Drinking from the Firehose #127: 💰 Internet money. 💰


Facebook* announced the Libra project this week. Many people far more qualified than I have commented on its prospects and implications. For one, I like Eric Voorhees' take here. I normally don't feature topics that I think are well covered by others in this newsletter, but Libra just seems too darn important to miss.

Cryptocurrency skeptics and enthusiasts alike will argue fervently over weedsy topics like whether or not Libra is a "real" crypto, or whether the protocol behind Libra will ever become permissionless. A lot of this misses the forest for the trees.

The forest is that an application used daily by 1.5 billion people will now have crypto built in as a native feature. Since inception, all cypto projects have suffered from a tough "chicken or the egg" problem. Libra could solve that problem and unlock the first consumer crypto use case, inside of Facebook. With Libra in their wallets, those Facebook users will no doubt look for other ways to use their Libra. They will turn to other partners in the Libra network (initially including Farfetch, Spotify, Ebay). As such, Facebook's ambition is to bootstrap an entire payments network with an open currency. Facebook benefits because those users will be more likely to use Facebook's core services, which are designed natively to support Libra.

Along those lines, Justine Humenansky noted in a blog post that Libra brings Facebook closer to its endgame for WhatsApp, for which it paid $19B in 2014 and has yet to directly monetize. She compares the potential to Tencent's WeChat service, which "integrates payments with messaging in a way that allows the near entirety of one's online behavior to occur within one app's ecosystem, resulting in habitual usage and dependence." Moreover, Facebook likely knows that advertising in WhatsApp is a near dead end, thanks to the privacy implications of serving ads based on the text of messages between private parties. A payment platform on top of Libra is a much more straightforward, and less controversial, way for Facebook to create a palatable business model for WhatsApp.

I'd encourage everyone to read the Libra white paper below. I personally think it's a big deal and am excited to see where the project goes from here.


Welcome to the Official Libra Whitepaper.

Now is the time to create a new kind of digital currency built on the foundation of blockchain technology. The mission for Libra is a simple global currency and financial infrastructure that empowers billions of people. Libra is made up of three parts that will work together to create a more inclusive financial system: (1) It is built on a secure, scalable, and reliable blockchain; (2) It is backed by a reserve of assets designed to give it intrinsic value; (3) It is governed by the independent Libra Association tasked with evolving the ecosystem.



Fancy pants.

Farfetch* has one of the strongest moats against Amazon of any commerce company due to the nature of the luxury goods industry. Amazon is ubiquitous, and ubiquity is antithetical to luxury. Luxury brands that venture online do so carefully, and in such a way that preserves an air of scarcity around their products.

Farfetch found a back door to bringing luxury brands online: boutiques. By bringing luxury boutiques online, it effectively brought those brands online too. The brands themselves have struggled to aggregate enough consumers to compete directly. With $1.5B of trailing 12-month GMV, Farfetch is likely the largest single destination for luxury online.

The company's CEO recently gave an interview with JP Morgan in which he revealed more on how Farfetch works with boutiques. He revealed that Farfetch represents 30-40% of its boutique parters' sales at an average take rate of 30%. Because Farfetch is such a significant portion of its partners' business, boutiques rarely, if ever, churn. He positions the 30% take rate as small relative to the 55% retail margins that are typical in luxury. Farfetch's partners still earn 25% after its rake, which is incremental profit to them.

Are cities overrated? (podcast)

Conventional wisdom is that, if you want a high paying job and a better standard of living, you should move to a major city. The study featured on this episode of Planet Money questions that logic for the non-college educated population.

Decades ago, you could truly move to a big city and "work your way up from the mailroom." The study's authors posit that attractive top level jobs were accessible through on the job training. Those jobs today, however, seem to be less accessible to the same people. In cities like NYC and SF, the "professional" economy (e.g. finance, tech) rarely hires from the pool of service economy workers who support it. The result is that cost of living tends to be set by the more wealthy professional class, but the wages of non-college educated service workers may no longer rise over time to meet that cost. Without a path to improving income, those workers are less incentivized to move to major cities and may do at a lower rate in the future.


A brief history of Spotify (video).

Today Spotify* has nearly 200M MAUs and 100M paid subscribers. I find it really instructive to look at such iconic companies in the early days and analyze the decisions it took to survive and, eventually, thrive. These candid stories are rare because most executives like to rewrite history and make success look inevitable.

That's why I was thrilled to see this video of Chief R&D Officer Gustav Söderström's recent talk at MIT on the history of the Spotify.

At launch in 2008, the magic of Spotify was streaming audio from the cloud in under 250 milliseconds. This only worked because Sweden, its country of origin, was one of the first to roll out high bandwidth connections to its population. Growth was rapid. Within two months of launch, Spotify had 200K MAUs. It hit 1M MAUs within 6 months and 3.2M MAUs within 12 months. At the time, Sweden had only 9 million residents. Excluding babies and the elderly, Spotify had penetrated half the country's addressable market in a year!

Its first subscription product allowed offline storage of files for $9.99/month. At the time, music streaming over 2G or 3G was so bad that this was a really important feature for power users. It resulted in outstanding conversion to paid upwards of 30% in 2010.

Spotify was almost killed by the labels in 2011, who insisted that the company gate usage for its paying subscribers. MAUs actually declined for a bit, until the decision was reversed. It almost died for the second time in 2013 when a decline in global PC users, who were increasingly "mobile first," started to reverse growth. The labels initially restricted Spotify from launching a free product to solve this problem. The company eventually convinced them to allow "shuffle play" for free in the U.S., which resuscitated growth.

The video has more interesting details. I am most impressed at how many times the company anticipated technology shifts (mobile, machine learning, smart speakers) far ahead of the curve.


Oh hi, OYO!

OYO* has a unique business model that has enabled it to grow into the leading hospitality brand in its native India. It's now applying that model in the U.S. The WSJ wrote the first local piece of media covering the OYO's journey, featuring an interview with my partner and investor Bejul Somaia.


Stairway to Neven.

The rate of progress in quantum computing may be the first significant, observable phenomenon to progress at double exponential growth. One factor comes from the inherent nature of quantum computation, which embeds an exponentially growing number of Qbits compared to classical bits. The other comes from traditional semiconductor scaling laws. Hartmut Neven of Google was the first to make this observation, so many are calling it "Neven's Law" in parallel to the famous "Moore's law," which dictates progress in traditional processors.


Sleep no more.

This "3 Month Old's Guide to Sleep Training Your Parents" from McSweeney's hits a bit close to home!

"Go for the jugular. Make your parents recognize the futility of trying to get sleep. Cry in their arms for two hours, then fall asleep in your bassinet a minute before the Amazon delivery man knocks on the door, sending the dog into a frenzy. Fall deeply into REM, lulling them into a false sense of security, waking up when they are mid-shower, or cooking risotto, or disputing a charge with the credit card company. The more inconvenient, the better. Wail inconsolably until you can see the will leave their bodies. If neither of them is crying, cry at a higher pitch. Once they are in tears, let them cry it out. It will be hard at first, but in the long run, it will teach them to self-soothe."