Questions to ask as a Manager that can make any idea better

 One of the key responsibilities of an engineering manager/direction/leader in any organization is to ask the right questions. Here is a list of questions to ask to make almost any idea better

1. What does success look like and how to measure/track it reliably ? 
2. How does this tie to the north star/mission/vission ? 
3. Who are we solving for ? 
4. What problem are we solving ?
5. Recursively ask why is this a problem till you find the root cause. Always treat the root cause rather than the symptom. 
6. What makes this challenging ?
7. What is the P50 goal - 50% chance of hitting success and why is this a P50 goal
8. What tradeoffs are we making ? 
9. What is the cost benefit analysis of the tradeoffs ?
10. What are the prioritization principles being used behind these tradeoffs ? 
11. What is the opportunity cost of doing this over something else ? 
12. What ideas are we not able to fund because we are funding this ? 
13. What are the constraints we are operating in and what happens if fundamentally relax/change any of those constraints ? What is the timeline of the change of those constraints. 

Volatile, Synchronized, Re-entrant lock, Condition variable

Volatile

• The volatile concept is specific to Java. Its easier to understand volatile, if you understand the problem it solves.
• If you have a variable say a counter that is being worked on by a thread, it is possible the thread keeps a copy of the counter variable in the CPU cache and manipulates it rather than writing to the main memory. The JVM will decide when to update the main memory with the value of the counter, even though other threads may read the value of the counter from the main memory and may end up reading a stale value.
• If a variable is declared volatile then whenever a thread writes or reads to the volatile variable, the read and write always happen in the main memory.
• As a further guarantee, all the variables that are visible to the writing thread also get written-out to the main memory alongside the volatile variable. Similarly, all the variables visible to the reading thread alongside the volatile variable will have the latest values visible to the reading thread.
• Volatile comes into play because of multiples levels of memory in hardware architecture required for performance enhancements.
• If there’s a single thread that writes to the volatile variable and other threads only read the volatile variable then just using volatile is enough,
• however, if there’s a possibility of multiple threads writing to the volatile variable then “synchronized” would be required to ensure atomic writes to the variable.

Synchronized

• Java’s primary tool for rendering interactions between threads predictably is the synchronized keyword. Many programmers think of synchronized strictly in terms of enforcing a mutual exclusion semaphore (mutex) to prevent execution of critical sections by more than one thread at a time. Unfortunately, that intuition does not fully describe what synchronized means.
• Synchronize is more than mutual exclusion : The semantics of synchronized do indeed include mutual exclusion of execution based on the status of a semaphore, but they also include rules about the synchronizing thread's interaction with main memory. In particular, the acquisition or release of a lock triggers a memory barrier -- a forced synchronization between the thread's local memory and main memory. (Some processors -- like the Alpha -- have explicit machine instructions for performing memory barriers.) When a thread exits a synchronized block, it performs a write barrier -- it must flush out any variables modified in that block to main memory before releasing the lock. Similarly, when entering a synchronized block, it performs a read barrier -- it is as if the local memory has been invalidated, and it must fetch any variables that will be referenced in the block from main memory.
• The proper use of synchronization guarantees that one thread will see the effects of another in a predictable manner. Only when threads A and B synchronize on the same object will the JMM guarantee that thread B sees the changes made by thread A, and that changes made by thread A inside the synchronized block appear atomically to thread B (either the whole block executes or none of it does.)
• Furthermore, the JMM ensures thatsynchronizeblocks that synchronize on the same object will appear to execute in the same order as they do in the program.

Reentrant Lock

• Java’s answer to the traditional mutex is the reentrant lock, which comes with additional bells and whistles.
• It is similar to the implicit monitor lock accessed when using synchronized methods or blocks.
• With the reentrant lock, you are free to lock and unlock it in different methods but not with different threads. If you attempt to unlock a reentrant lock object by a thread which didn't lock it initially, you'll get an IllegalMonitorStateException. This behavior is similar to when a thread attempts to unlock a pthread mutex.

Condition Variable

• We saw how each java object exposes the three methods, wait(),notify() and notifyAll() which can be used to suspend threads till some condition becomes true.
• You can think of Condition as factoring out these three methods of the object monitor into separate objects so that there can be multiple wait-sets per object.
• As a reentrant lock replaces synchronized blocks or methods, a condition replaces the object monitor methods. In the same vein, one can't invoke the condition variable's methods without acquiring the associated lock, just like one can't wait on an object's monitor without synchronizing on the object first.
• In fact, a reentrant lock exposes an API to create new condition variables, like so:

Lock lock = new ReentrantLock();

Condition myCondition  = lock.newCondition();

• Notice, how can we now have multiple condition variables associated with the same lock. In the synchronized paradigm, we could only have one wait-set associated with each object.

Differences Between Lock and Synchronized Block

• A synchronized block is fully contained within a method — we can have Lock API’s lock() and unlock() operation in separate methods
• A synchronized block doesn’t support the fairness, any thread can acquire the lock once released, no preference can be specified. We can achieve fairness within the Lock APIs by specifying the fairness property. It makes sure that longest waiting thread is given access to the lock
• A thread gets blocked if it can’t get an access to the synchronized block. The Lock API provides tryLock() method. The thread acquires lock only if it’s available and not held by any other thread. This reduces blocking time of thread waiting for the lock
• A thread which is in “waiting” state to acquire the access to synchronized block, can’t be interrupted. The Lock API provides a method lockInterruptibly() which can be used to interrupt the thread when it’s waiting for the lock

 

Amazon's 2020 review

 In the beginning of the pandemic, I had published a post on "How Coronavirus will expedite the Amazonification of the world" and also had put my money where my mouth was, ie invested in Amazon. The other day I was reading an article on how marketplaces faired in 2020 and here are few highlights on Amazon

• 3P Amazon marketplace added eBay's worth of sales to its GMV in 2020. 3P went from 200bn to 295bn. Overall GMV was up 42%. Marketplace share of Amazon GMV grew to 62%. 

• Revenue for all business categories grew by 25%+ in Q3 demonstrating across the board strength

• Amazon's marketplace is even more concentrated than Parreto principle would suggest. Just top 852 (.05%) of the sellers contribute 10% of the GMV, and top 12% of the sellers contribute 80% of the GMV. 

• 18% of US GMV came from 2015 seller cohort showing stickiness in the platform

• Investors pour 1bn into buying up small merchants on Amazon. The capital pouring is validation about the robustness of Amazon marketplace as a platform to grow businesses. Also as the 3P sellers get better capitalized, this line of business can continue to see eye popping numbers. 

• Amazon launched in 3 countries this year - Netherlands - March 10, Saudi Arabia - June 17 and Sweden - Oct 28th. Using auto-translated catalogues and reviews, the cost to launching new countries seeded with sellers already serving similar countries is too small. This will only expedite Amazon's international expansion. 
• Ads revenue came in at 20 bn and advertising revenue as a share of marketing expense is steadily rising. This is similar to what Chamath had previously pointed out, Amazon thinks about making every expense item a revenue bullet point. 

References

• Marketplaces year in review 2020
• Investors pour 1bn into buying up small merchants on Amazon
• Inside the whale : an interview with an anonymous Amazonia
  
  

The Amazing Vaccine Race

 Few astounding facts which will be table stakes for the future

• The digital copy of the virus from China reached America even before the biological/physical copy of the virus on Jan 11 thanks to the advances of Genomic Sequencing Tecnology
• Moderna was able to design a vaccine with just the digital copy without the actual physical copy in a matter of 48 hours
• The team at NIH and the team at Moderna had the exact same copy of the vaccine after they compared the results
• At the same time they started to make a clinical grade product that could go into Phase 1 trials
• The vaccine that was reviewed by FDA in December 17 was the exactly the same vaccine that was designed by Moderna in Jan

What is mRNA technology ? 

Its a molecule that exists in everyones cells that is like a xerox copy of the instructions of your genome to one gene at a time to make protein. 

When cells want to make insulin, it makes a zerox copy of the instructions from the DNA. Ribosome (like a 3D printer) will read the mRNA instructions and makes a protein adding one amino acid at a time. 

How to use mRNA ? 

Most viruses in life including COVID-19 contain mRNA. Through evolution mammals have developed mechanisms to recognize foreign mRNA. People abandoned and quit on trying to make mRNA as a drug. In 2010-2011, a set of academics at Harvard and MIT, played with mRNA. They believed that if you modify you can make an mRNA that is immunosilent. 

1 Protein of the virus - spike protein - this is the instruction that is being passed through the mRNA technology in the vaccine. The cell knows then how to create this protein. Then the body knows that this foreign protein is in the body and starts making antibodies. The spike protein sticking out the cell mimicks the real virus and the immune system starts fighting it.

How to know the Spike protein ? 

Spike is what the virus uses to get into the cells. Spike protein had been used to protect against MERS in animals. 

Future mRNA verticals

• mRNA for Tumours
• mRNA for heart 
• Vaccines

Networking Resources for Engineering Managers

Engineering Management can be a lonely job. Also, people tend to get better at Management only with experience. In the journey, it can be very helpful to have a peer group of other engineering manager, with whom one can discuss, debug and learn from their challenges. This group can be ones mentors, sponsors, past managers, network of weak connections who can help in the Engineering Manager leadership development journey.

Below are the list of resources to hone and develop a network as an Engineering Manager. These resources can be helpful for high TLs or TLMs looking to convert to an EM role, new EMs to learn from each other's experiences, seasoned EMs to grow their network outside work : 

• Engineering Management group on Facebook
• Rands leadership slack channel 
• https://engmanagers.github.io/ - eng managers slack channel
• Engineering Manager group on linkedin

Compounding M1 Money Supply

M1 acceleration(new)

1st trillion-80 years(1913–1993)
2nd trillion-18 years(1994–2011)
3rd trillion-5 years(2012–2016)
4th trillion-4 years(2017–2020)
5th trillion-6 months(Feb–Jul 2020)
6th trillion-4 months(Aug-Nov 2020)
7th trillion - 2 months?

 

Airbnb S1 Analysis

Summary

• Crazy effect of the pandemic shows what V shaped recovery looks like and revenue resilience
• Healthy marketplace dynamics shows growth on both demand and supply side
• Pandemic could still dampen growth rates in the coming 2 years
• Company poised to grow and do well over the next 10 years aligning with new trends and making improving lives of guests, hosts and communities
• Business has resilient revenue and durability due to 2 sided marketplace and local and global network effects

Key Metrics

• 4 million hosts
• 220 countries and regions, 100,000 cities, 55% hosts are women
• 110 million guests cumulative in 13 years
• 110 billion host earnings total in 13 years

COVID Impact

• Airbnb business declined significantly during COVID and rebounded after three months
• During Mar, Apr, May 2020, Airbnb Gross Nights booked crashed 42%, 72% and 50% respectively YoY. After that it has stabilized at -20% YoY.
• Even though international travel has been limited, domestic travel has rebounded showing resilience of the business
• As COVID, blurred the lines between work and life, it also blurred the lines between travel and living. A new trend has started "work from Airbnb" or a "Workation" which is Work + Vacation 

The below chart shows the crazy effect the pandemic had on Airbnb

Business and Strategy Commentary

• Two sided market place with guests(demand) and hosts(supply)
• Increase the supply of hosts
• 79% of the hosts coming to the platform organically to sign up in 2019
• 23% of the new hosts are converting from guests - this is a very healthy sign
• Attract new guests onto the platform(Acquisitions) + Get the current guests to use the platform more (retention and resurrection)
• 91% of the traffic came to AirBnB from organic or direct channels meaning acquisition cost can be negligible. The platform has good content which is driving users there
• 69% of revenue was generated from repeat guests meaning guests who use the platform once like it and continue to come back showing the platform is healthy
• These numbers really make the business likeable because there is low acquisition costs both on the supply and demand side. Hosts are getting economic benefits by monetizing assets, guests are getting choices to buy an wide array of experiences which is attracting both guests and hosts and making the marketplace stronger and stronger. 
• International Expansion will be a future driver for growth specially in countries where penetration is lower like India, China, LaTaM, SEA. Has both local and global network effects
• Categories  resilient to COVID
• Domestic
• Short distance stays - within 50 to 500 miles
• Long term stays > 28 days
• Travel outside the top 20 cities
• Investing and developing the brand 
• Create new experienes

Financials

• 2.5 bn revenue in 2017, 3.6 bn revenue in 2018, 4.8 bn in 2019
• For the 9 months revenue comparison is 3.6 bn vs 2.5 bn in 2019 vs 2020
• Net loss has been 70mn, 16mn, 674mn in 2017, 2018 and 2019
• Cash in hand 4.5 billion

Bay Area is Dead - Look at the IPO class of 2020

Look at the IPO class of 2020

• Airbnb - Airbnb S1 Analysis
• Asana - a work management platform
• Unity - a 3D video game dev platform
• Snowflake - a data warehousing startup
• Palantir - Peter Thiel's data products startup for gov
• Sumo Logic
• Doordash

If we just include Airbnb and Doordash, they have 150 billion worth of market cap after first week of trading. If we assume even 10% of that went to employees, thats 15 billion of liquidity coming to the bay area only from these two IPOs. Even at half the valuation, its will create a new crop of home buyers, new seed rounds, new angel investments that will plant the seeds of the future IPOs. 

Also there has been lot of talk about HP and Oracle leaving the bay area. In all fairness, Bay area left these companies long before they left Bay Area. It was hard for them to hire talent in a competitive market given their brand, vision and strategy. Also given their uncompetitive pay, very hard to see why folks will prefer them over companies with the chance of going IPO. I think in the short and medium term, it is great news as they will make room for other new startups to secure leases in the bay area. 

Influence over authority

To influence without authority, ask yourself:

1) What do I want them to Know ?

2) What do I want them to Do ?

3) How to do this ? 

4) What do I want them to Think About?

5) How do I want them to Feel?

Want to force ? Come across as authoritative or directive style of management

Fixate on 1, 2, 3. Yields short term results. 

Inspire and Influence ? Build for the longer term and hone everyone's drive, passion and agency

Focus on 4 and 5. 

Read more about Influence : The psychology of persuasion .

Common Lessons from running Growth teams, Investing and Genomic Biology - a hypothesis is a liability

 “When someone is seeking, it happens quite easily that he only sees the thing that he is seeking; that he is unable to find anything, unable to absorb anything, because he is only thinking of the thing he is seeking, because he has a goal, because he is obsessed with his goal. 

Seeking means: to have a goal; but finding means: to be free, to be receptive, to have no goal. You, O worthy one, are perhaps indeed a seeker, for in striving towards your goal, you do not see many things that are under your nose.” - Hermann Hesse

There is a hidden cost to having a hypothesis. It arises from the relationship between night science and day science, the two very distinct modes of activity in which scientific ideas are generated and tested, respectively. With a hypothesis in hand, the impressive strengths of day science are unleashed, guiding us in designing tests, estimating parameters, and throwing out the hypothesis if it fails the tests. But when we analyze the results of an experiment, our mental focus on a specific hypothesis can prevent us from exploring other aspects of the data, effectively blinding us to new ideas. A hypothesis then becomes a liability for any night science explorations. The corresponding limitations on our creativity, self-imposed in hypothesis-driven research, are of particular concern in the context of modern biological datasets, which are often vast and likely to contain hints at multiple distinct and potentially exciting discoveries. Night science has its own liability though, generating many spurious relationships and false hypotheses. Fortunately, these are exposed by the light of day science, emphasizing the complementarity of the two modes, where each overcomes the other’s shortcomings.

The gorilla experiment

Many of us recall the famous selective attention experiment, where subjects watch a clip of students passing a basketball to each other. If you have not seen it, we recommend watching it before continuing to read. 

As you watch the two teams in action, your task is to count the number of passes made by the team in white. About halfway through, a person dressed up as a gorilla enters the foreground. The gorilla pauses in the center, pounding its chest with its fists, before exiting to the opposite side of the frame. Surprisingly, half of us completely miss the gorilla, as we are focused on counting passes, even though hardly anyone overlooks it when simply watching the clip without the assignment.

We believe that a similar process happens in various phases of life. 

Growth teams

Growth teams are laser focussed on hypothesis building, executing on the hypothesis, tracking progress against goals measured using metrics assumed to be impacted by the hypothesis. Some of these experiments lead to metrics wins and some of them not so much. It is equally important to distill learnings from the wins and the failures and articulate them to build the next hypothesis. This kind of learning at the project level, aggregated produces the future direction of the team and informs the organization level strategy in growth companies. Hence, the failures even though they seem innocuous or wasted opportunity at certain points serve the valuable purpose of informing future roadmap direction and strategy for the organization. This helps form the future organization level hypothesis, goal and metric. Also in committing to a hypothesis based on these learnings, there is also a pseudo commitment about not investing on other hypothesis (in case of limited team bandwidth). Hence the process of distilling the learnings in growth teams and the quality of the learnings serve the lifeblood of the organization. 

Investing

Similar idea can also be applied to active portfolio managers who are making certain bets driven by hypothesis. It is very important to shed biases and question each and every assumption behind hypothesis building. 

Interestingly this concept formalized appeared in this journal on genome biology. I merely tied this concept to other spheres of life. 

Comprehensive vs Interesting tradeoff

 Being comprehensive vs interesting:

You might want to give all the details, explain all the things fully in one breath. But you risk overwhelming your audience.

Optimize for being interesting instead. If you hook their interest you’ll earn the chance to share more later. 

How to ask great questions as a PM ?

 1)

Less “How will we build this?”

More “How will we differentiate?”

2)

Less “How to enforce accountability?”

More “How to foster ownership?”

3)

Less “What problems can we solve?”

More “What problems are worthwhile?”

4)

Less “What is the 3-yr roadmap?”

More “What is the 3-yr strategy?”

5)

Less “How to run growth experiments?”

More “How to get more distribution?”

6)

Less “What is the process for X?”

More “What is the purpose of X?”

7)

Less “Does this team run well?”

More “Does this team learn well?”

8)

Less “What are top user requests?”

More “What are top user needs?”

9)

Less “What is the template for Y?”

More “What is my goal with Y?”

10)

Less “What is the schedule?”

More “What are the priorities?”

11)

Less “Are all stakeholders happy?”

More “Are all stakeholders aligned?”

12)

Less “How many resources do we need?”

More “What is the marginal impact?”

14)

Less “How can I use metrics?”

More “How can I use psychology?”

15)

Less “Who will write the blog post?”

More “How can we create excitement?”

16)

Less “What will get me promoted?”

More “What will get the buyer promoted?”

17)

Less “How did Google solve this?”

More “How are we different?”

18)

Less “What does the CEO want?”

More “What does the CEO know?”

19)

Less “What are competitors saying?”

More “What is their strategy?”

20)

Less “What is rational for users?”

More “What is natural for users?”

For the original content and discussion, please read Shreyas's thread

Ellsberg Paradox and Option value of investing

 Ellsberg Paradox 

People prefer to take risks in situations where the odds are known, rather than a scenario where the odds are unknown - even when the latter scenario has the guarantee of a positive outcome (its just that the magnitude of the outcome is unknown). Its often used to evaluate how people have an aversion to ambiguity. 

Where do the traditional financial tools fall short ? 

In traditional investing, the most common financial tool for valuing a company is DCF model (discounted cash flow model). Under this methodology, investors attempt to accurately model out the discrete financial metrics of a company over a finite period of time, and discount the cash flow generated to determine the appropriate valuation for a company.

The issue is though that in real investing, businesses have embedded options which have unknown outcomes everywhere. DCF is terrible at valuing these businesses which have both : 1. uncertain payoff magnitude and 2. uncertain timing as to when it will occur. Examples of such options : 

• Amazon is able to extend its dominance on one ecommerce category(books) to multiple categories
• Amazon is able to extend its dominance in ecommerce to build the largest retail search engine in the world and draw advertising dollars
• Google is able to leverage its technology prowess in building scalable and reliable search infrastructure to build Google Cloud and democratize building distributed systems technology. Same for AWS and Azure
• Apple launching Wearables segment (Watch, airpods, etc) using its expertise in building iphones

For each of these successful options, there are multiple failed investments like Fire Phone, Google's communication apps, etc. However, in net all these embedded options have generated multiples of shareholder value which would not have been able to be correctly valued through DCF. 

This paper articulates this concept in detail : Get real, using real options in security analysis

This diagram illustrates how the valuation breakdown of such companies progresses

Some of the parameters to evaluate such businesses are

Management 

• Superior capital allocation 
• Access to cheap capital from markets or through high profit margin businesses
• Execution track record
• Ability attract top talent in the industry

Business

• Strong moat
• Economies of scale
• Economies of scope

Evolving markets

• New trends of user behavior and patterns
• Uncertainty

No wonder FAANGs check several of the above boxes and continue to generate superior returns for their shareholders. 

References

1. Hayden Capital Quarterly Letter Q3 2020

2. Get real, using real options in security analysis

Return to India - Financial Guide

Scenarios

- Planning to return to India 

- Keeping US accounts as a non-resident alient

What are the implications of moving on brokerage accounts ? 

• Allow reasonable access, under non-resident alien terms (with a W-8BEN). There may be some restrictions on the type of thing you can trade, and you may have to send paper forms in for some things that could otherwise be done online, but overall the account functions relatively well. Buy and sell, and things like 401ks, IRA rollovers and Roth conversions are still available.

• Limit activity as noted upthread. That is, you can sell existing holdings on your own schedule but not buy any new ones. Over time, that makes the account hard to manage (rebalancing, for example).

• Force you to close your account and move your money out.

Which occurs seems to to depend heavily on which country you move to, its tax treaty status and regulatory framework, and so on.

What are the tax implications ? 

It depends which country you live in. If it has an income tax treaty with the US, you will pay US tax on dividends at the treaty rate. It's typically 15%, but could be higher, perhaps 25%. A few countries have a 10% rate. If it has no income tax treaty, you will pay 30% in US tax on dividends. Vanguard will take the correct rate automatically through withholding once you've sent them a W-8BEN. It's a flat rate tax, so you cannot recover any of it from the IRS. You might however be able to use it as a credit against local income tax on these dividends.

No US capital gains tax implications. The US does not tax capital gains for nonresident aliens. Watch out though for US estate tax. If you country lacks a US estate tax treaty, your heirs could face 26%-40% of the balance of your holdings above $60,000 should the worst happen. This applies also to any IRA or 401k accounts you hold in the US. The US estate tax treaties with Ireland and South Africa are reportedly deficient, so best not to rely on those.

What about tax returns ? 

If you're neither a US citizen (or resident) or a green card holder, if you have to file anything it's always a nonresident alien return, so a 1040-NR.

In general though, you should have no need to file one. Provided Vanguard applies the correct US tax withholding on dividends, your US tax liability will exactly match your US tax withholding, and in that case you don't need to send any US tax return. See 1040-NR instructions for more.

If for any reason you do have to file a 1040-NR, perhaps Vanguard overwithheld relative to your treaty rate, you'd only have to declare your US source income on that. You don't tell the IRS anything about your non-US earnings, interest, or anything else financial happening in your (non-US) country of residence on that form. 

The IRS will only issue tax refunds in USD, either as a cheque or ACH payment to a US bank account.

Some tips

1. Remittance from brokerage accounts to directly foreign accounts through transferwise if possible. I am not sure this works. Seems like a hassle. 

2. Interactive Brokers seems to have a global presence

3. 

Appendix Links

1. United states income tax treaties 

2. India tax treaty documents 

3. US Estate and Gift tax treaties

4. Interaction of Indian and US tax laws

5. Keeping vanguard accounts as a non-resident alien

6. Bogglehead link on non-residen alien taxation

7. Beware the taxation angle if you want to invest in direct foreign equities from India - Reddit thread

8. Taxation for international residents

9. Non-US investors guide to navigating US tax traps

10. A guide to selling US property for foreign residents and expats

Outlook for the Cloud Market

Satya's description of why Cloud is a secular trend from Microsoft earnings call

The way I think about the computing landscape going forward is if you sort of said at the highest of levels today as a percentage of GDP, tech spend is 5%. We think it will double in the next 10 years. And if anything, this pandemic perhaps has accelerated that doubling.

And in that context, what's the large — the most secular need, it's the need for distributed cloud infrastructure. It's both needed for modernizing existing applications you have, and so that's why, by the way, 20% penetrated so there's more 80% that needs to move. But more importantly, there's going to be new application starts which need infrastructure. And so if you sort of add those up, I think that we're still in early innings.

There will be, between quarters, volatility, all of the points that Amy made even earlier. But we think distributed cloud infrastructure is the most important layer. But the way we have approached it is not to just think of that layer in isolation but the data layer work we do composes, the AI layer composes and more importantly, our SaaS applications, whether business applications, Power Platform, Microsoft 365, all reinforce that same modern tech stack. So I would still say that digitization in its — this new tech stack is in its very infancy. - Satya Nadella

Milestone moment

This is the first year when Azure revenue is larger than Windows revenue. Azure revenue has grown to 17% of the revenue. 

Google's Cloud Strategy

• Six priority industries : healthcare, retail, media, finance, manufacturing, public sector
• 5 major geographies
• 4 customer segments
• Data processing, analytics, AI/ML - differentiation
• Migration of legacy datacenters to google cloud and cut in IT spending and infrastructure. eg : Nokia migrating 30 data centers to the cloud across 12 countries
• Google meet saw a peak of 7.5 billion daily video call minutes

How Big Tech Makes Money - Q3, 2020

Apple 

• Net sales : 64.7 billion
• Product : 50.1 bn
• iPhone : 26.4 bn (33bn in Q3, 2019)
• Mac : 9 bn (> Q4 2019 revenue 7.56 bn)
• iPad : 6.7 bn (> Q4 2019 revenue 5.9 bn)
• Wearables :7.8 bn
• Services : 14.6 bn (> Q4 2019 revenue 12.6 bn)
• Cost of sale : 40 bn
• Operating expense : 10 bn
• Operating Income : 14.7 bn
• Net Profit : 12.7 bn
• Cash in hand ~50 bn
• Commentary : Overall revenue and profit is constant from 2018 and 2019. However, stock price has gone up as the company has reduced the number of outstanding stocks through buy backs. However, PE has gone up from 12 to 35 and future growth seems to be priced in, until Apple opens up new revenue streams. Apple shared no revenue guidance for Q4

Amazon

• Net sales : 96 bn
• Operating Income : 6.1 bn
• Net income : 6.3 bn
• All the above figures are higher than Q4 2019
• Investing in Amazon provides significant international exposure : NA 62%, International 26% and AWS 12% of net sales. YoY net sales grew by 39%, 37%, 29% across those 3 categories
• Online stores 48 bn, Physical stores (includes whole foods) 3.7 bn, Third party seller services 20 bn, subscription services 6.5 bn, AWS 11.6 bn, Ads 5.3 bn
• Online stores grew at 37%, physical stores at -10%, third party seller services at 53%, subscription services at 32%, AWS at 29% and ads at 49% YoY
• Commentary 
• Physical stores slowdown is expected due to Covid, but all other categories grew faster than AWS. Net sales increased 37% YoY. With Covid-19 expediting Amazonification of the world, this will be an unprecedented holiday season. Amazon expecting 112bn in net sales in Q4. 
• Some of the demand from COVID like grocery, gloves may not recurring next year. But COVID is increasing engagement, retention of the Prime membership program and usage is growing across different categories
• Prime videos is a very good acquisition channel for Prime members - > higher membership renewal rates and higher engagement
• International has been profitable for 2 quarters now

Microsoft

• Revenue 37.2 bn (12% growth)
• Productivity and business processes : 12.3 bn (11% growth)
• Office commercial products revenue growth of 9% (Office 365)
• Office consumer products revenue growth on 13%
• Linkedin 16% up
• Dynamic products up 19%
• Intelligent cloud 12 bn (20% growth)
• Azure up 48%
• Personal Computing 11.8 bn (6% growth)
• Windows OEM down 5%
• Windows commercial products up 13%
• XBox content and services up 30%
• Surface revenue up 37%
• Search advertising down 10%
• Net income : 13.9 bn
• 9.5 bn returned to shareholders, 5.3 bn share repurchases and 4.2 bn dividends

Google

• Revenue - 46 bn (up 14% from Q3 2019)
• Search ads 26.3 bn
• Youtube ads 5 bn
• substantial growth in direct response
• brand advertising (create brand and interest)
• YTs strong watch time growth enables advertisers to reach people who they cant reach on TV
• Google network 5.7 bn
• Cloud 3.4 bn
• Data processing, analytics, AI/ML
• Migration of legacy datacenters to google cloud and cut in IT spending and infrastructure. eg : Nokia migrating 30 data centers to the cloud across 12 countries
• Google meet saw a peak of 7.5 billion daily video call minutes
• Six priority industries : healthcare, retail, media, finance, manufacturing, public sector
• 5 major geographies
• 4 customer segments
• Other revenues (YT subscriptions, Google play) 5.4 bn
• Other bets 178 mn
• TAC : 8.16 bn
• Net income - 11 bn (operating margin 24%)
• Losses on other bets : 1 bn, nicely masks the operating margin from 27% to 24%
• Cash on hand : 133 bn
• International exposure : US 21.4 bn, EMEA 13.6 bn, APAC 8.4 bn, Other americas 2.6 bn
• Google has greater than 50% ex-US revenue exposure
• PE (1.1 tn / 50 bn) ~ 22. Alphabet looks cheap going forward given it is growing revenue at double digits compared to the rest of big tech. 
• Youtube music has 30 mn paying customers and Youtube TV has 3 mn paying customers
• In Q4, 2020 GCP will be its own reporting segment
• Commerce
• Shopping listings available in 48 countries
• Google checkout option opened the platform to paypal and shopify integration
• advertisers in Youtube at the mid-funnel level
• Earnings call transcript

Facebook

• Revenue : 21.4 bn (22% growth)
• Operating income : 8 bn (operating margin 37%)
• Cash on hand : 55 bn
• 200 mn businesses, 10 mn active advertisers, 40 mn people viewing a business catalog every month, more businesses are doing things online due to COVID
• Earnings call transcript

Commentary

• Big Tech companies are generating revenue worldwide. All of them have considerable international revenue exposure. 
• Amazon (37%), Facebook (22%), Google (14%), Apple (~flat) - revenue growth
• Facebook (37%), Google (24%), Microsoft (37%) - operating margin
• 265 bn in revenue, net profit of 52 bn, operating margin of ~20%, market cap ~7.5 tn
• Cloud is an extraordinary secular trend and all cloud players are gaining from IT spending and cost cutting
• Digital advertising continues to be a secular trend - TV dollars, CPG marketing budgets, retail goods moving online all driving this trend, organic small businesses trying to reach customers online

The Firehorse Effect

• In the early 20th century, a horse pulling a wagon would all of a sudden gallop towards a burning building nearby, endangering itself, the driver, & the passengers. What was the reason for this peculiar phenomenon? And why does this matter today?
• Fire has been one of the biggest dangers that humans have faced ever since they started living in structures. In the olden days, firefighting was a community responsibility. When a fire broke out, the people in a town or a village would form a "Bucket Brigade". 
• Visualize a double-line of people passing buckets of water from a nearby water-source to the fire, and sending empty buckets back to be refilled. That’s a Bucket Brigade. Eventually, tanks of water, hand pumps, & hoses became the preferred firefighting equipment. 
• Then came steam pumps, more powerful & efficient. Great, but the equipment got heavy & it became difficult for firefighters to pull it to wherever the fire was. What to do? 
• Enter horses. It wasn’t an easy change. Horses had to be trained to reliably run towards a fire. They also had to be strong. As horses became more commonly used to pull fire engines, places like Detroit even created a Horse College, along with report cards for each horse (!) 
• What happens when one is trained to do a job well, is systematically evaluated, & rewarded or punished based on job performance? One often gets rather good at that job. Same with these firehorses. 
• At some point, a firehorse would be retired from the firefighting job & given the job of pulling wagons on the street. A new job! Things were generally fine in the beginning. 
• But whenever a former firehorse heard a fire alarm or felt the presence of a fire nearby, it instinctively galloped towards the fire: terrorizing its drivers, passengers, & owners. 
• This firehorse did a rather poor job in its new context due to the very behaviors and patterns that were reinforced in its old context. Twitter, this is the Firehorse Effect. 
• The Firehorse Effect is why some accomplished business leaders fail to inspire, create a compelling strategy, eliminate drama, or rebuild the culture when they take on a leadership role in a new organization. The Firehorse Effect is also why a manager with a string of prior successes is just unable to execute at your startup. 
• What can be done about the Firehorse Effect? As with most such things, the solution needs to be grounded in Self-awareness, Organizational-awareness, & Sound Management. Leaders in a new setting should regularly think about the Firehorse Effect. 
• Leaders in a new setting should take the time to observe & learn. Doing is important, but doing without a deep understanding of one’s new role can be frustrating or even fatal. 
• The leader should invite people to challenge them. Ask "X has worked for me in the past, will it work here?" And the leader's managers & mentors need to help accelerate this process by candidly sharing organizational context & feedback during the leader's early days. 

This term is coined by Shreyas Doshi, I heavily recommend reading his writing for organizational and leadership development.