customers to pay. And in our shareholder letter, I mentioned the 2 further areas we’re targeted on in 2023. The primary is generative AI, which is able to energy a higher-tier Duolingo Max subscription, in addition to assist us cut back content material creation prices; and the second is bettering how we train and monetize English learners.

I might like to emphasise that we’re enthusiastic about these initiatives, however they’re nonetheless within the early phases. And I additionally need to contact on profitability. Coming off a powerful 2022, we’re well-positioned to develop customers and bookings quickly whereas additionally delivering larger profitability. We plan to proceed to compound the returns we have seen from our previous R&D investments, so we do not plan to rent as many individuals as we did final 12 months.

Simply to remind you, we have by no means gone nuts with hiring. And actually, we have grown income at about twice the speed of headcount up to now 4 years. This 12 months, we plan to indicate working leverage throughout all prices, R&D, gross sales and advertising and marketing, and G&A. Given these strengths, we’re guiding to a ten% to 12% adjusted EBITDA margin for the 12 months, up from about 4% in 2022.

And with that, I will flip it over to Matt to speak extra about our monetary outlook.

Matt Skaruppa — Chief Monetary Officer

Thanks, Luis. To shortly recap the highlights of our spectacular This autumn and full-year 2022 outcomes. Within the fourth quarter, we delivered 39% bookings development 12 months over 12 months, which was about 46% on a constant-currency foundation. We noticed a 42% income development 12 months over 12 months, which was 47% on a constant-currency foundation.

We had a web lack of $13.9 million in comparison with a web lack of $17.5 million within 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 money circulation margin of about 11%. For the complete 12 months, we delivered 46% bookings development 12 months over 12 months, which was about 52% on a constant-currency foundation. We noticed 47% income development, which was about 51% in fixed forex.

We had a web lack of $59.6 million in comparison with a web lack of $60.1 million final 12 months. We noticed adjusted EBITDA optimistic $15.5 million in comparison with an adjusted EBITDA lack of $1 million final 12 months, and we had a free money circulation margin of about 12.5%. In 2023, we count on to proceed delivering sturdy bookings development, and we’ll accomplish that whereas making progress towards our long-term profitability purpose of an adjusted EBITDA margin of 30% to 35%. For the complete 12 months 2023, we’re guiding to $530 million to $542 million in complete bookings; $486 million to $498 million in income; and an adjusted EBITDA margin of 10% to 12%, which interprets to a few 32% incremental margin on the midpoint.

And for Q1 2023, we’re guiding to $127.5 million to $130.5 million in complete bookings, $111 million to $114 million in income, and an adjusted EBITDA margin of 9% to 10%. Our steering assumes present prevailing overseas alternate charges. As a reminder, roughly half of our income comes from outdoors the U.S., so each 1% enhance or lower within the worth of the greenback versus our basket of currencies has a few $2 million headwind or tailwind, respectively, on full-year complete bookings. On the prevailing alternate charges we have utilized in our steering, overseas alternate is predicted to be a three-percentage-point headwind to Q1 2023 year-over-year bookings development.

By way of quarterly cadence of our bookings for the rest of the 12 months, we count on our year-over-year development for complete bookings in Q2 to be about the identical as Q1. We count on Q3 year-over-year development charge to be barely larger, and we count on This autumn to be our largest quarter when it comes to greenback bookings. Given our outperformance in This autumn of this 12 months, it is going to be a barely decrease development charge. After pricing in 2022, we lowered pricing — costs in a number of nations to get our costs extra in keeping with every nation’s GDP per capita.

This initiative and our intentional combine shift from month-to-month to annual plans was to extend the general LTV of our platform, and that drove down subscription income per paid sub. However we’re finished, for probably the most half, with reducing costs globally. And actually, primarily based on sturdy conversion developments, we consider we now have the chance to boost costs in some locations and are experimenting with pricing as we converse. We’ve not included any materials quantity of bookings or income from Duolingo Max in our steering since we have solely simply began testing it, however we’ll give you an replace after we report our Q1 outcomes.

I might like to emphasise that we’re targeted on managing the enterprise in order that we obtain our full-year adjusted EBITDA margin of 10% to 12%. We’ll be making spending selections all year long primarily based on our income and gross margin developments to make sure that we’re on this vary for the complete 12 months. Additionally to remind you, given the seasonality of our income, our Q2 and Q3 margins shall be barely decrease than Q1, and This autumn shall be larger. We’ll proceed to run the enterprise with self-discipline and prudently handle our bills.

Beginning in Q1, we count on to see significant leverage in complete non-GAAP opex in comparison with the fourth quarter of final 12 months. And in 2023, we count on to realize about seven to 9 share factors of working leverage in complete non-GAAP opex. For each durations, the advance is generally in gross sales and advertising and marketing and G&A and, to a lesser extent, R&D. We ended the 12 months with roughly 48 million absolutely diluted shares excellent utilizing the year-end closing value.

In 2023, we count on to finish the 12 months with about 2% dilution from fairness issued to workers, which is lower than the roughly 3% dilution we had in ’22. And with that, I will flip it again to Luis.

Luis von Ahn — Co-Founder and Chief Government Officer

Thanks, Matt. Earlier than we get into Q&A, I might prefer to thank our very proficient crew of Duos, who proceed to seek out new methods to innovate and delight our learners. And now, we might be glad to take your questions. I will flip it again to Debbie to handle the queue, and right here we go.

Debbie Belevan

OK. Thanks, Luis. And as I discussed earlier, when you have a query, you possibly can simply use the raise-hand characteristic.  And the primary query comes from Ralph Schackart of William Blair.

Ralph Schackart — William Blair and Firm — Analyst

Good afternoon, and thanks for taking the query. First query, the shareholder letter, I believe you referred to as it a “tremendous” finish to 2022. I believe that will incorporate among the beginning-of-the-year marketing campaign that you just run to draw new customers. Simply curious, what drove the actually sturdy efficiency within the year-end? After which how did your New 12 months’s marketing campaign evaluate to expectations? And I’ve a follow-up.

Luis von Ahn — Co-Founder and Chief Government Officer

Thanks. That is an amazing query. So, sure, we had an amazing This autumn, nice year-end. The principle motive — there’s two predominant causes.

The primary one is simply our compounding product enhancements. I imply, we have simply been getting higher and higher at making our merchandise stickier and attracting extra customers. We’ve a really high-performing development crew that does that. That is one factor.

After which one other one is a few of our advertising and marketing campaigns, we simply did rather well. A very good instance of that was our year-end assessment marketing campaign the place towards — round December 14 or so or December 10 to December 14, all of the customers obtained a abstract of what they did all through the entire 12 months. And so they’re anticipated — properly, they’re inspired to share it. We obtained a whole lot of shares, tens of millions and tens of millions of shares.

And we hit prime trending subjects on Twitter, and stuff like that, by simply having individuals share their Duolingo stats over a number of days. So, stuff like that basically helped us do the outperformance. By way of New 12 months’s marketing campaign, we began — only for reference, we begin a New 12 months’s marketing campaign on December 28 normally yearly, and it goes all the best way to January. So, for this quarter, it is about three days-ish of marketing campaign, they usually carried out higher than we anticipated.

And the rationale for that outperformance simply has to do with the truth that yearly, we take the learnings from the earlier 12 months of the marketing campaign, and we simply get a little bit higher. So, yearly, we run just a few A/B exams to attempt to make that marketing campaign simpler, and we take these learnings for the following 12 months. So, that labored out fairly properly for us.

Ralph Schackart — William Blair and Firm — Analyst

Nice. Thanks. Perhaps only a fast follow-up. Simply when it comes to DAUs, I believe you reaccelerated for six consecutive quarters, which is fairly robust to do.

However simply possibly some coloration what’s driving the sturdy acceleration and development. I believe within the ready remarks, you talked about kind of broad-based. However had been there any kind of one or two areas that is stronger development than others? Thanks.

Luis von Ahn — Co-Founder and Chief Government Officer

Yeah. It is an amazing query, too. So, when it comes to areas, it truly is rising in each single area on the planet. If you wish to know stronger areas, the U.S.

is especially sturdy. Asia is especially sturdy, and Western Europe is especially sturdy when it comes to rising DAUs. And when it comes to why it’s, I imply, it is once more the identical two causes. It’s continued product enhancements that simply compound over time.

The product is simply higher and stickier, after which we simply get higher and higher at determining what strikes a chord with advertising and marketing. So, issues like our social media campaigns like our TikTok campaigns and our Twitter campaigns, and in addition among the influencer work that we do. And among the stuff, we do not even do. For instance, you simply noticed this video at the start of this name.

These are completely natural movies that different individuals make. It is nearly each single day. I do not know if it’s a truth that each single day on the planet, however nearly each single day, there’s an natural video that goes viral that mentions Duolingo. So, all of that simply retains getting — retains rising our customers.

Debbie Belevan

All proper. Thanks, Ralph. And the following query comes from Justin Patterson of KeyBanc.

Justin Patterson — KeyBanc Capital Markets — Analyst

Nice. Thanks very a lot. Luis, lately, there was a very fascinating put up on the Duolingo weblog about simply utilizing knowledge science to optimize DAUs. I believe present consumer retention charge was that the metric you’ve got been optimized up to now few years.

Would love to listen to extra about simply the way you’re utilizing a whole lot of these top-down statistical fashions to essentially drive DAU development, hyperlink the product crew collectively. After which now that you’ve got gone fairly far with the present metric, the way you’re eager about tailoring that to particular person cohorts? So, extra of a bottoms-up method to benefit from issues like, say, Gen Z, having a lot shorter consideration spans and ensuring they’re nonetheless having wholesome engagement. Thanks.

Luis von Ahn — Co-Founder and Chief Government Officer

Nicely, I am glad you learn that weblog put up. We’re very happy with our development mannequin. Through the years, we have simply gotten considerably extra refined at analyzing our consumer base and making adjustments that enhance sure cohorts. On this put up, we talked about what has been — I might say the only most vital metric for this firm during the last a number of years, which is known as — one thing we name CURR, or present consumer retention charge.

And what that’s, it’s the charge of those that in the event that they had been right here at present they usually had been right here during the last one different time, no less than during the last seven days, what’s the chance they are going to come again tomorrow, that is present consumer retention charge. That quantity is round 80% proper now, and we have elevated it. During the last two, three years, we have elevated that quantity from about 65% to about 80%, and each 1% enhance of CURR elevated our DAUs by quite a bit. And so, that mannequin has been — actually helped us, and we maintain getting increasingly refined.

You’re proper. We’re beginning to — this was a mannequin — at first, while you’re simply — earlier than this development mannequin, we might simply deal with all our customers in a single cohort. Then we began getting extra refined, and we deal with sort of new customers, present customers, and many others. Now we’re getting much more refined than that.

I do not suppose we now have a particular factor for Gen Z that we do. However we’re — relatedly, we’re making our classes shorter and shorter as a result of, you recognize, such as you stated, they’ve a shorter and shorter consideration span.

Debbie Belevan

Nice. The following query comes from Andrew Boone of JMP.

Andrew Boone — JMP Securities — Analyst

Nice. Thanks for taking my questions. Can we begin off with simply any learnings from in-app purchases? I noticed that it doubled this final quarter. Are you able to discuss what successes are there? After which how we should always take into consideration that for 2023? After which secondly, on profitability, Matt talked about actual contribution margin sort of particularly within the again half of this 12 months.

Is there any motive to suppose that that steps down as we take into consideration sort of a three-year plan or one thing past this upcoming 12 months? Thanks a lot.

Luis von Ahn — Co-Founder and Chief Government Officer

I can take the IAP query, Matt, after which you possibly can take the following one. So, for in-app purchases, sure, we now have a really high-performing crew behind this that’s simply making in-app purchases increasingly distinguished, increasingly profitable within the app. As you noticed, they have been rising very quickly. We count on them to proceed rising quickly this 12 months.

And when it comes to learnings, the principle approach we do in-app purchases, by the best way, is by promoting gems. So, there’s this in-app forex referred to as gems which you could both earn by doing stuff on the app or it’s also possible to purchase them. After which you should use gems to purchase different stuff, sort of power-ups within the app. One of many predominant power-ups that we use within the app is the timer boosts that assist for sure timed classes.

So, we now have these — most classes on Duolingo should not timed, however there are some classes that you are able to do — some apply classes which might be timed. And the concept is that for those who’re not very quick, you run out of time or you should buy these timer boosts with gems. And people timer boosts get lots of people to spend on in-app purchases. That is the kind of stuff you are going to get to see us do.

So, basically, for us, for in-app purchases, we’re not going to be promoting content material inside our purchases, however primarily we’ll be promoting gamification issues that individuals actually like. And so, far, that technique has been working fairly properly.

Matt Skaruppa — Chief Monetary Officer

Nice. After which on profitability, so Andrew, the profitability elevated in 2022 by about 450 foundation factors from 2021 on an adjusted EBITDA foundation, which was good progress towards our long-term profitability purpose, of that 30% to 35%. This 12 months, we’re guiding to 600 to 800-basis level enchancment over that. And so, as we glance out sort of into the long term, our purpose hasn’t modified.

We need to be at a 30% to 35% adjusted EBITDA margin. And between final 12 months and what we’re guiding to this 12 months, we predict we’re exhibiting the best progress towards that purpose.

Andrew Boone — JMP Securities — Analyst

OK. Thanks.

Debbie Belevan

Thanks. After which the following query comes from Ryan MacDonald of Needham.

Ryan MacDonald — Needham and Firm — Analyst

Thanks for taking my questions, and congrats on a powerful near the 12 months. Luis, let me begin with you on the Duolingo English take a look at. There was lately a ban in China from the federal government on on-line levels, sort of a Chinese language citizen basically taking a completely on-line diploma from a world college. They’re now not recognizing that diploma.

And it looks like that is going to start out creating kind of a shift in an inflection level, and extra Chinese language learners seeking to go in individual abroad. I am simply curious for those who’ve seen any impacts of this to date on kind of functions or sign-ups for the Duolingo English take a look at and the way you suppose — or what your expectations for which might be in 2023 right here.

Luis von Ahn — Co-Founder and Chief Government Officer

Nicely, thanks for the query. We’ve not seen any affect primarily based on this, and I believe there’s many causes. I imply, for one, sure, China is a comparatively giant chunk of the Duolingo English take a look at takers, however we actually have take a look at takers all around the world. The opposite factor is the explanations for taking the exams are very diversified.

A few of them is people who find themselves making use of for worldwide undergraduate admissions. There’s additionally graduate admissions. There’s transnational training. There’s every kind of various pockets.

And possibly one pocket has some kind of coverage implications to it. However basically, the pockets are fairly wholesome. That is simply sort of one factor. The opposite factor is we’re nonetheless — we aren’t the chief on this market.

We’re coming into this market with a web based take a look at, and it’s a market that normally has had all these offline exams. So, on condition that we’re nonetheless within the sort of development section of this, we do not — it would not have an effect on us an excessive amount of when there’s one thing that adjustments the complete quantity of it as a result of we’re coming in with no matter share of exams we now have, which is considerably below 25%. We’re nonetheless rising fairly a bit and fairly quick, so this isn’t one thing that we’re seeing but.

Ryan MacDonald — Needham and Firm — Analyst

That is useful. Recognize it. After which possibly as a follow-up. You talked quite a bit about generative AI and using that, and I am very curious across the content material creation.

You talked about an elevated give attention to kind of extra superior English content material for non-English audio system or primary English audio system. How shortly do you suppose that the generative AI can compress your content material creation timeline? And the way do you handle for high quality in that? After which for Matt, are you assuming any — throughout the leverage that you just’re exhibiting this 12 months on the revenue margin aspect, are you assuming any leverage from that content material creation from generative AI? Thanks.

Luis von Ahn — Co-Founder and Chief Government Officer

Yeah. Thanks for the query. So, we’re very enthusiastic about generative AI. And simply to place issues into context, at Duolingo, we have all the time leaned into AI from the start.

I imply, a whole lot of our merchandise are constructed primarily based on AI, and many others. During the last couple of years, generative AI has turn out to be considerably higher, and we’re positively going to lean into this as properly. You talked about content material creation. That is one of many locations the place we predict that it may well assist us quite a bit and velocity issues up fairly a bit.

It’s kind of early for me to present you a exact estimate of how a lot you possibly can velocity it up, however we’re sure — I imply, it is already occurring. We’re already rushing issues and in addition reducing prices. One of many locations the place we’re utilizing that’s to have the ability to create a way more English studying content material. So, you may see us create much more of that.

By way of high quality assurance, we’re nonetheless going to have people go over it. However the timeframe, it was the case that each one of this was nearly completely generated by people. By now, it simply must be reviewed by people, and it is quite a bit faster to assessment than to generate. So, when it comes to time-frame, we’ll get an enormous enhance in that, and I assume it is the identical for value.

So, sure, that is one thing we’re very enthusiastic about. And the opposite factor I ought to point out about generative AI, so content material creation is certainly one of them, and we’ll positively use that. Nevertheless it additionally goes to assist us simply create higher experiences to our customers, sort of extra on-line conversational apply that we’re going to have the ability to do. And that is the kind of stuff that we’re placing into our new higher-tier subscription, Duolingo Max.

So, that is one thing we’re very enthusiastic about. As an organization, we have partnered with open AI, and we’ll be working very laborious on this.

Matt Skaruppa — Chief Monetary Officer

Yeah. And, Ryan, Luis answered the query for the information. I imply, we solely included within the information issues we now have direct line of sight into. It isn’t — generative AI is so new.

We do not have a ton of line of sight into direct financial savings but. However as Luis stated, over time, that is most likely the best way to guess.

Debbie Belevan

Thanks, Ryan. And the following query comes from Aaron Kessler of Raymond James.

Aaron Kessler — Raymond James — Analyst

Nice. Thanks. A few questions. Perhaps for Matt.

On the paid conversion charge, it was very sturdy in This autumn. Are you able to simply discuss that? Was that household plan part of that? After which how ought to we take into consideration paid conversion in 2023? After which possibly any replace on a given take a look at, possibly sort of human tutoring as properly. Is there any sort of ideas on that course as properly, particularly as possibly it pertains to Duolingo Max, the way you’re speaking about — eager about integrating that as properly?

Matt Skaruppa — Chief Monetary Officer

Certain. Sure. I will let Luis reply that final half. For the payer penetration, we had talked about, I believe all through most of final 12 months, so we anticipated a 1% to 2% enhance every quarter 12 months over 12 months in that metric.

And that is coming from all the issues that Luis has described earlier than, that each one the product enhancements that compound over time to make our free-to-pay conversion go up, basically. There wasn’t anyone specific merchandise like household plan or this or that. It was all of them compounding over time. I believe going ahead, as we talked about on, I believe, the Q2 or Q3 name final 12 months, given the expansion in MAU, in order that denominator of that quantity is simply rising actually, actually properly, and the retention developments, we might count on it to truly as an alternative of including one or two factors each quarter 12 months over 12 months, be nearer to 1% this 12 months, simply because, once more, that denominator is rising actually properly.

So, it should finally flip into subscribers over time. However within the brief time period, it is going to be about one level. After which, Luis, you need to discuss Max and tutors?

Luis von Ahn — Co-Founder and Chief Government Officer

Yeah. On human tutoring, I imply, the primary query was about human tutoring — or the second query was about human tutoring. Through the years, we have waffled with human tutoring. At first, we launched Duolingo with no people concerned.

Then we tried it. We tried placing it in. We tried — and during the last 12 months, we have made the choice that I believe is the best choice that we aren’t going to get into human tutoring in any respect. We have tried it.

We’re out. There’s a variety of causes for that. There’s nothing — it is not like we now have something towards human tutors. It is simply on our finish, we’re a know-how firm, and that is what we excel at.

And along with that, for those who have a look at it, human tutors are nice, however they are not getting any higher. They’re simply there. Whereas know-how, particularly AI, and particularly a whole lot of the stuff coming from generative AI, is simply getting higher and higher yearly. So, we’ll guess on that.

A lot in the identical approach, like 10 years in the past, we had been informed we couldn’t train a language with synthesized computerized voices. We guess on that, despite the fact that 10 years in the past, synthesized computerized voices nonetheless sounded sort of robotic. At the moment, they sound glorious. We really feel the identical with AI that in 5 years, I do not know, 5, 10 years, we shall be very glad that we’ll have that on know-how moderately than people.

Aaron Kessler — Raymond James — Analyst

Thanks.

Debbie Belevan

And subsequent query comes from Nat Schindler from Financial institution of America.

Nat Schindler — Financial institution of America Merrill Lynch — Analyst

Sure. Thanks. I really need to simply go a little bit deeper on among the questions you had earlier than, actually about how the customers have modified. You noticed a dramatic acceleration in each customers and paid subscribers over this 12 months versus the prior 12 months.

Has there been an actual change within the demographics of these? Is there something considerably totally different about customers this 12 months than there have been final 12 months?

Luis von Ahn — Co-Founder and Chief Government Officer

It is an amazing query. If there was, it is small. Usually, our consumer base may be very extensive. We’ve customers from each single nation on the planet.

Each single nation on the planet is represented. We’ve customers from each single socioeconomic standing from all ages so long as they know the way to learn. So, it is sort of from six years to about 100. We simply actually have customers from in every single place, and it is a very extensive consumer base.

So, it hasn’t modified an excessive amount of. The one factor we have seen is it has gotten barely, ever so barely youthful, and that most likely has to do with our — a whole lot of our prominence on social media like TikTok, however that is it. So, we’ve not actually seen an excessive amount of of a change.

Nat Schindler — Financial institution of America Merrill Lynch — Analyst

And as these adjustments towards going barely youthful, has that occurred in each free customers and paid?

Luis von Ahn — Co-Founder and Chief Government Officer

The trustworthy reply is I do not know. I do know that it is — free customers is getting barely youthful. I do not know for paid.

Matt Skaruppa — Chief Monetary Officer

Sure. I might have to return and double-check, Nat, on the paid. I believe what we have seen on paid is that the conversion developments have not actually modified. Some have accelerated.

For instance, simply general conversion of sort of cohorts throughout the board has elevated over the course of final 12 months. So, of us who’re new to the platform within the first couple of months, they transformed extra quickly than in 2021. People which have been on the platform for an extended time period, they transformed a little bit bit sooner than that they had up to now. We have all the time seen that individuals who use the app extra convert extra.

That stayed the identical or obtained a little bit bit higher. So, sure, it was sort of a broad base. Similar to our customers grew world wide in a reasonably broad-based approach, the identical development on the free to paid conversion.

Nat Schindler — Financial institution of America Merrill Lynch — Analyst

Thanks very a lot. And as soon as once more, very spectacular.

Debbie Belevan

Thanks, Nat. And subsequent query comes from Mark Mahaney of Evercore.

Mark Mahaney — Evercore ISI — Analyst

Two questions. You had this line within the launch about how the variety of returning customers is roughly the identical because the variety of brand-new customers. I might simply have assumed that that has all the time been the case, or it has been the case for some time. However is that one thing new that is occurred? As a result of I get your level about how individuals will use it, after which they will both come again to it once they understand significance in studying a brand new language or there will be one other alternative.

Perhaps they’re touring to a unique nation. So, is {that a} new development or not? What is the so what out of that? And simply actual shortly, Duo Math, although, that math product, simply the place is that? Are you continue to tweaking it? Is it removed from being the place you need it to be? Simply discuss sort of adoption, how a lot curiosity you’ve got actually seen in that. Thanks.

Luis von Ahn — Co-Founder and Chief Government Officer

To your first query, this isn’t a brand new factor. It has been some time since returning customers are the identical as new customers. I will assume on the very early of Duolingo, that was not the case. Nevertheless it’s been some time that has been like that.

And also you’re proper. It is as a result of it’s uncommon for us to see a consumer that makes use of Duolingo goes away and by no means comes again. That’s uncommon. Normally, individuals might even go away for 2 years after which come again.

It is simply because there’s such a little bit friction. So, we see that. And the rationale we put that there’s primarily to emphasise that it’s not proper to think about our customers as coming, having churn, and been gone ceaselessly. That’s uncommon.

It is extra that they arrive. They keep for just a few months or one thing, then they might take a break for a few months, they usually come again or one thing. So, one thing like that. By way of math, we’re very pleased with the progress.

It is rising quickly. We’re pleased with that. Simply to remind you of the plan, we’re — proper now, the purpose is to develop our math product organically for it to develop by itself. As soon as we get to a reasonably respectable variety of customers, name it, 1 million day by day energetic customers, that is sort of — it would not need to be precisely 1,000,000.

As soon as we get there, we’ll say, OK, nice. We’ve good product market. It’s time to begin monetizing. We’re not there but, however it’s rising fairly properly, and so we’re pleased with the outcomes.

By way of tweaking company, we’re working very laborious on tweaking the app. We’re including a whole lot of the options that we all know work properly on the principle Duolingo app to the mathematics product. So, enhancements to the Road. We’re considering of including issues like leaderboards, higher notifications.

So, a whole lot of the stuff that we have realized during the last 10 years on Duolingo and that we now have the code for, we’re beginning to add to the mathematics app to get it — to have larger retention. We’re additionally including extra content material to it. One of many issues that occurs with the mathematics app proper now could be individuals really run out of content material as a result of in case you are very — it seems we now have some very heavy customers that use the entire thing, after which they sort of run out of content material. So, we’re including extra content material to it.

So, you may see us over the following 12 months simply make the product higher and higher. And due to that, it should proceed rising.

Mark Mahaney — Evercore ISI — Analyst

OK. Thanks, Luis.

Debbie Belevan

Subsequent query comes from Eric Sheridan of Goldman Sachs.

Eric Sheridan — Goldman Sachs — Analystb

Nice. Thanks. Perhaps first query could be following up on one thing that was implied in Mark’s first query. Journey has clearly been a giant tailwind, and the shift to companies over items has been a giant tailwind within the macroeconomic setting.

Are you able to speak a little bit bit about what you’ve got seen with journey probably as a catalyst for utilization, each prime of the funnel and the underside of the funnel as that is continued to construct momentum and the way possibly on account of cross-border journey persevering with to return, that might be a pleasant tailwind for the enterprise going ahead over 2023 and past. After which the second dynamic could be you talked quite a bit about virality in TikTok. Are you able to speak a little bit bit about the place you’re excited to experiment and take into consideration advertising and marketing or constructing virality on platforms on the content material aspect or the distribution platform aspect, possibly even away from TikTok, and the way we ought to be eager about that constructing within the years forward? Thanks a lot.

Luis von Ahn — Co-Founder and Chief Government Officer

Nice query, Eric. Let’s have a look at. By way of journey, we do not see in our metrics that increasingly individuals are utilizing Duolingo for journey. We did see a really momentary dip within the peak of COVID a few years in the past.

That is sort of a fewer — barely fewer. Since then, it has been about the identical. And we simply — journey is fascinating. For lots of people like us which might be, to be trustworthy, from the worldwide perspective, fairly rich.

I believe journey is sort of a massive motivation for utilizing Duolingo. However for lots of our customers, they’re utilizing it to study English or as a interest or one thing. So, we do not see this stuff that is like, “Oh, my god, journey has opened up, and due to this fact, Duolingo is exploding.” That simply — we simply have not seen a lot of that. We do ask when individuals enroll, why they’re utilizing Duolingo.

And a few fraction inform us they’re utilizing it for journey. I do not know the quantity off the highest of my head, I will assume it is one thing like 15% to inform us that they are utilizing it for journey. Nevertheless it’s simply not one thing that we glance an excessive amount of to. By way of your second query about virality.

So, to this point, we now have experimented with just a few social media platforms to get our phrase out, by the best way. And that is primarily natural, as in we do not pay for these things. We simply make actually good content material. TikTok has been actually good for us.

Not simply TikTok within the U.S., we have really localized our TikTok to Portuguese and Spanish and different languages. So, one of many issues that we’re enthusiastic about is TikTok accounts within the different languages. They’re really rising actually quick. And it is really wonderful how — by the best way, that is simply wonderful factor about Brazil.

Something you do with social in Brazil does properly. And so, as quickly as we put TikTok in Portuguese, it simply sort of went nuts. And for those who go see these movies, they’re wonderful. We simply had Duo dressed up as Wednesday from the Netflix present, Wednesday.

And we obtained, I do not know what number of, 10 million views or one thing. So, we’re enthusiastic about platforms like TikTok and truly additionally Twitter. We’re doing fairly properly in Twitter. Now, while you ask about virality basically, one other factor that we’re doing is we’re simply including much more issues contained in the product.

This isn’t social media or something, contained in the product to get extra individuals to share issues. And we now have a crew behind it. It is a actually good crew, too, and they’re able to get much more individuals to share issues. One of many predominant issues that individuals share, there’s now a strategy to share sentences rather more simply.

One of many predominant issues individuals share are bizarre sentences that they see on Duolingo. So, among the programs, for one motive or one other, have a really bizarre sentences. I do not know if I can consider certainly one of them. Just like the German course has one thing like — the sentence has one thing like, “proper subsequent to the mattress, lay the physique of the husband,” or one thing like actually bizarre stuff that individuals share.

And that simply will get lots of people. So, we’re engaged on getting individuals to share like that, and it is figuring out fairly properly. I need to say Matt most likely wrote that query in regards to the — that sentence in regards to the husband. However OK.

Debbie Belevan

Subsequent query comes from Mario Lu of Barclays.

Mario Lu — Barclays — Analyst

Nice. Thanks for taking the query. So, the primary one is on app retailer charges. Simply curious to your thought on the current development of apps bypassing the app retailer.

For instance, video games like Fortnite and apps like Spotify, Netflix have finished this for some time. However extra lately, extra conventional cell video games are implementing like a direct-to-consumer cost methodology. Simply curious your ideas on any initiatives that you just guys have in your finish.

Luis von Ahn — Co-Founder and Chief Government Officer

We’re not at present engaged on that. For us, we love the app shops. Clearly, it is laborious to say that we would not find it irresistible in the event that they lower their charges. After all, we might.

However we do rather well with the App Shops. They’re an incredible distribution mechanism. And the factor while you go off app shops is that it is tougher to distribute your product. And likewise, it has fairly a little bit of friction.

One of many wonderful factor in regards to the app shops is that individuals have already got the bank card in there, they usually simply need to faucet as soon as and it will get to there. So, you might be able to decrease your charges, however you add a lot friction to your customers. Nevertheless it’s not clear to us that we might make that rather more cash. It is one thing that we’re all the time eager about and will ultimately do it, however it’s not one thing that we’re at present engaged on an excessive amount of.

Matt Skaruppa — Chief Monetary Officer

And bear in mind, Mario, we did get a fairly important bump from Google reducing their app retailer charges final 12 months.

Mario Lu — Barclays — Analyst

Obtained it. Useful. After which second one simply on sort of adjustments to the Duolingo app itself. I consider previous few quarters, you talked in regards to the change to the only studying observe being at parity.

Any updates to that? And simply basically, how do you stability between sort of optimizing for studying and versus having a extra informal app that sort of attracts the plenty? So, I believe simply consumer feedbacks, might be like a lab minority. However the elimination of like issues like oral prompts or sort solutions, I consider obtained some destructive suggestions. I simply needed to listen to your ideas there. Thanks.

Luis von Ahn — Co-Founder and Chief Government Officer

Mario, you learn Reddit.

Mario Lu — Barclays — Analyst

I do.

Luis von Ahn — Co-Founder and Chief Government Officer

Yeah. So, it is an amazing query in regards to the adjustments to the app. So, we made some fairly important adjustments. We’re all the time making the app higher.

I imply, that is what we work on. For this reason we spent a lot on R&D. We’re actually all the time making the app higher. During the last 12 months, we made this huge change to our house display, going from — giving a whole lot of alternative about what to study subsequent to this linear path.

The principle motive we did that linear path, by the best way, was to show higher. We seen that we had been — after we had been giving individuals a lot alternative, what was occurring was a major fraction of them, we’re simply coming in to increase the streak with very easy classes that we’re not really getting them to study. Whereas now with the linear path, we have inspired a considerably bigger fraction of customers to every time they arrive to extend their streak, they really additionally study one thing. So, the principle motive we did that was really to show higher and in addition to simplify our product, which has actually helped us.

However that was really to show higher. The best way we consider it’s after we’re making adjustments, we have a look at — we now have three issues that we’re bettering. We’re bettering how properly we train. We’re bettering how partaking the app is, so mainly, how a lot time individuals spend on it, how — retention, and many others., and we’re bettering how properly we monetize.

And we now have giant groups behind every certainly one of this sort of group of metrics. And what occurs is that every crew, for instance, there’s the monetization crew, what they’re making an attempt to do is get extra individuals to subscribe. Each time they run a take a look at to attempt to get extra individuals to subscribe, what we implement is that they can not screw up the opposite metrics. So, sure, you will get extra individuals to subscribe, however you can not lower our variety of DAUs.

You can not lower the time individuals spend studying, otherwise you can not lower the training final result. Equally, after we get individuals to — the group of individuals that’s devoted to show higher, we inform them, OK, you may make adjustments that train higher, however you can not screw up the monetization otherwise you can not screw up the consumer engagement. That is how we stability it. And we discovered that that works fairly properly.

And most each change within the app, you possibly can really hint it down. That change was made by the group or by the crew of individuals that’s making an attempt to make us develop sooner or make us earn more money, and many others. And so, that is how we do it. Hopefully, that offers you some coloration.

Mario Lu — Barclays — Analyst

Yeah. Very useful. Thanks.

Debbie Belevan

Our subsequent query comes from Arvind Ramnani from Piper Sandler. You are muted, Arvind.

Arvind Ramnani — Piper Sandler — Analyst

Yeah, I caught it right here. I assume my first query is, among the many administration crew, who has the longest streak when it comes to utilization? However let me now ask that.

Luis von Ahn — Co-Founder and Chief Government Officer

However I do not know the reply to the query, however I am keen to guess it is me, however I do not know the reply to the query.

Matt Skaruppa — Chief Monetary Officer

We may be fairly positive about that.

Arvind Ramnani — Piper Sandler — Analyst

All proper. Thanks. Yeah, yeah. So, a very good metric with consumer engagement and sort of the income development and bookings.

However I do know there’s a complete lot of different metrics you observe that — past what you may share with us. So, are you able to — conceptually, for those who’re not in a position to share laborious numbers, can you conceptually share a bit extra beneath on among the metrics that is supplying you with confidence in sort of the form of the expansion for this 12 months?

Luis von Ahn — Co-Founder and Chief Government Officer

Yeah. I imply, I do not know if we are able to share exact numbers of — there’s sure issues that we share and stuff that we do not share, however we do observe a whole lot of metrics. Usually, as I used to be saying earlier than, there are three teams of metrics. There’s what we name the expansion metrics, which is simply normally our complete consumer base.

Are they utilizing the product extra? Do we now have extra DAUs, and many others.? There’s the monetization metrics, which is how a lot cash we make, mainly. After which there’s the metrics about how properly we train. In every certainly one of them, we now have a ton of metrics. By way of development, among the issues that no less than I have a look at, I have a look at time spent studying.

I have a look at DAU to MAU ratio. I have a look at resurrection charges, so sort of variety of individuals which might be being resurrected. I have a look at the retention of resurrected customers. I have a look at the retention of present customers.

I have a look at the retention of latest customers. So, that is the kind of stuff. In monetization, the sorts of stuff we have a look at are issues like conversion from free to paid, LTV, retention of paying customers. Every day bookings is one thing we have a look at quite a bit, variety of customers or a fraction of customers which have an IAP or do a purchase order on an IAP.

In studying, we have a look at a variety of issues, however one of many predominant issues we have a look at is problem of the content material that we’re in a position to give out with out scaring customers all away. So, there’s a whole lot of metrics that we observe, and I do not even know all of them off the highest of my head.

Arvind Ramnani — Piper Sandler — Analyst

Yeah. That is useful. After which, you recognize, sort of bookings development, sort of sitting externally in some sense, it’s kind of of a black field. However can you share like while you mission for bookings development, what are the inputs that go into sort of coming as much as the kind of projections for bookings development? I imply we do not even know the particular numbers.

However simply conceptually, what goes into predicting consumer development — reserving development, sorry?

Luis von Ahn — Co-Founder and Chief Government Officer

Yeah. We simply take a ruler and go up like that. It really works fairly properly. I am kidding.

Matt has a significantly better reply than that.

Matt Skaruppa — Chief Monetary Officer

Nicely, you recognize, possibly not. I imply, the principle factor to consider with bookings development is that there is simply the 2 dimensions. There’s new bookings development, new subscribers, and there is renewals, and so — on the subscription aspect. After which you possibly can add in provides, clearly, and IAP.

On the brand new customers, Luis mainly talked about how we now have a very difficult mannequin, a markup mannequin, if you wish to learn our weblog put up on our development mannequin, how we predict customers. An enter into that’s new customers, and that is a mix of natural development and advertising and marketing spend. So, we predict new customers. That flows into some factor of latest bookings.

After which we’re simply mainly doing cohort math for our cohorts of latest customers over time to see what we predict they may convert into. After which we’re including into that for the subscription enterprise, the renewal cohorts, which we now have actually good knowledge on at this level. We really feel fairly assured about. So, you add that collectively, you get your subscription bookings.

Add in provides and IAP, which once more are a bit much less detailed within the cohorts, make rather less sense for that. And then you definately obtained bookings. It is actually that easy, Arvind.

Arvind Ramnani — Piper Sandler — Analyst

Excellent. That is a mouthful. Thanks.

Debbie Belevan

Nice. Nicely, that is all we now have for questions. So, I will flip it again over to Luis.

Luis von Ahn — Co-Founder and Chief Government Officer

Thanks for all the wonderful questions. And please, please, I encourage you. Do your Duolingo lesson.

Period: 0 minutes

Name individuals:

Debbie Belevan

Luis von Ahn — Co-Founder and Chief Government Officer

Matt Skaruppa — Chief Monetary Officer

Ralph Schackart — William Blair and Firm — Analyst

Justin Patterson — KeyBanc Capital Markets — Analyst

Andrew Boone — JMP Securities — Analyst

Ryan MacDonald — Needham and Firm — Analyst

Aaron Kessler — Raymond James — Analyst

Nat Schindler — Financial institution of America Merrill Lynch — Analyst

Mark Mahaney — Evercore ISI — Analyst

Eric Sheridan — Goldman Sachs — Analystb

Mario Lu — Barclays — Analyst

Arvind Ramnani — Piper Sandler — Analyst

Extra DUOL evaluation

All earnings name transcripts