DUOL EPS is expected to grow by 86% Y/Y this year and by 29% Y/Y next year. Those growth estimates will rise following this report.

The new, GenAI-fueled Duolingo Max subscription tier is already contributing materially to bookings. This is a big step in its mission to teach as well as a human tutor. Unsurprisingly, this is bolstering customer lifetime value (LTV), retention and word-of-mouth growth. Its financial impact will merely rise as awareness grows and it rolls it out to the remaining 50% of its DAUs. Among the most popular tools is the live video chat that Duolingo debuted at Duocon in September. As a reminder, this product allows users to practice natural conversation. This is enabled by powerful, OpenAI-augmented, back-end algorithms responding through its familiar Lily character on the front end. The product has memory, and can bring up previous conversations to enhance individualization. So far, English learners are using it at 2x the pace of all other languages; again, this supports Duolingo’s large aforementioned advanced English opportunity.

Duolingo showed Lily’s new talents off to start its own earnings call. She filled in for Luis to kick things off in her typically sarcastic tone. This technically could have been a pre-recording, but when you actually use this new tool in-app, it will become clear that this character doesn’t need a fixed script.

This video product led to a surge in Duolingo Max bookings, which powered part of the 8% outperformance it posted. Duolingo did say it isn’t banking on this same tailwind next quarter. But? If history is any indication at all, this is it leaning pessimistic to tee itself up for more outperformance. They tell us all the time about why growth cannot possibly stay as strong as it is… and then it stays as strong as it is. Under-promise, over-deliver.

Duolingo Max generates more gross profit dollars per plan, but lower gross margin. This is related to incremental large language model (LLM) querying costs, which is a trade-off the company is happy to make. That’s understandable, considering Max also generates more cash flow per plan than other tiers. And while Duolingo Max LLM and amortization costs will lead to 100 bps of Q/Q GPM contraction next quarter, that will not be permanent. As leadership explained, it tasked its engineers with rolling out great products as quickly as possible. Per the team, they told these builders “not to worry too much about cost.” That optimization work will happen over time. The priority was driving expeditious product-market fit amid this rapidly evolving landscape. Leverage is coming.

As a related aside, tanking inference costs should embolden Duolingo to pursue more complex conversational tools. The trend will also make existing tools like the FaceTime product affordable for lower income nations.

Duolingo’s unmatched combination of DAU growth and scale is being powered by the same things as always. What you may ask? Consistent, thorough product split-testing to directly observe what works & craft products around those insights. It has accelerated its ability to run these tests in recent quarters, with the same success rate as before. This means ever faster product iterating and improving. It gets as precise as which shade of a certain color works in a certain setting to boost engagement by some microscopic amount. As leadership always says, these improvements compound over time, and they’re right. This fantastic quarter was simply more evidence. Duolingo is not an external marketing-led growth engine like its competition… It's an R&D-led growth engine. 

This time around, leadership credited the strong DAU results to making the core app more social. For example, it extended user streaks to multiple learners with its friends streak debut. The user streak created an observed sense of competition and urgency to use the app more frequently, so engagement gains were large. By making this same idea more social, it’s notching incremental boosts. This may be a cute, cuddly owl on the outside... but it is a lean, mean, outcompeting machine on the inside. The team has a lot to do with that, as education technology is a very hard place to sustainably win. 54% DAU growth while lapping 60%+ DAU growth last year (and its margins) is the result. 

Notably, the company enjoyed half of its active user growth coming from previously lapsed MAUs. The new social streak and other similar tools like friend quests, it thinks, are the reasons for the success. USA DAU expansion is also “similar” to overall 54% Y/Y growth, so this success isn’t being powered by lower monetizing counties.

“How did we pull this off? You know, the usual stuff, product improvements and social marketing. It just works.”

Lily

As expected, the rollout of 20 new advanced English courses is receiving a warm welcome. Duolingo has already racked up 2 million DAUs here in a handful of months. To the team, there’s a long runway for making even more complex content to tap into the “enormous global demand” and to “unlock more user growth.”

This was a great quarter. The stock could easily cool off after a monstrous 3-month run, but the company is on a phenomenal path. It’s executing at a level of growth, scale and margin that nobody else can match. Nobody. I’ve published two pieces recently that included two separate Duolingo trims. The sales had nothing to do with fundamentals. They had everything to do with soaring estimates and expectations that were becoming wildly difficult to surpass. It was also a response to modest multiple expansion; this firm’s PEG ratio has risen from around 1.2x to probably 1.5x-1.6x following profit revisions. I don’t think it’s unfairly expensive, but I do think it’s now fairly priced.

When investors get this excited and this uniformly positive on a name, great quarters (like this absolutely was) are routinely shrugged off. This stock probably needs to take a breather, and I’d love to add some of the sale proceeds back if that occurs. Who knows what will happen tomorrow. Zooming out, things look great.

Celsius missed greatly lowered analyst revenue estimates by 0.7%.

Celsius doesn’t offer much formal guidance. It was asked if “low-to-mid-single digit revenue growth” for next quarter before the expected Pepsi impact (much more later) is about right. It said “yes.” I don’t know why they can’t just tell us this before being asked, but I was grateful for the response. Guidance was roughly in line with expectations.

Celsius EPS is expected to be flat Y/Y this year and grow by 29% Y/Y next year.

To tee this conversation up, here is how Circana defines various key market share metrics:

Considering Amazon and club channels are key sources of growth for Celsius, I prefer using the last two metrics. Still, all four are valuable.

For MULO+ W/C, retail dollar share trends remain reasonably fine on a Y/Y basis, and not so fine on a Q/Q basis. It jumped from 11.2% to 11.8% Y/Y, and fell quite materially from 12.3% to 11.8% Q/Q. Most recently, through the month of September, market share was 11.6% vs. 11.5% Y/Y and fell Q/Q. That is not what I want to see to justify making this a more legitimate piece of my portfolio. This screams to me “do not add… do nothing.”

It is worth noting that mix shift to convenience store (c-store) revenue is hurting market share for metrics that include c-stores vs. those that don’t. It enjoys lower overall market share there vs. Amazon and club channels. Still, that’s a minor factor in terms of overall Q/Q market share weakness.

I know… It’s probably confusing to see positive Y/Y market share dynamics, yet deeply negative Y/Y revenue growth. A small part of the reason is its observed rise in promotional activity from retail partners, which affects its own dollars per can. More importantly, this is mainly related to how Celsius collects revenue. It does not earn revenue when a can is sold at a Costco or a 7/11. It collects revenue as it delivers orders to its main distribution partner, Pepsi, in the USA. As we’ve written about extensively, Pepsi has been cutting the amount of Celsius inventory they hold. Either they became too excited with orders while times were fun and needed to fix that blunder… or? Perhaps they’re trying to squeeze what is still the best growth story in energy drinks so they can just buy Celsius at a discount. That’s pure speculation, but it does make sense to me.

Fortunately, retail sales growth, which measures how much Celsius is selling in stores, continues to outperform the sector comfortably. CELH posted 7.1% Y/Y growth vs. 2% for the field. This points to rapid brand decay not being the root cause. Unfortunately, Pepsi inventory reductions were ever sharper than the high end of estimates, and that fostered the revenue shortfall. Specifically, Pepsi was a $124 million revenue headwind Y/Y and the main reason for the decline.

CEO John Fieldly had somewhat encouraging things to say about this process perhaps peaking in intensity:

“Pronounced supply chain optimization by our largest distributor, which we believe has largely stabilized, had an outsized and adverse impact on our operating results during an otherwise solid quarter.”

CEO John Fiedly

During the Q&A, leadership added more color. It has a “handful of weeks of Q4 visibility” and sees retail-level sales growth converging with Pepsi order growth. They seem to be wrapping up inventory cuts, and, based on current trends, that process is expected to be fully finished before the end of this year. For Q4, it expects the revenue impact to shrink from $124 million to anywhere between a “positive benefit” to a $15 million headwind. The impact should be gone thereafter. This was by far the most important and most positive piece of the call. So needed and so good to hear. It seems like Pepsi pulled forward a lot of the inventory shrinkage into this quarter, which bodes well for upcoming periods.

“I think we have good visibility, and we've been working closer and closer together [with Pepsi].”

CFO Jarrod Langhans

Margins also suffered from lower fixed cost leverage enjoyed from the Pepsi order cuts. Furthermore, margins were greatly hurt by the fully ramped and revamped incentive program with Pepsi. Celsius says this “further aligns interests,” which is a nice way of saying it’s trying to motivate Pepsi to work harder for them through more financial incentives. CELH was able to offset some of this impact with operating efficiencies and lower freight costs, but only a bit.

Sugar-free MULO+ W/C sales as a percent of overall sector volume continues to jump. It moved from 46.8% to 51.8% Y/Y, with Celsius taking meaningful market share. Specifically, its piece of this attractive segment rose from 16.6% to 22.3% Y/Y. The world is embracing a sugar-free future, with more than 25% of consumers calling “better for you” a difference maker in purchases; that bodes well for this firm. It’s hard to argue that the drinks are actually healthy. It’s easy to argue that they’re healthier than direct alternatives. That’s what it does.

All of this is why Celsius has powered 37% of category MULO+ W/C growth year-to-date. That’s undeniably impressive, but again less so when we look at more recent data. This quarter, the 37% category growth contribution fell to 19%. It’s still overindexing vs. its sector, but that outperformance is dwindling.

The graphic below shows you how healthy its portion of the beverage industry is vs. the overall sector:

In club, Costco sales growth remained strong at 15% Y/Y. For Sam’s Club and BJs, promotional timing and “innovation loading in the year-ago period” led to tough comps there. That fostered -4% Y/Y club channel revenue growth.

For Amazon, sales growth came in at a robust 21% Y/Y. It has 20.4% market share there vs. 19.7% year-to-date. Good result for this segment. Next, food service revenue represented 12.3% of sales to Pepsi this quarter vs. 12.1% Q/Q. Lodging and restaurant distribution points rose 46% Y/Y and 27% Y/Y respectively.

Sales across all 6 international expansion markets are “exceeding the company’s initial expectations for Q3. While many want faster scaling here, we must keep CELH’s approach to growth in mind. It always goes slowly, and starts with local gyms and fitness communities to build a base of brand awareness. That’s the process currently unfolding. When that critical mass of awareness is in place, it will accelerate marketing dollars knowing that those dollars will come with compelling returns. Simple enough, and international sales still did grow by 37% Y/Y.

I covered recent CELH M&A news here.

The company added this quarter that it’s open to more vertical integration, but not necessarily committed to it. It also has no interest in becoming a copacker and prefers its more asset-light, partner-based approach.

This quarter changes nothing about my views on the company. It’s encouraging to hear Pepsi commentary and means revenue growth will likely accelerate meaningfully going forward. At the same time, I’ve said that I want this to be a tiny position and I will not add until I see Q/Q market share trends look better. That did not happen this quarter. So? I’m not adding. The team is offering me enough optimism to stick around and see if they can weather the storm, but not nearly enough tangible progress for me to make this a more meaningful part of holdings. It’s a moonshot for now and I’m treating it as such. Broken record alert: consumer brands come and go constantly. I want more evidence of this brand being healthy and durable.

Mercado Libre is an e-commerce, logistics and payments giant in Latin America with a quickly broadening product offering.

Note that Meli made a series of changes to reporting disclosures starting last quarter. First, Mercado Pago Interest Income and Expense was moved from below the EBIT line to above it. For its shipping business (Mercado Envios) it changed its position from an agent to a principal. Previously, it netted shipping costs out of gross revenue. Now, it reports revenue as gross revenue and puts shipping expenses in cost of revenue line. Finally, it removed peer-to-peer volume from total payment volume (TPV). All of this means higher net revenue and margin dilution as a result. Q/Q and Y/Y margin comps reflect this.

Mercado Libre beat revenue estimates by 0.6%. Its 40.5% 2-yr revenue CAGR compares to 39.8% Q/Q & 38.8% two quarters ago. Fintech monthly active users (MAUs) rose by 33% Y/ while unique active buyers rose by 22% Y/Y. Accounting changes added $513 million in revenue during the quarter. Without this help, revenue rose by 28% Y/Y. This added 24 points to its 48% Y/Y commerce and fintech revenue growth results.

A ton of EBIT margin context is needed this quarter. The historical data uses pre-accounting change metrics. Without this change, EBIT margin would have fallen Y/Y by an additional 200 bps. Of the 770 bps of reported EBIT margin contraction, 740 bps was related to credit card growth, fulfillment center investments and higher stock comp due to its rising stock. It was also related to a 110 bps impact from a one-off refund to fintech users due to a policy change. None of these reasons should be overly concerning. I’ll explain why throughout this piece.

All in all, unique items led to 940 basis points of headwinds in the Y/Y EBIT margin comp.

MELI EPS is expected to grow by 62% this year and by 34% next year. The 2024 growth estimate is likely going to fall following the report.

Despite Mercado Libre being the e-commerce staple across Latin America and existing for more than two decades, the opportunity remains vast. That has a lot to do with Latin America lagging U.S. e-commerce penetration by “almost a decade.” It’s at about 15% per MELI and somewhere around there, according to other sources. Regardless of which data vendor we look at, the point remains. MELI still calls less than 10% of Latin Americans its customers. This, paired with continued strong value creation and execution, is why it continues to enjoy rapid GMV and TPV growth on a large base. Here’s an overview of FXN GMV growth and items sold growth by geography:

Market share rose across all geographies and MELI posted its 2nd straight quarterly record for post-pandemic unique buyer growth (+21% Y/Y). To keep optimizing the experience, deepening delight and bolstering engagement, the company added several new commerce menu items this quarter. It debuted appointment booking for auto part purchases and installation, virtual makeup try-on and new dynamic pricing tools for sellers, such as bulk discounts.

1st party commerce was another highlight for this quarterly report. 1st party GMV growth surpassed post-pandemic peaks. It’s taking advantage of its existing fulfillment capacity and marketplace traffic to drive more revenue (just like Amazon does). MELI’s push to plug 3rd party assortment gaps for certain products, through partners like Apple and also directly, is clearly working.

“1st party commerce is turning MELI into a destination for categories in which we have historically had less presence in.”

Shareholder Letter

As discussed last quarter, Meli split its Meli+ loyalty program into two tiers. Meli+ Essentials is $2/month and features more access to free shipping, loyalty rewards etc. Its Meli+ Total tier offers everything in Meli+ Essentials, as well as Disney+, Deezer and more streaming discounts. This is where MELI is successfully emulating the Amazon Prime playbook. It’s affordably passing on savings across a wide array of categories to build “the best subscription deal out there.” That’s what Amazon always says… that’s what MELI is doing in Latin America. It’s relying on incremental, unique delight to raise retention, improve revenue quality and offer this deep value with strong margins. Expect these programs to consistently become more valuable over time. I think this is a big reason why retention is improving across all segments.

Last month, the company added installment purchases and more cash back rewards for fintech users.

As briefly mentioned, there were a series of EBIT margin headwinds that led to sharp Y/Y de-levering. In addition to aforementioned accounting changes and refunds, building 6 distribution centers in Brazil and 1 in Mexico to meet demand hurt short term margin. This also let it raise self-fulfilled order penetration by 450 bps Y/Y, which it knows raises conversation rates, customer satisfaction and retention.

As MELI doesn’t offer forward guidance, these investment decisions often creep up on analysts and lead to temporarily lower capacity utilization, more deadweight loss, higher costs and sizable profit misses. That seems to be what happened this quarter. As long as the company maintains its pristine track record of strong returns on invested capital, this short term profit blip should be well worth it. It’s the right decision.

MELI has also fixed a previous risk for 1st party GMV proliferation through logistics improvements. Its 1st party margin used to be significantly lower than its 3rd party business. That gap has actually entirely closed as “1st party did not act as an income from operations drag.” That’s notable. Not even King Amazon gets as much margin from their 1st party business as their 3rd party segment.

Note that same & next day shipment penetration fell from 54% to 51% Y/Y and fell Q/Q as well. This is related to customers opting into cheaper, slower delivery and momentum for Meli Delivery Day. As a reminder, this product lets customers select a pre-set day of the week to get all of their packages.

Advertising represented 2.0% of GMV vs. 1.7% Y/Y. This is still being powered by product ad growth. Through Disney partnerships and off-platform arrangements, it should find growth across brand marketing campaigns for placements across channels like streaming.

Mercado Pago (payments platform) enjoyed robust 35% Y/Y MAU growth to reach 56 million total. Engagement trends are uniformly positive, as growth among its power users comfortably surpasses that of the overall base. This is resulting in it increasingly becoming the principal financial service provider for more customers. Sticky, sticky. Its newer credit cards and high yield savings accounts are the two large contributors to this thriving traction. The credit card portfolio overall rose 172% in size Y/Y to reach $2.3 billion. This leads us to EBIT (and overall) margin headwind #2. Meli is growing increasingly confident in its underwriting algorithms and how credit card performance is shaping up across vintages. Some of these vintages are ever turning net profitable ahead of schedule. So? MELI is leaning in, with its same long term value maximization over short term target chasing mindset. It invested an incremental $76 million in this product to support growth here, and to motivate more cross-selling, where the credit card is a key tool. This decision reduced EBIT margin by a full 340 bps.

This relates to growth for its Mercado Pago payments suite for merchants to offer to maximize customer convenience and conversion rates. Some of the volume happens within Meli’s platform and some occurs at merchant sites. MELI made a series of software-level investments in Brazil to better serve its merchants and build traction for its marketplace and checkout gateway API (so merchants can use it on their own sites). It added seamless product catalog building, tax receipts and inventory management. Much more is coming to augment the already strong 30% Y/Y FXN Acquiring TPV growth in Brazil.

Net Interest Margin after Losses (NIMAL) fell sharply to 24.2% vs. 31.1% Q/Q and 37.4% Y/Y. This is due to a series of factors, none of which are concerning. First, it’s expanding to longer-duration loans, as well as higher credit quality customers and businesses, where profit spreads are lower. Secondly, its credit mix shift is moving heavily towards credit cards and away from personal loans, which is also a profit spread headwind. Finally, as mentioned, it’s accelerating issuance here. Considering front-loaded CECL-style loss provisions, that hurts NIMAL in the short term. For more evidence of this benign entire fine, 15-90 day non-performing loan (NPL) rate improved to 7.8% vs. 8.2% Q/Q and 10.6% Y/Y. 90+ day NPL rate also improved Q/Q and Y/Y to 17.9%.

I’m sure many will read this number and assume the worst, but there is nothing at all alarming here. I’m not a shareholder; I have no inherent bias here; just my view.

As an aside, the ramp in Mexico is going much faster and with lower early losses compared to Brazil. This is thanks to the growing pains it learned from when first launching its credit card in Brazil.

I don’t think anything in here changes the long term fundamental investment case. They should be accelerating investments to capture new growth at strong returns. They should not be holding back on optimizing shareholder value creation by trying to meet quarterly profit numbers. This team’s mind is in the right place. Zooming out several years, I think that will serve them and shareholders extremely well as it has to date. On a quarterly basis, that may mean surprise investments that lead to misses here and there.

To traders, that matters a lot. To long-term investors, that often presents a compelling opportunity to generate more alpha. Growth is stellar, new bets are working, underwriting strength is apparent and this team is world-class. Fine quarter.