Welcome to the free Friday edition of The Diff! This newsletter goes out to 23,830 subscribers, up 304 since the last Friday free issue two weeks ago. In this issue:

• Magalu: Inside Brazil's Omnichannel Powerhouse

• The Privacy/Security/Centralization Axes

• Factors of Production

• VPNs and Streaming Price Discrimination

• The Belt and Road Alternative

• Privacy and Tradeoffs

Magalu: Inside Brazil’s Omnichannel Powerhouse

One of my favorite categories of company is the kind that has had a near-death experience—call it a 90% drawdown in market value—and since recovered. Levered cyclical companies do this sometimes (Sleep Number has done it twice, trading as a penny stock in 2001 and 2008 but recovering both times). Magazine Luiza, better known as Magalu, is a notable example. The Brazilian retailer went public in 2011, and by late 2015 it had lost 94% of its value. But since its IPO, the company has produced a total return of 4,200%, or 44% compounded. It sports an enterprise value of $27bn, and is still primarily focused on the Brazilian market. Magalu's growth was accelerating pre-Covid, with revenue up 28% in 2017, 36% in 2018, and 39% in 2019. And then, during the pandemic, growth exploded to nearly 60%.

Retail companies don't look like tech companies, but that's deceptive; the industry is one where technology adoption has been a competitive advantage even if the tech originates somewhere else. In the late 19th and early 20th centuries, electric lighting and cellophane wrapping made it much easier to sell goods based on visual appeal, while cash registers increased employee efficiency and reduced shrinkage. Later on, stores that adopted air conditioning and escalators early had a competitive advantage against those that didn't. Walmart was an early adopter of bar codes, which let it dramatically reduce inventory costs. Retail turns out to be a bad industry to work in if you like developing new technologies, but a very fun one if you like winning by deploying them: each of these waves of tech adoption led to a temporary advantage before becoming table stakes.

Magalu is also a tech adopter; it created one of the first digital celebrities, Lu, who now has 5.5m followers on Instagram and 14.5m on Facebook. The company is also testing new kinds of shopping, like creating streaming playlists that let listeners buy the instruments used to make the songs. Earlier, the company was using digital kiosks for in-store orders in the early 90s.

The technology front-end is the most visible part of a retailer's tech stack, but as the historical examples above show, the backend is more important. Retail is ultimately a business driven by asset turnover and cost control: turnover is a way to amortize more gross profit over a given fixed cost base, and also a way to avoid obsolete inventory. Retailers' inventory strategy can be modeled as something akin to advantage gambling: making a series of bets with positive edge but uncertain outcomes, and knowing that the way to win is to scale into situations where the odds are unusually favorable. A multichannel retailer like Magalu, with physical stores, a website, and apps, has more ways to scale, but also more problems to coordinate. The tradeoff between these two risks was ambiguous in recent years, but once Covid hit, the companies that were used to selling across multiple channels had a vast head start. Magalu used this head start in two interesting ways: they expanded their product offering, growing grocery delivery to 40% of units sold online, and they launched Partner Magalu, which let other retailers access their infrastructure to take and fulfill online orders. I'm always a fan of tacking "as-a-Service" onto concepts when it's even tangentially relevant, and the company buys into this, too: they refer to the entire suite of offerings to third-party merchants as Magalu-as-a-Service.

The grocery category is a particularly interesting one: it's hard to get packing and fulfillment costs low, because so much of the inventory is delicate and prone to spoiling. On the other hand, it's a high-frequency purchase, and a recession-resistant one. Magalu started out focusing on hard goods like electronics (today, the two categories their site pitches first are furniture and iPhones), which is a cyclical category. And for a seller in Brazil, it's important to worry about cycles. Brazil has had two deep recessions in the last ten years, including both the Covid drop and an 8% decrease in real GDP from 2014 through 2016. For a company that sells imported goods, that's doubly dangerous, since recessions hit Brazil's currency hard and made imported goods like smartphones more expensive.

So it makes sense for the company to diversify within retail. But it also makes sense for them to diversify vertically. One thing Amazon demonstrated was that the technology deployment dependency that characterizes retail can mean that a sufficiently effective retailer is a good seller of other services retailers want. In their annual report, the company notes that there's a vast addressable market in retailers that could use some or all of their logistics and e-commerce solutions, and an opportunity to do more in payments. (This, too, has been done before: credit cards are a very challenging two-sided network to crack; Visa and Amex's card businesses date back to 1958, and Mastercard's predecessor was founded in 1966. Discover is a comparative newcomer, launching in 1985, but it was launched by Sears, so it kickstarted its network effect by being the preferred card of what was then America's largest retailer.) They're also working on transitioning their app to being a super app, by building in bill paying, restaurant reservations and ordering, and other categories in addition to e-commerce. Once an app has good distribution, decent selection, and in-house payment options, the optimal strategy seems to be making it an app for everything.

Physical locations were a liability at some stages of the pandemic, when stores were either shut completely or people were voluntarily avoiding them. But in the adjustment period, they've turned into an asset: since Magalu's app is popular, and lists third-party goods, and since its stores are ubiquitous, it's been able to use the stores for in-store pickup for third-party goods. So Magalu is using an app that sells everything to accelerate a model that sells everywhere (or at least everywhere in Brazil).

Building a business that has multiple growth levers, and that can thrive in a return-to-brick-and-mortar environment or another Covid variant, a company that can sell phones and appliances or that can sell services to other companies that sell different products entirely, that sells staple products as well as cyclical goods—all of this sounds like a hedge. And Magalu is hedged, in a way. But running a complex model like this is not just a way to avoid worrying about uncontrollable events in the future; it's a way to direct the more controllable ones. A traditional retailer is vulnerable to economic cycles and execution risk, but a retailer that reimagines itself as a super app with a dense distribution network accepts higher execution risk in exchange for the chance to determine its own destiny.

Share

A Word From Our Sponsors

Secureframe helps hundreds of companies get enterprise ready by streamlining SOC 2 compliance. Secureframe allows companies to get compliant within weeks, rather than months and monitors 40+ services, including AWS, GCP, and Azure. Secureframe continuously collects audit evidence, runs security awareness training, manages vendors, monitors infrastructure, and more, all automatically.

Our team of compliance experts and auditors are happy to help answer any questions and give you an overview of SOC 2, even if you don't need it today!

Our customers save an average of 50% on their audit costs and hundreds of hours of their time. Schedule a demo and learn how!

Schedule a Demo

Elsewhere

The Privacy/Security/Centralization Axes

Two recent tech stories show an interesting set of privacy tradeoffs: Google has been banning location-tracking SDKs for privacy reasons, which has led at least one location-tracking company to switch to helping developers gather the data themselves. And Amazon is planning to monitor some workers' keystrokes and mouse movements to stop them from accidentally or deliberately accessing customer data they shouldn't have. What both of these illustrate is that centralized companies with recognizable brand names are much better equipped to keep data safe, and have a stronger incentive to do so. A headline that starts as "X-Mode tracks Android users' locations and sells data," is probably going to be rewritten to a much more clickable one that puts "Android" first. This doesn't mean that centralized companies are strictly better than the alternatives—perhaps no one should be gathering that much information—but it suggests that as privacy rules get more restrictive, bigger companies will have an easier time navigating them.

Factors of Production

The Economist draws attention to an obscure change in a Chinese Communist Party policy document ($): last April, "data were named as a “factor of production” alongside capital, labour, land and technology." One of the useful things about tracking China is that the CCP is a huge organization with 95 million members. There can be secret plans and secret policies, but it's hard to keep them too quiet if so many millions of people will be involved in implementing them.

Data sort of makes sense as a factor of production, although it's extremely non-fungible; Google's search data is worth a lot to Bing, but a lot less to Netflix and almost nothing to the average person. Encapsulating the data in an operating business that's designed around using it is a very complicated version of the same sort of convenience embodied in oil futures, which are much easier to handle than a few swimming pools full of West Texas Intermediate Crude. While worrying about companies handling data is a good pretext for regulating them, data economics are tricky, and the most reasonable way they map to other factors of production is that the state has an interest in controlling who benefits from data and how.

VPNs and Streaming Price Discrimination

VPNs have been a way to stream Netflix shows outside one's home country for years; since the catalogue (and pricing) vary by location, it's a way to choose from 190 bundles instead of just one. Netflix has been playing a very slow cat-and-mouse game with VPN providers, because allowing them is a form of price-discrimination—some of these users wouldn't pay full price, but they'll pay something. But recently, Netflix seems to have determined that VPNs, password sharing, and other revenue leaks are worth plugging, so it's started blocking them. But some VPNs use a different approach, which I wrote about briefly here ($): they use browser extensions to route traffic through random computers owned by real people instead of VPNs themselves, so it's hard to distinguish from regular traffic. It's not, as it turns out, indistinguishable, and Netflix is blocking some of that traffic, too. This is leading to collateral damage, at least according to the VPN companies, but it makes sense for Netflix's long-term plans: since evading a site's rules at scale is detectable, at least in a statistical sense, they can decide how much VPN use they're willing to tolerate, and block most of it beyond that.

The Belt and Road Alternative

Japan is a country that's had a high savings rate for an exceptionally long time, has low returns on domestic fixed-income assets, and a stock market that has historically not done a great job of returning profits to shareholders ($). As a result, Japan does a fair amount of outbound investment. Some of this is a complement to the country's existing economic strengths: as Japanese companies got priced out of labor-intensive work, they moved the lower value-added parts of their supply chain to Taiwan and then China. It also functions as a way to counter China's Belt and Road Initiative ($, Economist) by promising capital that's cheap and doesn't have strings attached. When export-driven growth runs out of steam, it often leads to an oversized financial system with a lot of capital that can't find good returns, but in this case the country's accumulated savings can help accomplish its strategic goals.

Privacy and Tradeoffs

Eric Seufert of Mobile Dev Memo has very interesting thoughts on Facebook's approach to privacy: FB can think in sophisticated terms around tradeoffs, but policy issues don't get salient because some efficient frontier has slightly shifted—they're more all-or-nothing. A cynical reading of the company's shift away from "move fast and break things" might be that it's not just caution that's a feature, but slower launch time. If the time it takes to implement one set of rules is longer than the life of privacy as a live political issue, the company can wait politics out. Facebook's stock is certainly priced as though either its model can survive a radically more restrictive approach to user data, or like it won't face one in the immediate future.