Firehose #195: 🍲 The kitchen is open. 🍲

Announcing Lightspeed’s Series C investment in All Day Kitchens — the leading distributed platform for restaurant infrastructure.

Remember me?

A few months have passed since my last newsletter. Every now and then, I get into a pattern at work that so consumes my schedule that I don’t have time to sit down and write this newsletter each week.

I’m definitely in one of those patterns now. I can’t recall a time when I was busier. That said, I did announce a new investment this week in All Day Kitchens and wanted to share some of the thought behind it with all of you. I’d welcome your inquiries and feedback.

I hope to get back into more of a regular writing cadence before the end of 2021. Until then, Firehose is going to lean heavily on the “ish” in “weely(ish).”

One Big Thought

Local restaurants form the character of the communities they serve. Because food is the only we retail product we ingest, restaurants are deeply personal establishments. The late, great chef, author, and food culture documentarian Anthony Bourdain spoke eloquently on this topic:

Food is everything we are. It’s an extension of nationalist feeling, ethnic feeling, your personal history, your province, your region, your tribe, your grandma. It’s inseparable from those from the get-go. Even before we get into food professionals or food bloggers or food nerds, you’re already talking about something that people identify very closely with their identities.

A dark cloud and a silver lining.

The COVID-19 pandemic has wreaked havoc on public health and the global economy. Less discussed is pain it has inflicted on the social fabric of communities, of which restaurants are a critical part. When restaurants shut down, communities suffer too. From the beginning of the pandemic through the end of 2020, 17% of all US restaurants (over 100,000 establishments) closed their doors permanently. 16% of those restaurants had been open for more than 30 years. Local heroes like Pok Pok of Portland, Mission Chinese of NYC, Cliff House in SF, and City Tavern in Philadelphia were among the list of casualties.

The silver lining during these dark times for restaurants has been a booming e-commerce business. Online delivery had admirably grown to 8% of US restaurant sales in 2019, but COVID-19 pulled its market penetration forward by 2–3 years, according to Morgan Stanley. We now expect 20–25% of the of the US restaurant industry ($900 billion at its peak) to be fulfilled through delivery by 2025, with 53% of Americans now saying that purchasing take-out or delivery is “essential to the way I live.”

The demand-side of the food ordering equation would not be possible without delivery networks like Doordash (DASH, $74B), Uber Eats/Postmates (UBER, $89B), and more. Self-ordering is also possible thanks to Square (SQ, $121B), Toast (TOST, $28B), and other vendors who allow restaurants to manage those orders themselves. Middleware vendors like Olo (OLO, $5B) allow restaurants to connect those solutions into other systems of record, creating an omni-channel experience for restaurants.

Yet, despite this new source of e-commerce demand, restaurants have struggled to keep up. The dominant reason is that restaurant demand is inherently “peaky.” It’s even more peaky now that delivery and pick up are turning restaurants into omni-channel retailers. Restaurants have fixed labor costs that do not scale up or down with volatile demand throughout the day and week. Kitchens also have operating constraints defined by physical space, in addition to staff levels. While some QSR chains like Chipotle and Wendy’s have built out their own large commissary kitchens to serve the mass market, the roughly half of the US restaurant market that is independently owned is not operationally set up to scale like internet businesses.

A scalable back-end for independent restaurants.

Ken Chong and Matt Sawchuk saw these problems first-hand during their time at Uber, where they helped grow Uber Eats’ GMV from $0 to $6B in GMV run-rate in only a few years. They understood that the critical bottleneck to the growth of delivery was the kitchen itself.

The duo left Uber to start All Day Kitchens to address this problem. Today, we’re announcing that Lightspeed is leading the company’s $65M Series C financing. I delighted to join the board on the firm’s behalf.

Ken and Matt quickly noted a few problems with the conventional “ghost kitchen” model. First, traditional warehouse kitchens require significant scale in real estate, so the cost per square foot needs to be much lower than a typical restaurant. Those spaces don’t typically exist near where people live. So delivery orders fulfilled by warehouse kitchens in cheap, industrial zones often take significant time to reach the end-customer — resulting in a poor customer experience. Second, restaurants do not want to manage additional staff in a warehouse kitchen in order to scale. They have enough trouble hiring line chefs at competitive rates and can only manage so much labor complexity in a small business. Third, each restaurant manages its own menu and assortment in a traditional warehouse model, working with a restricted menu to streamline operations. It therefore lacks the ability to cross-sell into a longer tail of products, and participating in multi-restaurant baskets is impossible.

All Day Kitchens was designed with a novel operational model in mind. Instead of large centralized warehouse kitchens, it operates a dense network of distributed small kitchens that sit in neighborhoods close to the customer.

Restaurants prepare food in their own kitchens during down time and ship product “nearly finished” into the All Day Kitchens network. Each All Day Kitchens location (today in SF, Chicago, LA, and Dallas) serves food from dozens of these “local heroes,” and consumers can order through their favorite delivery app like DASH or UBER, or directly on Users can add their favorite CPG products to orders, and even place multi-restaurant orders in the same cart. Ever feel like a bit of curry or dumplings with your burger, maybe with a bit of ice cream to finish? All Day Kitchens has got you covered!

All Day Kitchens has performed remarkably well, growing 4x YoY with industry leading profit margins and service levels. The average All Day Kitchens order arrives in 20 min — less than half of typical delivery times. Its newer locations are performing even better than its older ones, a sign of network effects from density of locations and restaurant partners, as well as growing operational sophistication. The company has built significant tech to operate locations remotely, an essential element for fast and profitable new market entry.

All Day Kitchens marks Lightspeed’s 12th global investment at the intersection of food and commerce across the US, Europe, India, China, Southeast Asia, and LATAM:

We are thrilled to partner with Ken and Matt on their journey to help the best local restaurants operate at internet scale and delight their customers with tasty, satisfying cuisine. We are grateful to work alongside friends and co-investors, Andrew Chen at Andreessen Horowitz, Keith Rabois at Founders Fund, Adeyemi Ajao at Base10, and Jeremy Kranz at GIC, all of whom are investing in this round.

All Day Kitchens is hiring aggressively to hit some ambitious goals. If you’re excited about using technology to scale restaurants and want to get on a rocket ship, check out the job listings page:


Tweet of the Week

Enjoyed this newsletter?

Getting Drinking from the Firehose in your inbox via Substack is easy. Click below to subscribe:

Have some thoughts? Leave me a comment:

Leave a comment

Or share this post on social media to get the word out:


Disclaimer: * indicates a Lightspeed portfolio company, or other company in which I have economic interest. I also have economic interest in a number of public equities, including, but not limited to AAPL, ABNB, ADBE, AMT, AMZN, BLK, CCI, CRWD, FB, GOOG/L, HD, MA, MCD, MELI, MSFT, NEE, NET, NFLX, NOW, NVDA, PYPL, SE, SHOP, SNAP, SNOW, SPOT, SQ, TMO, TWLO, TXG, V, and VEEV. (Last updated: 10/3/2021)

Firehose #194: ⚖️ Operating leverage. ⚖️

Why product and services marketplaces earn profits differently at scale.

One Big Thought

I’ve been spending time lately thinking about how operating leverage functions differently in product- vs. services-oriented marketplaces.

Operating leverage measures the degree to which a company can increase its profits by growing revenue. Companies segment their costs into both variable and fixed costs. The former grow proportionally with revenue, while the latter do not — at least in the short term. The split between variable and fixed costs in a business determines its ability to achieve operating leverage.

For example, consider a SaaS company that puts fixed R&D dollars into a new version of its product. Whether it ships that product to 10 or 1,000 current users, the marginal cost of doing so is small relative to the up-front investment in R&D. While the company certainly needs to spend some amount of R&D to maintain its code base, a big chunk of that initial R&D should be considered fixed costs. The SaaS company therefore has strong operating leverage on its R&D spend.

To the contrary, consider a consulting firm with hundreds of employees billing their clients on an hourly basis. If the firm wants to onboard 1 or 100 new clients, it must hire a proportional number of new consultants. It can make those consultants marginally more efficient by standardizing Powerpoint templates, or building an internal knowledge base; but, the cost of these new consultants is almost entirely variable, and therefore the firm has low operating leverage in its core service.

Why does operating leverage matter? Because high leverage businesses show faster operating profit growth per dollar of additional revenue. Operating leverage is akin to the mechanical advantage you get from a physical lever. In a high leverage business, you push down with a little more revenue and you get a lot of profit on the other side:

Operating leverage can be applied to various line items on the P&L. Typically most of the costs included in COGS are variable (although not always), and most of the costs “below the line” in OpEx are fixed (although, again, not always). In the simplified case that all COGS are variable and all OpEx are fixed, then gross profit margin is a good proxy for operating leverage. Software businesses, which have high leverage, also have high gross profit margins, and the opposite is true for consulting businesses.

Where this gets interesting is in complex “low margin” businesses like marketplaces. By definition, the vast majority of marketplace GMV goes to participants in the marketplace. What is left behind is there to cover the remaining variable and fixed costs. These marketplaces can generate operating leverage from several different sources, with a significant difference between product- and services-oriented marketplaces.

Product-oriented marketplaces (e.g. EBAY, TDUP*, Faire*) primarily match supply and demand in goods and then enable the payment for and fulfillment of those goods. The way these marketplaces achieve leverage is by driving the cost of the transaction down. For example, TDUP lists a line item in its OpEx called “operations, product, and technology,” as shown below from the company’s S-1 filing:

This cost line has declined from 52% of revenue in 2018 to 50% in 2019. It went back up in 2020 to 54%, but that’s likely related to COVID complicating operations. Importantly, TDUP also breaks out its non-GAAP contribution profit per order, which increased from 23% of revenue in 2018 to 36% in 2019 (and back down to 27% in 2020, again likely due to COVID impact). In addition to payment processing costs, contribution profit per order deducts so-called “distribution center operating expenses” from gross profit:

“[Distribution center operating expenses] includes inbound shipping, inbound labor, distribution center fixed costs and management labor [emphasis added], excluding stock based compensation expense, which are included within our operations, product and technology expenses.”

We can see the overall calculation here:

So, contribution profit per order breaks out fixed costs of running distribution centers. As TDUP scales more volume in its distribution centers, contribution profit should go up disproportionately. Moreover, its inbound labor costs benefit from fixed investments in automation. While those investments are likely treated as CapEx, and are therefore not included in the contribution profit calculation, the benefits of those investments should improve the inbound labor costs included in contribution profit:

We believe that average contribution profit per order, when taken collectively with our GAAP results, including gross profit per order, may be helpful to investors in understanding our order economics. We believe that if we are successful in scaling and automating our platform pursuant to our strategy, our results will show (i) an increasing average contribution profit per order over time and (ii) a growth rate in average contribution profit per order that exceeds the growth rate of average gross profit per order due to our ability to reduce inbound processing costs [emphasis added]. Such results would likely mean that our average order economics are becoming increasingly attractive and that our investments in technology and automation in our distribution centers are having a positive impact on our average order economics.

So, this line item is crucial to understanding operating leverage at TDUP. The company believes that contribution growth should outpace gross profit growth because the former includes the benefits of operating leverage, which is exactly what the “lever” analogy would imply. Indeed, the former grew 83% from 2018 to 2019, vs. 44% over the same period for the latter.

The nature of operating leverage in a business like TDUP is that it comes from a lower cost structure. Investing in automation, for example, should enable inbound processing labor to be more efficient and therefore amortize more volume over the fixed cost of distribution center employees. TDUP can then choose whether to allow those gains in profitability to flow to the bottom line, or to reinvest them, for example, in better payouts for sellers, or lower unit prices for buyers. Regardless, the source of the leverage comes from lowering the unit cost at the contribution level.

Services-oriented marketplaces (e.g. FVRR, UPWK) primarily match supply and demand in services provided by humans. Because humans in a given field and market require a certain pay rate, less flexibility exists with respect to lowering the cost of a transaction (other than typical geo-based labor arbitrage). Instead, these marketplaces put an emphasis on high quality matches and lower cognitive overhead of running a search process for labor.

For example, in its F-1 filing, FVRR discusses its innovative approach to productizing the “Gig”:

At the foundation of our platform lies an expansive catalog with over 200 categories of productized service listings, which we coined as Gigs. Each Gig has a clearly defined scope, duration and price, along with buyer-generated reviews. Using either our search or navigation tools, buyers can easily find and purchase productized services, such as logo design, video creation and editing, website development and blog writing, with prices ranging from $5 to thousands of dollars. We call this the Service-as-a-Product ("SaaP") model. Our approach fundamentally transforms the traditional freelancer staffing model into an e-commerce-like experience. Since inception, we have facilitated over 50 million transactions between over 5.5 million buyers and more than 830,000 sellers on our platform.

Still, FVRR’s take rate in these transactions has remained around 25% for several years. Its gross margins are some of the highest of any marketplace, at over 80% today. The main way in which FVRR achieves operating leverage is not by lowering costs, but by increasing the spend per customer meaningfully over time. At the time of its IPO, FVRR had more than doubled that figure over a 6 year period:

Increasing spend per buyer essentially provides operating leverage against a fixed marketing expense. The shape of the curve means that older cohorts contribute more and more to FVRR’s overall quarterly revenue, without significant additional marketing support. The below demonstrates this methodology on a cohorted basis:

One other method to increase operating leverage in a services marketplace is to invest in labor differentiation and education. I’m not aware of a public company that does this yet, but expect to see more in the coming years.** Services marketplaces will eventually provide education and additional resources to help their workers “level up” to higher pay rates — perhaps in exchange for some sort of temporal exclusivity. The trade school of the future, in fact, may not be a school at all. It could be a labor marketplace with embedded education. Because the marketplace in question earns a fixed % take rate of billings, the incentives of all parties would be aligned.

In summary, operating leverage works differently for product- vs. services-oriented marketplaces:

  • A product-oriented marketplace should focus on driving more volume to cover fixed costs at scale, lowering overall unit costs.

  • A services-oriented marketplace should focus on both driving spend per customer up in a given customer cohort and investing in supply differentiation and capabilities.

  • Some overlap exists in these methodologies, but firms will get the most “bang for their buck” by prioritizing the top source of operating leverage in their respective categories.

Thanks to Lenny Rachitsky for reading a draft of this post.

** If you hear of any good examples, please let me know!


Tweet of the Week

Enjoyed this newsletter?

Getting Drinking from the Firehose in your inbox via Substack is easy. Click below to subscribe:

Have some thoughts? Leave me a comment:

Leave a comment

Or share this post on social media to get the word out:


Disclaimer: * indicates a Lightspeed portfolio company, or other company in which I have economic interest. I also have economic interest in AAPL, ABNB, ADBE, ADSK, AMT, AMZN, BABA, BRK, BLK, CCI, COUP, CRM, CRWD, GOOG/GOOGL, FB, HD, LMT, MA, MCD, MELI, MSFT, NFLX, NSRGY, NEE, NET, NFLX, NOW, NVDA, PINS, PYPL, SE, SHOP, SNAP, SPOT, SQ, TMO, TWLO, VEEV, and V.

Calling in sick. Sorry!

I'll be back in a couple weeks.

The last year or so have spared me from all manner of the normal colds/flus that afflict a household with two preschool age children. However, this past week our entire household got our first cold since COVID started. The kids are fine now, but this weekend I was down for the count. Father’s Day basically consisted of hanging out in bed listening to podcasts — not so bad actually.

As such, I don’t have a newsletter for you today. I also plan to skip next week because I have some unrelated family obligations. So I’ll be back in your inboxes in a couple weeks!

Firehose #193: 🦸‍♂️ Mega Marketplaces 🦸‍♂️

Why some marketplaces can drive their flywheels faster.

One Big Thought

Marketplace IPOs are commonplace today, but just 5 years ago they were a rare commodity. To quote myself in Firehose #187:

Of the publicly traded, independent US marketplaces that exist today, only 6 were public companies in April 2016 — EBAY ($30B), GRUB ($1.8B), GRPN ($1.8B), ETSY ($715M), CHGG ($321M) and OSTK ($234M). These companies were worth a combined $34B, 88% of which was EBAY — a company that was already 20 years old back then.

Since then, the public markets have embraced 16 new consumer marketplaces ranging from travel (ABNB*), logistics (DASH, UBER, LYFT), apparel (POSH, REAL, TDUP*, FTCH), real estate (OPEN), e-commerce (WISH), automotive (CVNA, VRM), events (EB), and services (ANGI, FVRR, UPWK). EBAY has shrunk from 88% of aggregate marketplace enterprise value to only 10% today. Aggregate enterprise value has exploded to $474B across these 22 listed companies. Of the $440B of gain in enterprise value over the last 5 years, only 15% came from the original 6 stocks. These 16 new marketplaces have been responsible for nearly all the incremental value creation.

With all these additional public companies, we can finally make some generalized observations about the marketplace business model. The most important characteristic of a marketplace is its flywheel. The flywheel dictates not only how fast a marketplace grows, but what constitutes its moat over time. Below are some of my favorite examples from DASH, UBER, Faire*, FTCH, and ABNB*.

Flywheels give marketplaces inherent momentum. Like a powerful locomotive, they are painful to get started, but difficult to slow down. The fact that Craigslist — a product that has made very few changes since the late 1990’s — still dominates certain markets in the US is evidence that liquidity, which governs the flywheel, is the most important component of a marketplace. Everything else is secondary.

Some flywheels spin faster than others. I spent time this weekend looking at the path to IPO for a number of publicly traded marketplaces. I found that the companies can be split neatly into two groups based on rate of growth since inception.

The grouping is tighter when we plot annual gross profit instead of revenue, since it puts businesses like OPEN (6% GM at IPO) on level footing with those like FVRR (79% GM at IPO). I’ll call the upper grouping in the gross profit chart the Mega Marketplaces (EBAY, DASH, UBER, LYFT, OPEN, WISH, ABNB*). They roughly track the ascendence of EBAY in blue. I’ll call the lower grouping Steady Marketplaces (GRUB, ETSY, CHGG, POSH, REAL, TDUP*, FTCH, VRM, EB, FVRR, CARG). They roughly track the path of GRUB in red.

The differences between these two groups are significant:

  • Time to IPO. The average Mega gets to IPO two years earlier (8 vs. 10 years after founding).

  • Gross margin at IPO. The average Mega has a significantly lower gross margin at IPO (49% vs. 61%). That’s because the Mega’s are mostly in “commodity” industries like food, transportation, and real estate.

  • Faster to $1B+ in gross profits. The average Mega achieves $1B in gross profits 8 years after its founding, at which point the average Steady has around $100M of gross profit. That order of magnitude difference is counterintuitive since the former actually has lower margins on a percentage basis than the latter.

The qualitative differences between Mega’s and Steady’s are a bit murkier. At the risk of seeming reductionist, it might all come down to a market size measured in trillions, not billions. UBER declared in its S-1 that its current serviceable addressable market (SAM) was $2.5 trillion:

ABNB had a similar feature in its S-1:

We have a substantial market opportunity in the growing travel market and experience economy. We estimate our serviceable addressable market (“SAM”) today to be $1.5 trillion, including $1.2 trillion for short-term stays and $239 billion for experiences. We estimate our total addressable market (“TAM”) to be $3.4 trillion, including $1.8 trillion for short-term stays, $210 billion for long-term stays, and $1.4 trillion for experiences.

OPEN positioned itself to address the 87% of US homes within the $100-750K buy box, which constitute a $1.3 trillion addressable market:

DASH cites third party reports in its S-1 that peg $1.5 trillion of food spend in the US, of which $600 billion is out-of-home:

We are the category leader in U.S. local food delivery logistics today and have an enormous market opportunity ahead of us in food alone. In 2019, Americans spent $1.5 trillion on food and beverages, of which $600.5 billion was spent on restaurants and other consumer foodservices.

While Steady Marketplaces must also address sizable markets, their respective paths to scale are likely slower because those markets are “only” measured in billions. FTCH, for example, is an admirable company with nearly $2 billion in net revenue, but it addresses 10-25% of a Mega’s addressable market. See below from its F-1 filing:

According to Bain, the global market for personal luxury goods was estimated to reach a record high of $307 billion in 2017, growing at a 6% CAGR since 2010, and in 2017, the personal luxury goods market experienced growth across all regions. Bain also states that online has become a larger percentage of the overall market, growing at a 27% CAGR since 2010.

Obviously, everyone wants to go after a big market. If you had to choose a trillion dollar market or a billion dollar market, why would you ever choose the latter? Well, smaller markets often have less competition and, therefore, higher margins. Those margins can present an attractive entry point for an online marketplace. It’s much more common, and probably more advisable, to attack a fat margin pool if you see one. Yet, the biggest markets of all (food, transit, healthcare, energy, etc.) are very competitive and low margin. The data clearly shows that an effective flywheel in one of these markets produces the biggest outcomes fastest.

To paraphrase a famous Jeff Bezos quote, two kind of companies exist — “those who work hard to try to charge more, and those who work to charge less.”

The Mega’s are the latter.


Tweet of the Week

Enjoyed this newsletter?

Getting Drinking from the Firehose in your inbox via Substack is easy. Click below to subscribe:

Have some thoughts? Leave me a comment:

Leave a comment

Or share this post on social media to get the word out:


Disclaimer: * indicates a Lightspeed portfolio company, or other company in which I have economic interest. I also have economic interest in AAPL, ABNB, ADBE, ADSK, AMT, AMZN, BABA, BRK, BLK, CCI, COUP, CRM, CRWD, GOOG/GOOGL, FB, HD, LMT, MA, MCD, MELI, MSFT, NFLX, NSRGY, NEE, NET, NFLX, NOW, NVDA, PINS, PYPL, SE, SHOP, SNAP, SPOT, SQ, TDUP, TMO, TWLO, VEEV, and V.

Firehose #192: 🍽 Campo a la mesa. 🍽

Lightspeed invests in Frubana, doubling down on LATAM. Plus: pumping up PTON, anime going mainstream, the NFT bust, and more.

One Big Thought

Latin America (LATAM) is the fastest growing e-commerce market in the world. It has 640M residents with $5.8T of GDP (6–7% growth). GDP per capita is $9K, or roughly the same as China. The region has near ubiquitous mobile and internet penetration, yet it has a fraction of the tech investment dollars of comparable developing regions like Southeast Asia. One sign that capital is flowing into the region now is the meteoric rise of MercadoLibre (MELI)*, which has increased its stock price by 4x in the last 3 years, buoyed by LATAM’s digital transformation during COVID.

Private capital flows have accelerated too. 2021 is tracking to an all-time high of ~$15B of venture capital in LATAM companies, a 3x increase since 2018. The secret is getting out that LATAM will be the next global region for breakout tech startup growth.

While the macro factors are attractive, one of the more practical aspects that makes me bullish on LATAM is a new generation of experienced talent emerging to start their own companies. MercadoLibre (founded in 1999) is the poster child of the 1st generation. Companies like Nubank (founded in 2013) and Rappi (founded in 2015) exemplify the 2nd. This 3rd generation cut their teeth at these foundational startups and got an invaluable, real-world education in company building. Their market timing is even better than that of their predecessors.

In late 2019, I met a compelling founder from this 3rd LATAM generation. Fabián Gómez was one of Rappi’s earliest employees and top performers. He had recently founded Frubana — a wholesale marketplace for local restaurants. In our initial meeting, I was blown away by Fabián’s incredible passion, market knowledge, hustle, and attention to detail. I had some good context for his business since I led Lightspeed’s initial investment in Faire in 2018. My experience with Faire, as well as our portfolio companies Udaan (India) and Ula (Indonesia), informed my belief that B2B/wholesale would be the next wave of online marketplaces.

While we weren’t prepared to invest in Frubana at that time, my partner Mercedes Bent and I set out on a mission to map the LATAM ecosystem and build relationships with entrepreneurs and investors in the region. We met hundreds of companies in the following 18 months. In early 2021, we finally decided the timing was right to make our first few investments.

We happily announced our first LATAM investment in Stori a few months ago. This week, we announced that our second is Frubana!

Frubana* is a wholesale marketplace serving the restaurant industry in LATAM. 1.5M restaurants in LATAM spend over $100B procuring goods each year. The market is transacted today through a complex supply chain of middlemen in local stalls in town plazas. The below is a typical produce market in Mexico City, for example. Farmers and restaurant owners meet at these markets several times a week to exchange goods and money. To reach these markets, product passes through multiple layers of middlemen in a highly inefficient process with little transparency or consistency.

Frubana has vertically integrated the supply chain for produce, giving farmers more value for their products, reducing waste, and delivering transparent pricing to restaurants. Frubana purchases directly from farmers, aggregates produce in regional distribution centers, and fulfills deliveries using intelligent routing. While building a produce supply chain is no easy task, doing so has enabled Frubana to create a highly recurring purchase behavior with its customers, leading to best-in-class retention for a wholesale marketplace.

Frubana also sells proteins and packaged goods through a 3rd party marketplace where sellers compete for Frubana’s customers. While most customers start buying produce from Frubana, over time they grow their “share of wallet” by adding the other categories to their deliveries. Frubana has some of the best retention we’ve seen in the wholesale category. In cities like Bogotá, Mexico City, and São Paulo, tens of thousands of restaurants essentially run on Frubana.

Compared to in the US, the restaurant industry in LATAM is significantly more fragmented. 86% of restaurants are independently owned, similar to developing nations like China (81%) and India (81%). The US restaurant industry is only 47% independent, and as such large food distributors (Sysco, US Foods, Performance Food Group — in aggregate, $59B market cap) have grown to serve corporate restaurant chains. Because of their local economies of scale in distribution and ability to cross-sell, these distributors are incredibly hard to disintermediate in the US. However, in LATAM, no such oligopoly exists. By building a supply chain for produce first, Frubana has a unique opportunity to aggregate LATAM’s restaurant market.

COVID hit the restaurant industry hard in LATAM. Approximately 70% of establishments shut down or slowed operations. Despite these headwinds, Frubana grew active customers 6x and sales 3x in the last year — establishing its position as the “everything store” for the LATAM restaurant industry. This resilience in the face of adversity demonstrated how capable Fabián and his team truly are. With little digital penetration in local restaurants, we think Frubana will be well positioned to bring these businesses online, help facilitate a transition to electronic payments, provide working capital, and in general drive efficiencies throughout the entire food ecosystem.

Frubana’s mission is to make food cheaper and more accessible in LATAM. It reduces food waste thanks to its innovative supply chain. Frubana only looses 1–2% of tonnage to waste compared with over 50% in the existing supply chain. That means more money in the pockets of farmers, more efficiencies for the restaurant owner, and cheaper end-product for consumers. Any great platform gives away more value than it captures, and that’s certainly the case for Frubana.

Lightspeed is delighted to partner with Fabián and the Frubana team to build a category defining company in LATAM. We’re excited to co-invest with friends Carlo Dapuzzo at Monashees, Hans Tung at GGV, Tiger, Softbank, and more.

If you’re interested in hearing more of my thoughts on Frubana’s market opportunity, and global wholesale marketplaces in general, check out this interview I did last week at a virtual customer event for Frubana. It’s even translated into Spanish!

And, if you’re interested in working for Frubana and reside in Bogotá, Mexico City, Buenos Aires, or São Paulo, check out their careers page.

Note: I originally published this post on the Lightspeed Medium here.


Tweet of the Week

Links I Enjoy


Pumping up Peloton.

Peloton was about as counterintuitive an early stage consumer investment as you can find. CEO and founder John Foley famously struggled to raise the first $400K angel round. A $2,000 exercise bike, with negative margins, in an industry built on breakage? Conventional wisdom said, “No way.”

But, Foley led Peloton through its IPO and built the most addictive exercise service ever to exist, with 90%+ annual retention of members. Its margins are world class on both hardware and subscription, and its market position continues to grow.

This week I read an insightful bull case on the company. It hits on some of the same factors that make Netflix such a successful business, namely that Foley regards the subscription price of $39/month as “sacrosanct” and that over time he wants to “make it irresponsible to not have a Peloton membership.”

The management has stated that the goal is to have 4 or 5 fitness platforms that are instructor-led, music, and socially driven. What initially started as a bike to democratize access to boutique fitness classes will transform into a multiplatform fitness ecosystem that will radically improve our approach towards fitness.

In addition, the post touches on another element of Peloton’s strategy, which the author calls “premiumization”.

Our day is composed of hundreds of experiences and if any given company is able to improve a relevant aspect of our life for a reasonable (but higher than other alternatives) price, we will eventually make the transition towards its products or services.

The author is making the argument that gyms used to be gritty and inconvenient (Gold’s Gym, etc). Then they got fancier, but still inconvenient (Crunch, Planet Fitness). Then they got social and fancy, and slightly less inconvenient (SoulCycle, CrossFit). Peloton is the inevitable end-state: a premium, social gym experience with all the convenience of working out at home.

This evolution of the fitness industry from gritty gyms to $2,000 connected home bikes reminded me of the premiumization coffee industry. My mother drank Folgers’ brand, freeze dried coffee every morning growing up (and so did her mother). Then Starbucks came along and turned coffee into an “experience” — charging multiples per cup of what you could make at home. Keurig made similar quality coffee even more convenient at home. Then came along even higher end home coffee machines like those made by Nespresso. Finally, we saw the rise of the “third wave” coffee companies like Blue Bottle that make a truly artisan product available commercially through all points of purchase. The price of coffee for a consumer can literally span an order of magnitude today. It turned out the market for coffee was much bigger than anyone expected as it became de-commoditized, or “premiumized.”


Anime is going mainstream.

One of the oddest trends in media during COVID has been the western rise of anime in popular culture. According to The Information, Netflix has seen viewership of anime double over the last year in the US. In Chile and Peru, anime shows are consistently in the top 10. Netflix apparently pays 50% more to anime studios compared to its usual deals, due to the demanding technical nature of developing these shows. The best IPs are also proven out by the comics (or “manga”), which means they are highly competitive and generate bidding wars. Netflix acquired the first season of Demon Slayer, for example, on a non-exclusive basis — highly unusual.

The subsequent Demon Slayer movie, released in late 2020, has now grossed over $500M globally, making it the highest grossing film of 2020! For those of you playing along at home, that worldwide gross is more than Christopher Nolan’s Tenet brought in at the box office. Furthermore, the movie starts chronologically exactly where Season 1 of the anime stopped, with no recap whatsoever. It makes no sense at all if you aren’t familiar with the series. And it’s in Japanese with subtitles! I gotta believe that everyone who saw that movie was already a severe Demon Slayer fan.

I fully expect we’re going to see more anime movies opening in wide release post-COVID and, given the prior anime TV series and manga popularity, we’ll see them beat out major live action releases in US theaters.


Boom and bust (and boom again?). →

NFTs are one of the most exciting trends in tech this year, but it’s hard not to notice the rapid decline in enthusiasm in nearly every NFT category in the last month. I point to the data in this post not as a nay-sayer, but as a conscientious observer who is interested in the underlying long-term trend.

One nevertheless bullish observation is that, compared to June 2020, active NFT wallets are still up YoY in most categories. In addition, the biggest category of decline is collectibles, which I’ve personally viewed as the most speculative category. The time it takes to establish an item traditionally as collectible is at least a generation (nostalgia needs time to set in), so a whole raft of creators claiming they’re creating new collectibles in 2021 feels premature to me. On the other hand, NFTs that are attached to experiences like games, meta-verses, or sports, have retained much better than collectibles.

My prediction is that we’re about to enter the “trough of sorrow” for NFTs, but that more productive use cases will rise from the ashes.


Twisted black holes.

Black holes are the mystery machines of physics. They sit at the edge of what we know and what we want to know to understand the cosmos. The scale and energy of these systems simply defies all intuition. This article is about a particular, supermassive black hole called M87 that its at the center of the Milky Way galaxy:

“The black hole in M87 is about the size of our solar system,” Issaoun said, yet it produces a 5,000-light-year-long current of white-hot plasma. That’s like the Statue of Liberty popping out of a marble. Some 3 trillion trillion trillion joules of energy flow up the jet each second — 500 trillion times more energy than the entire human population burns in a decade. “How could something so tiny be so powerful?”

In particular, the article discusses recent explanations of black hole jets and goes back to review old theories. The above video does a good job reviewing the basic concepts.


Here’s Johnny!

Like many “old millennials,” I grew up watching MTV. Jackass debuted in 2000 and ran for 3 seasons on the air. It redefined the age old art of the daredevil in an irreverent, vulgar modern context. The first few seasons of the show were addictive, and Knoxville, along with his peers like Bam Margera and Steve-O, became iconic symbols of the Gen X, IDGAF, skater punk generation.

This GQ profile of Johnny Knoxville is nostalgic for me, and I bet it will be for some of you. Crazy to see your childhood heroes get a full head of grey hair.

Enjoyed this newsletter?

Getting Drinking from the Firehose in your inbox via Substack is easy. Click below to subscribe:

Have some thoughts? Leave me a comment:

Leave a comment

Or share this post on social media to get the word out:


Disclaimer: * indicates a Lightspeed portfolio company, or other company in which I have economic interest. I also have economic interest in AAPL, ABNB, ADBE, ADSK, AMT, AMZN, BABA, BRK, BLK, CCI, COUP, CRM, CRWD, GOOG/GOOGL, FB, HD, LMT, MA, MCD, MELI, MSFT, NFLX, NSRGY, NEE, NET, NFLX, NOW, NVDA, PINS, PYPL, SE, SHOP, SNAP, SPOT, SQ, TMO, TWLO, VEEV, and V.

Loading more posts…