Weekly readings – 19th September 2020

What I wrote

I reviewed a few books such as: The Anatomy of The Swipe, Tape Sucks, The Motley Fool Investment Guide, 7 Powers: The Foundation of Business Strategy

I put down some thoughts on Apple Fitness+ and Apple One


A deep dive analysis into Snowflake

A study on the effect of Wikipedia on businesses

Our estimates show that adding about 2,000 characters (approximately two paragraphs) of text and one photo to a city’s Wikipedia page increased the number of nights spent in this city by about 9% during the tourist season compared to cities in the control group.6 The effect comes mostly from pages that were initially relatively incomplete. In particular, the treatment increases hotel stays by about 33% in cities which initially had very short pages in a particular language, while there was no effect on city-language combinations where the pages were well developed.


A review of Microsoft Duo by WSJ. It’s quite concerning that a $1,400 phone has a subpar camera and a buggy software

What I found interesting

A brief profile by BBC of Freiburg, a green and futuristic city in Germany

An interactive map of the Earth some 240 million years ago

A damning memo of a Facebook employee on how the leadership turned a blind eye on election manipulations. She wrote “I have blood on my hands”

The US is almost at the bottom among advanced countries when it comes to well-being of children

The Three Year Rule: How To Stay Motivated Working On A Long-Term Project

A WSJ profile on Trevor Noah and his journey from South Africa to America

According to Census, Asians had the highest median income in the US in 2019 and foreign born folks grew median income at a faster rate than native-borns

Weekly readings – 29th August 2020

What I wrote

A quick thought on Uber/Lyft potentially leaving California and Apple’s fight with Epic

I applauded the NBA for taking a stand

My breakdown of Snowflake’s S-1

I also wrote about Wix, a popular website creation tool, with more than 180 million users


A 3-part series on Microsoft’s acquisition of Forethought, the company that invented PowerPoint

A short story on Goodreads

In its lawsuit against the Trump Administration, TikTok reported 50 million daily active users in the US, 20.3 billion views on posts related to #BlackLivesMatter

This should sound ominous for Amazon’s rivals and the likes of UPS

Faster Than Fast: SMB Retailers Move to Shopify

A study on Gen Z’s attitude towards brands in 2020

Apple Pay can be a multi-billion dollar business

A great Twitter thread on the bull case fo Fiserv

How Roku Built Itself Into a Major Gatekeeper in Premium Streaming

A piece on Adobe’s bull case

In Defense of the IPO, and How to Improve It


India, Israel and Kyrgyzstan have the cheapest plan for every GB of mobile data in the world

A primer on Apple’s new security and privacy features in the upcoming iOS

Amazon Halo

Stuff that I found interesting

The inside story fo the $8 million heist from the Carnegie Library

New Swing State Georgia Could Decide Control of the Senate

How cute is this?

Get to know Snowflake

What is this company about?

Every business needs data to be competitive nowadays. The challenges related to data are daunting. First, data grows exponentially every day. Companies need to invest and maintain the right infrastructure to store a large amount of data. Second, the data infrastructure needs to support the business use cases. It must enable data staff to access data quickly and efficiently in order to build data-driven applications as well as carry out data analytics. As business leaders have to make quick decisions as a response to changing environments, it requires fast and accurate analytics.

For instance, I work at a bank and at the beginning of the pandemic, we needed to build monitoring dashboards quickly so that the management team could make quick decisions and have a finger on the pulse of the business. It was a challenge because our data warehouse’s structure made it so time-consuming to complete complex queries. And when multiple users tried to execute queries at the same time, the problem was exacerbated. I often had to work late at night to make sure my queries could finish faster. Our bank is just a regional one. I imagine that our data problems would be magnified if we had a bigger operation.

Snowflake is built to solve data problems. It offers cloud-based data storage and analytics services through which businesses can have a single truth of data, consistency, resiliency, flexibility and fast access to data. Snowflake products work on major public cloud providers such as AWS, Azure or GCP. On top of Snowflake, companies can build out a proper data warehouse, enable data analytics, power data applications and facilitate data exchange with other entities.

Figure 1 – Overview of Snowflake. Source: Snowflake S-1
Figure 2 – Use Cases from Snowflake. Source: Snowflake

As a Software-as-a-Service, Snowflake allows customers to pay only for the resources and services they use. Since Snowflake relies on public cloud providers, their unit price mirrors that of such providers as in that prices change depending on which region you are located. Snowflake’s customer base is impressive. They count as clients corporations such as Akamai, Capital One, Neiman Marcus, AXA, McKesson, Hubspot. In terms of technology partners, Snowflake’s website lists 17 major partners, including GCP, AWS, Azure, Salesforce and Looker.

How has the company been doing?

In S-1, Snowflake revealed its financials all the way back to the quarter ending in October 2018. The company hasn’t been operationally profitable as it recorded losses in every quarter. Revenue increased from $29 million in October 2018 to $133 million in July 2020. Operating loss stood at $78 million in July 2020. Before Jan 2020, Snowflake’s operating loss was even bigger than its revenue every quarter, but since then the loss has gone down while revenue has accelerated.

While the company hasn’t been profitable, the good news is that its gross and operating margin have been improving. If the company can continue to keep its gross margin and increase operating leverage (spend less on acquiring customers & be more efficient), it’s slated to be profitable in the near future. The sign is already there. 94% of the company’s revenue in 6 months ending July 2020 was from existing customers that expanded their usage. That implied a customer stickiness, satisfaction with the products and cost savings in acquiring new customers to increase the top line.

The past 12 months were positive for Snowflake. The company doubled its customer base from 1,547 in July 2019 to 3,117 in July 2020, as well as the number of customers whose trailing 12-month product revenue was greater than $1 million increased from 22 to 56 in the same time frame. Remaining Performance Obligation (RPO), which is a performance metric that includes unrecognized revenue in the near future, grew from $221 million in July 2019 to $688 million a year later. Net revenue retention rate has been always higher than 150%, with the latest quarter seeing 158%. Snowflake expects this rate to decrease in the near future.

Regarding platform usage and customer acquisition, Snowflake has been fairly impressive. Its platforms are used by 7 of the top Fortune 10 (4% of revenue in July 2020) and 146 of the top Fortune 500 (26% of total revenue). In the month of July 2020, the company reported an average of 507 million daily queries, compared to 254 million daily queries in July 2019. Net Promoter Score (NPS), a metric that indicates how willing a customer is to promote a brand, stands at a very good 71.

Capital One is an important client for Snowflake. The bank made up 17% ($16 mil) and 11% ($29 mil) of Snowflake’s annual revenue in fiscal year ending Jan 2019 and Jan 2020, respectively. While its share of revenue was less than 10% in the quarter ending July 2020, it still made up 22% of Snowflake’s account receivables ($33 mil). The trend lowers risks for Snowflake and investors as the company is now less reliant on this one particular customer.

What are the tailwinds behind Snowflake?

In its S-1, Snowflake reported a Total Addressable Market of $137 billion for the company as of 2020. Given the annual run rate of just $520 million in revenue, Snowflake has a lot of room to grow in the future. Furthermore, 12% of the company’s revenue came from customers outside the US. That indicates a big opportunity internationally for Snowflake. Since the company is built on top of public cloud providers whose infrastructure spans the globe, it already has the base infrastructure to expand internationally.

More importantly, the digital transformation that is underway in the corporation world is positive for Snowflake. As companies become more agile and digital, they need data to make informed decisions and be competitive. Hence, products and solutions that live in the cloud like Snowflake’s are well-positioned to capture this trend. From a personal perspective, my company is still hosting data on legacy infrastructures which present a bottleneck to our data analytics and operations. I can’t tell you how many times I had hours off as our data warehouse went offline. I can’t count how many hours I wasted because it took too long for complex queries to run. My developer colleagues reported happiness with the upcoming transition to Snowflake. Hence, this gives me a bit of confidence in the company as I know that my employer isn’t the only one running on legacy data infrastructure.


Among the risk factors listed in its S-1, there is one that I think stands out for Snowflake: its reliance on public clouds. Here is what it wrote:

We currently only offer our platform on the public clouds provided by AWS, Azure, and GCP, which are also some of our primary competitors. Currently, a substantial majority of our business is run on the AWS public cloud. There is risk that one or more of these public cloud providers could use their respective control of their public clouds to embed innovations or privileged interoperating capabilities in competing products, bundle competing products, provide us unfavorable pricing, leverage its public cloud customer relationships to exclude us from opportunities, and treat us and our customers differently with respect to terms and conditions or regulatory requirements than it would treat its similarly situated customers. Further, they have the resources to acquire or partner with existing and emerging providers of competing technology and thereby accelerate adoption of those competing technologies. All of the foregoing could make it difficult or impossible for us to provide products and services that compete favorably with those of the public cloud providers.

Source: Snowflake’s S-1

This is really a genuine concern. I don’t see Snowflake will build out its own underlying infrastructure in the near future. Mirroring the scale and sophistication of these public providers, especially if they want to expand overseas, will be expensive and resource-consuming. While such a reliance presents a risk, particularly when all AWS, Azure and GCP have competing products with Snowflake, the company also brings a lot of revenue to these cloud providers. So it creates an interesting dynamic in which I also suspect the Big Three will do anything to harm its startup customer.

In summary, my personal experience gears me towards investing in this company. I also observed some folks lauding the business on Twitter. The company itself has seen impressive growth and has a lot of room to grow as well as tailwinds. I personally look forward to the IPO debut of Snowflake.

CrowdStrike looks impressive in their IPO filing

Today, the security startup CrowdStrike filed to go public and the numbers look impressive in my opinion.

Company background

Founded in 2011, CrowdStrike is a cybersecurity startup that offers their services mainly on a subscription basis. The primary offerings include endpoint security, vulnerability management, threat intelligence and a PaaS solution for cybersecurity.


  • Subscription customer base grew from 1,242 at January 31, 2018, to 2,516 at January 31, 2019 – a 103% increase
  • Customers include 44 of the Fortune 100, 37 of the top 100 global companies, and nine of the top 20 major banks
  • Total revenue grew from $52.7 mil in 2017 to $119 in 2018 and $250 mil in 2019, an increase of 125% and 110% respectively
  • Subscription revenue grew from $38 mil in 2017, to $92.6 mil in 2018 and $219.4 mil in 2019, an increase of 144% and 137% respectively
  • ARR growth is impressive as shown below
  • Dollar-based net retention rate grew from 104% in Q4 2017 to 147% in Q4 2019
  • 23% of the company’s revenue came from customers outside of the US in 2019, up from 13% in 2017

Partner & Customers

  • CrowdStrike is deployed on AWS GovCloud after receiving FedRAMP recently
  • Dell & SecureWorks use CrowdStrike’s endpoint security solution
  • Customers include AWS, HSBC, ADP, Hyatt Hotels, The Pokemon Company

Listed Competitors

  • McAfee, Symantec, Palo Alto Networks, FireEye, Cylance and Carbon Black


  • The company grew the top line significantly while the operating loss had a much smaller increase
  • The increasingly profitable subscription that already has higher gross margin than professional services makes up a bigger piece in the revenue while expenses are better leveraged


The company claimed to have a TAM of $24.6 billion and $29.2 billion in 2019 and 2021 respectively. It is a huge market and as companies go digital and have increased exposure due to more endpoints, more data, more cloud environments and more applications, the cybersecurity need will be there. With that being said, it is also a competitive market. Not only are there quite tough competitors such as Palo Alto Networks, McAfee, Symantec, FireEye, but there are also some smaller ones and on top of that, public clouds such as AWS or Azure also have their own security offerings.

I don’t know how they will compete moving forward, but the numbers look pretty good so far. Like many enterprise SaaS companies filing to go public, the company hasn’t been profitable operationally yet, but the situation looks promising with increased revenue and leveraged expenses. Their growth in ARR, negative churn and customer base is impressive. At least, there is a reason to believe that they are heading to the profitability land.

The partnership with Dell, I think, will be very helpful moving forward.

What I learned from Uber S-1

Its filing is packed with a lot of information. Below are my take-aways so far from reading it

It’s doing a lot of things

Apart from the ride-hailing business that it has been known for, Uber also offers Uber Eats, Uber Freights and New Mobility, including e-bikes, e-scooters. Additionally, it has been investing in autonomous driving cars as well.

Personal Mobility

In the quarter ended December 31, 2018, the average wait time for a rider to be picked up by a Driver was five minutes.

The rapid growth and scale of our Ridesharing products, which to date have accounted for virtually all of our Personal Mobility offering, demonstrates the size of our opportunity:

• Revenue derived from our Ridesharing products grew from $3.5 billion in 2016 to $9.2 billion in 2018.

• Gross Bookings derived from our Ridesharing products grew from $18.8 billion in 2016 to $41.5 billion in 2018.

• Consumers traveled approximately 26 billion miles on our platform in 2018.

Uber Eats

Our Uber Eatsoffering allows consumers to search for and discover local restaurants, order a meal at the touch of a button, and have the meal delivered reliably and quickly. We launched our Uber Eats app just over three years ago, and we believe that Uber Eats has grown to be the largest meal delivery platform in the world outside of China based on Gross Bookings. For the quarter ended December 31, 2018, the average delivery time was approximately 30 minutes.

Of the 91 million MAPCs on our platform, over 15 million received a meal using Uber Eats in the quarter ended December 31, 2018, tapping into our network of more than 220,000 restaurants in over 500 cities globally.

Uber Freights

We serve shippers ranging from small- and medium-sized businesses to global enterprises by enabling them to create and tender shipments with a few clicks, secure capacity on demand with upfront pricing, and track those shipments in real-time from pickup to delivery. We believe that all of these factors represent significant efficiency improvements over traditional freight brokerage providers. Since Uber Freight’s public launch in the United States in May 2017, we have contracted with over 36,000 carriers that in aggregate have more than 400,000 drivers and have served over 1,000 shippers, including global enterprises such as Anheuser-Busch InBev, Niagara, Land O’Lakes, and Colgate-Palmolive. Uber Freight has grown to over $125 million in revenue for the quarter ended December 31, 2018

Impressive growth has slowed down

Really impressive growth, but a further look reveals that the growth seems to slow down

Their market map indicates that there is not much room for further horizontal expansion. What Uber can do is to dig deeper in each market to gain more market share. Uber said that as of the quarter ended December 31st, 2018, 74% of their trips and 52% of their Gross Bookings were from outside of the US.

It hasn’t made money operationally yet

Operationally, Uber hasn’t made any money. A positive sign is that their revenue grew faster than their operating loss. In 2018, their operating loss was more or less at the same level as it was in 2016 even though revenue grew significantly in the same period

Regulations, Regulations, Regulations

Throughout the filing, regulatory challenges are repeatedly mentioned and for a good reason. Uber’s struggle with authority bodies around the world has been well documented. Below is what Uber said specifically how regulations restrict their ride-sharing operations in a few countries

We plan to grow our current SAM by expanding further into our six near-term priority countries, Argentina, Germany, Italy, Japan, South Korea, and Spain, where our ability to grow our Ridesharing operations to scale is currently and may continue to be limited by significant regulatory restrictions

In 2018, we derived 24% of our Ridesharing Gross Bookings from five metropolitan areas – Los Angeles, New York City, and the San Francisco Bay Area in the United States; London in the United Kingdom; and São Paulo in Brazil. Over the same period, we generated 15% of our Ridesharing Gross Bookings from trips that either started or were completed at an airport, and we expect this percentage to increase in the future.

If some regulations are imposed in those important markets or around airports, however likely or unlikely, it may meaningfully affect Uber’s revenue.

Another aspect related to laws is how Uber classifies its drivers. Here is what it said specifically on the matter

Our business would be adversely affected if Drivers were classified as employees instead of independent contractors

If politicians in the markets where Uber has operations decide to force the company to treat its drivers as employees and give them minimum wage, it may be an issue


The first of the norms Uber laid out in “How We Approach The Future” section reads: “We do the right thing. Period”. Not really surprising, but a welcoming sign from a company that endured a public backlash symbolized by the hashtag #DeleteUber not so long ago.

In the filing, Uber promises to “release a transparency report, which will provide the public with data related to reports of sexual assaults and other safety incidents claimed to have occurred on our platform in the United States.” this year. Another welcoming sign.

Furthermore, unlike other tech companies, Uber won’t have dual share structure which is implemented to give the founders, usually, more voting rights. For example, Mark Zuckerberg has more than half of the voting rights at Facebook.

Their stakes in strategic partnerships

In August 2016, we completed the sale of our operations in China to Didi in exchange for an approximate 18.8% interest in Didi, which, based on our current information, we estimate to be 15.4% as of September 30, 2018. In February 2018, we consummated a joint venture with Yandex whereby we and Yandex each contributed our operations in Russia/CIS to a joint venture which we refer to as the Yandex.Taxi joint venture. We received a 38.0% interest in the Yandex.Taxi joint venture at the closing of the transaction, which, based on our currently available information, we estimate to be 38.0% as of December 31, 2018. In March 2018, we completed the sale of our operations in Southeast Asia to Grab in exchange for a 30.0% interest in Grab, which, based on our currently available information, we estimate to be 23.2% as of December 31, 2018. We measure our interest in each of our minority-owned affiliates based on the outstanding shares of capital stock on an as-converted basis but without taking into account securities exercisable or exchangeable for shares of capital stock or its equivalent (including outstanding vested or unvested stock-based awards and any reserved but unissued stock-based awards under any equity incentive plan of our minority-owned affiliates).

Its business deals with Google

According to the filing, Uber paid Google from Jan 1, 2016 through December 31, 2018, $631 million, $70 million and $58 million for Marketing & Advertising (Ads), technology infrastructure & enterprise services (which I believe is Google Cloud Platform), and Google Maps respectively.

What I learn from Pinterest IPO Filing

Pinterest, that photo bookmarking tool, filed this Friday to go public. Here is what I learned from reading their S-1:

Their main users are women

Pinterest reaches more than 250 million monthly active users, two thirds of whom are female. In the United States, our total audience includes 43% of internet users, according to an independent study by Comscore based on total unique visitors to our service. This includes eight out of 10 moms, who are often the primary decision-makers when it comes to buying products and services for their household, as well as more than half of all U.S. millennials. We expect to continue to grow our user base over time, especially in international markets

Pinterest S-1
An example of successful marketing with Pinterest

The majority of their Monthly Active Users is outside the US, but most of their revenue comes from the US

Quarterly Revenue (in $ thousands)
Monthly Active Users (in millions)

They did make money on two quarters and their business seems to have some seasonality

Their non-cancelable long-term contractual commitments

In addition, as of December 31, 2018, we had approximately $731.1 million of long-term contractual commitments that are not cancelable. In March 2019 we also entered into a lease for office space to be constructed near our current headquarters campus for which we will be subject to total non-cancelable minimum lease payments of approximately $420.0 million beginning in 2022 if certain contingencies are met.

Some particularly concerning risks

• We may not be able to develop effective products and tools, including measurement tools, for advertisers
• We may not succeed in further expanding and monetizing our platform internationally.
• We have incurred operating losses in the past, anticipate increasing our operating expenses, expect to incur operating losses in the future and may never achieve or maintain profitability.

Reliance on Facebook and Google Single Sign On

A significant number of Pinners use their Facebook or Google login credentials to access their accounts on our service. If security on those platforms is compromised, if Pinners are locked out from their accounts on those platforms or if those platforms experience an outage, Pinners may be unable to access our service. As a result, user growth and engagement on our service could be adversely affected, even if for a temporary period. For example, in the second quarter of 2018, Facebook changed its login authentication systems, which negatively impacted our user growth and engagement in that period. Additionally, if Facebook or Google discontinue single sign-on or experience an outage, then we may lose and be unable to recover users previously using this function, and our user growth or engagement could decline.

Personal Thoughts

Personally, I am not very bullish on Pinterest’s prospect. The good news is that it managed to make more money from its US base more in 2018 than it did in 2017. The bad news is that it has failed to monetize its majority of its MAUs which are outside of the US. To really grow and consistently be profitable in the future, I think Pinterest needs to rectify the situation overseas soon.

It acknowledged the filing that it may not be able to achieve profitability in the future. While it may not be uncommon to have a startup acknowledge that, it is unsettling to have it from an advertising company. Facebook might be in the same boat for a while after its IPO, but back then, Facebook didn’t have the same competition as Pinterest does now. In addition to Facebook and Google, there is a little problem called Amazon, which has grown to be a force to be reckoned with in the advertising world with its superior ability to connect advertising dollars with actual sale.

So far, I haven’t come across any report on privacy violations or scandals related to fake news, propaganda or human rights abuse, the likes of which took place on Twitter and Facebook. It remains to be seen how Pinterest would handle such issues in the future once and if it grows bigger.

Zoom filed to go public

Here are a few notable points I got from reading the S-1 by Zoom today:

  • Revenue grew more than 5 times from fiscal year ending 31 Jan 2017 (FY2017) to fiscal year ending 31 Jan 2019 (FY2019). Top line grew by more than 118% from FY2018 to FY2019. Gross profit was 81% in FY2019. Cost of Revenue as % of Rev continued to decline from 20.5% in FY2017 to 18.5% in FY2019
  • Operating income was positive for the first time in FY19 and made up 1.9% of revenue
  • As of January 31, 2019, remaining performance obligations was $311.7 million, with billed consideration of $125.8 million and unbilled consideration of $185.9 million. Zoom expected to recognize 67% of the remaining performance obligations (equivalent to roughly $209 million) in 2020. The average term for multi-year contracts is 2.4 years.
  • Customers that contribute more than $100,000 in annual revenue make up 30% of Zoom’s revenue in FY19 (about $99 million). There are 344 customers in this bucket as of January 31, 2019, compared to 143 a year before. On average, a customer in this bucket brought $288,000 in annual revenue to Zoom
  • Regarding customers with more than 10 employees, this segment made up 78% of Zoom’s revenue in FY19 or $257 million. With approximately 50,800 customers in this segment, each contributed on average $507 to Zoom. Customers with more than $100,000 in annual revenue contribution are a subset of customers with more than 10 employees
  • For the fiscal year ended January 31, 2019, greater than 50% of the Fortune 500 had at least one paid Zoom host, compared to only 4% that contributed more than $100,000 of revenue.
  • New customers accounted for 44% of the increase in revenue in FY19
  • 74% of FY19 revenue came from annual and multi-year subscriptions
  • Zoom operates in the Hosted / Cloud Voice and Unified Communications, Collaborative Applications and IP Telephony Lines markets. Estimated Total Addressable Market is $43.1 billion (by IDC)
  • 18% of FY19 revenue came from EMEA and APAC

A few notable customers:

  • Chan Zuckerberg Initiative: 100 Zoom Rooms
  • F5 Networks: 190 Zoom Rooms
  • National Australia Bank: 33,000+ users
  • Uber: 2,000 Zoom Rooms, 14 million Zoom minutes/month in 2018 (20,000 employees) and CAST >= 95
  • VMWare: 19,000 Zoom users and 41 million meeting minutes in January 2019 alone, by a million attendees, CSAT: 96%
  • Prior to 2014, BAYADA Home Health Care averaged 100 meetings a week. In 2018, the figure rose to 2,000 Zoom meetings a week

I have used the product extensively during my 2.5 years in the US, either in an academic or professional capacity. It’s great and easy to use. Zoom’s high CAST and impressive revenue growth indicate that my assessment of the tool is shared by a lot of users. That is a great sign for a growth business. In addition to the numbers, it’s crucial to have a great product that users love and trust.

Regarding the numbers, impressive revenue growth! Plus, only 18% of Zoom’s revenue came outside of the US and there is a lot of TAM to gain more ($330 million in revenue in a $43bn market). I won’t be surprised if Zoom continues its hot streak of revenue growth in the next 2-3 years and generate a lot of profit now that the company is already profitable. Further than that, I have no clue. The market Zoom is in is competitive with whales such as Cisco and Microsoft, in addition to many other competitors. I wish they had disclosed their total number of customers as it would have shed more light on the business.