The Amazon Empire And What Better Might Look Like

a16z Board Partner and former Microsoft executive Steven Sinofsky had a great post on Medium (originally a tweet storm) in which he shared an alternative view on Amazon’s journey. Instead of the usual narrative about world domination, he makes the case that this Amazon’s moves are Amazon just becoming a traditional retailing empire – like Sears, WalMart, et al.

For Amazon’s logistics mastery, think of Walmart in the 1980s. For the loyalty driven by Amazon Prime, think of the Sears store credit card. For Amazon Basics, look at house brands of nearly every retailer. And, for Amazon Advertising, look at the channel structure (rebates or discounts for shelf placements) of every retailer.

Again, it is a great post and is right on most counts.

That said, while all of this points to Amazon becoming a retailing behemoth just like those of the past, I think technology available today enables a behemoth of hitherto unprecedented size, scale, and strength.

The obvious counter point to all this would be to look at the past. Walmart’s dominance felt unbeatable – until it wasn’t. But, let me explain why I look at Amazon’s dominance differently.

(1) The online-offline combination at scale is formidable. Anyone who’s taken a class in Supply Chain will tell you that it is impossible to be both broad and efficient on one channel. Online distribution and centralized warehouses are great for expensive products with high demand uncertainty (TVs, diamonds). They suck for steady demand, low value products.

Put simply, Amazon’s online business is very efficient at shipping you books or CDs or even televisions – but is very inefficient at shipping diapers. So, it was only a matter of time before Amazon expanded into physical retail (3000 stores are on the way by 2021).

Omni-channel retailing makes the prospect of a profitable “Everything Store’ very real.

(2) 4 Stars is made possible by the retailer – Go is made possible by the technology giant. Amazon announced a new store called “4 stars” in New York City. The store will feature the most popular products bought by New Yorkers on Amazon. It makes complete sense and is something Amazon is well positioned to pull off. The challenge with physical stores is optimizing sales per sq foot. Here, Amazon will only sell the most popular items. Social proof works.

You can imagine other retailers with massive online footprints (there aren’t many – but you can still imagine it) pulling something like this off.

Amazon Go, however, is about more than that. Go brings together mastery in computer vision and deep learning that has been honed for years thanks to an ecosystem that features projects like Alexa and AWS. There isn’t another retailer on the planet who can pull this off. In fact, I’d wager that there are less than ten companies on the planet that have the talent required to do this. Amazon is part of a very elite group of technology leaders.

(3) In the race to build the OS for the home, Alexa is the one to beat. Amazon has been going after the home for a few years now. After a few failed experiments, the combination of Alexa and Echo have definitely been a home run. But, in typical Amazon fashion, they’ve sought to build on that. Whether it is attempting to get to know your wardrobe via the Echo Show, be your home security vendor via Ring (acquired), and the recent Alexa microwave, Amazon is the clear frontrunner.

Again, the technology chops required to make Alexa work rules out every retailer. We’re thus left with 3 other companies with working (to various degrees) personal assistants – Apple, Google, and Microsoft. Facebook could be added to the list and will likely give it a shot too.

Microsoft is (mostly) focused on the workplace and Apple’s DNA is decidedly not mass market retail. All this brings us to Google and maybe Facebook.

We can begin by ruling out Facebook’s chances. Ads is a problematic business model when you’re creating home devices with personal assistants that listen to people’s most intimate conversations. And, Facebook has repeatedly proved it will use any and all data available (including what you share for security purposes) to target ads. That’s not going to be helpful in their quest for the home.

That, then, brings us to Google. The biggest challenge Google faces is the same as that of Facebook – it is an advertising company. But, luckily for Google and in no part thanks to its many moon shots, it has a brand that stands for more than that – at least relative to Facebook. This is getting undone with the various recent revelations about excessive tracking. That said, I’d posit that they still have hope.

And, that hope is critical because securing a foothold in retail has existential implications for Google’s growth. Steven Sinofsky’s point about Amazon’s ad business being the equivalent of retailers charging for great shelf space indicates why this is so. Since retail searches in the past two decades have generally started on Google, Google has taken home that money. But, as Amazon becomes that default place to begin the retail journey, Google will lose out heavily on future growth.

So, unless Google does what all successful retailers do and builds a successful house brand (Google Shopping), it is going to lose a massive portion of the e-commerce pie. If Google is to succeed in this, I believe Google will need to draw a strong and visible separation between its ad monetization and Google Home. If people perceive a strong connection, they will not buy.

I stress “perceive” because Amazon has a rapidly growing ad business too. Customers, however, think of Amazon as a retailer first. And, perception matters.

(4) The internet has created global (i.e. world minus China) winner take all effects – especially for platforms. The internet has enabled global domination by a few corporations at an unprecedented scale. Steven Sinofsky’s post points to Walmart’s scale in the 80s and 90s. But, I’d argue that Amazon is different because it is both a retailer and a platform of retailers (Fulfillment by Amazon).

It means Amazon’s domination is not limited to its ambitions as a retailer. Just as Facebook is the global internet’s (again, global = world minus China) default town square, Amazon is the default market with merchants from all across town. That’s something Walmart or any major retailer in the past never had.

I think that is also why the narratives around retail and ridesharing seem to be running parallel. Just as we’re seeing massive global ridesharing networks consolidate into two opposing forces (Uber, Softbank, Didi, Grab vs. Lyft, Go-Jek, Waymo), retail pits Amazon versus Walmart, Google, and The stakes are much higher compared to the 1980s and 90s.

(5) Amazon’s momentum can likely only be stopped by two forces – the government and itself. In sum, I think Steven Sinofsky is right in saying that a lot of what Amazon is doing isn’t as disruptive as is often portrayed. But, I do think equating Amazon to every other retail empire doesn’t do it justice. Its mastery of omni-channel retailing, access to cutting edge technologists who are building for Alexa and AWS, and its formidable platform chops make it really hard to beat.

As a result, I believe there are only two possible ways Amazon’s expansion will be stopped – regulation or poor strategy/execution on its part. As long as Jeff Bezos is around, my guess is that we’re effectively left with government/regulation as the only realistic option.

(6) I’m hopeful Amazon will help build out a better global retail empire. While they’re at it, I’m hopeful team Bezos will help build a better global retail empire in the coming decades. There’s benefits to massive consolidation and I’m hoping some of these benefits will be used for good.

3 areas where I’d love to see Amazon innovate –
1. Ethical global supply chains. We live in a world where everyone owns a smartphone – however, only a tiny minority of which are ethically built. Amazon’s heft means it can make ethical standards mandatory for its 3rd party retailers as well as its own supply chain. Amazon’s internal recruiting famously has “bar raisers” – folks who form part of the interview panel to ensure they’re raising standards. I’m hopeful they’ll raise the bar on supply chains too.

2. The Amazon recycling program. What if Amazon created  program to recycle its famous cardboard boxes?

What if, for example, this program recycled both boxes and goods purchased on Amazon?

3. More sustainable products. Building on the above two points, what if Amazon pushed for more sustainable products by making more of them available on Prime?

At Amazon’s scale, every step made toward sustainability will be a giant leap forward.

I’m well aware Amazon’s past record in these areas hasn’t been great. Their treatment of their warehouse employees has come under fire in the past. Better often makes margins worse and it is, after all, still day 1 at Amazon.

However, I’m hopeful there’ll come a time when Team Bezos reconciles the fact that it is also year 20 at Amazon. While they’ve (admirably) retained their scrappiness, they’re also among the most powerful corporations on the planet.

And, if responsibility is proportional to power, they certainly have a whole lot of it.

The World Cup, Google, And The Omnipresent Sorting Theory

France were crowned the new world football champions in the summer. Croatia were the runners up.

The interesting question, then – can we confidently say Croatia was the second best team in the tournament? Would they have beaten Belgium or Brazil if either team had been in their half of the bracket?

While it is likely to be just a battle of your opinion versus mine, I’d hazard a guess that Belgium and Brazil were probably better teams. The challenge is that elimination via brackets does a good job identifying the winner. But, it does a poor job of finding second place. The best approach to finding the stack rank of the teams would be to do a comprehensive round robin where everyone plays each other – equivalent to the English Premier League for example. But, we don’t have time for that in the world cup. So, brackets are the way to go.

Sorting theory
In Computer Science, a sorting algorithm puts the elements of a list in a certain order. Computers, in many ways, are just sorting machines. They take vasts amounts of data and find ways to sort them intelligently and make them available when we need them.

Sorting theory is a key part of introductory computer science and it tends to be a popular interview question topic. Barack Obama was famously asked a question about the most efficient sorting algorithms in an interview at Google. That’s how mainstream sorting questions are. :-)

The efficacy of sorting algorithms is judged by the big O notation. It is the worst case scenario of time taken to sort “n” items. The first sorting aglorithms had a worst case performance of N^2 square. This means the performance gets really bad as the number of items increases. As a result, sorting and, by extension, computation was very expensive.

This was until the Merge sort came along.

Merge sort and life applications of sorting theory
The merge sort has legendary status among sorting algorithms because it made “linearithmic” sorting possible. The worst case is NlogN – effectively MUCH faster than previous sorting algorithms. (Wikipedia has a simple illustration of how merge sort works – if you are interested)

Merge sort is what you do when you split a large pile of laundry, split it into little piles, sort those piles, and then merge them together. The process of elimination in competitions via brackets is, thus, a variant of merge sort.

As Brian Christian and Tom Griffiths share in their book “Algorithms to Live By,” there are plenty of other such life applications of sorting. For example, sorting theory teaches us to never sort email into folders if searching is easy. Sorting is expensive (read: giant waste of time).

Sorting theory also teaches us to use “bucket sort” when possible. Bucket sort or bin sort is what you do when you just separate your laundry into various “bins.” Assuming a small number of bins, it gets sorting done in linear time.

We also see sorting theory play out in every organization via the pecking order. Pecking order was first observed in a brood of chickens who figure out their leadership order by pecking away at each other. The most powerful peckers are higher up in the order. Humans do it too – it is called politics and usually just requires two humans, a room, and some incentives.

Okay, so what does all this have to do with technology and the future?
Brian Christian and Tom Griffiths have a couple of intriguing questions for us here – What is Google? The default answer would be “search engine.”  But, is it really a search engine? Or, should it be called a “sort engine?”

The power of Google is not just in the searching but in its ability to do the sorting. In fact, we could make the argument that it is the efficacy of their sorting algorithm that makes Google the massive money making machine that it is. Google rose to power sorting results into pages with 10 relevant results. Over time, they added ads to the side and then on top and became one of the most valuable companies on the planet.

Let’s take a look at what Google’s sorting looks like right now. These are screenshots as I scroll down for a search for “mattress” – I’ve captured 8 screenshots from beginning to end and I’ve encircled the ads in red.

3 observations –

1. It took me 2 entire screen scrolls before I saw the first completely organic result.

2. I counted roughly 16 organic results/groups and 7 ads.

3. I don’t think I ever scroll past screen 2 – my guess is that I’m not the exception here.

Google, then, is still a sort engine. For most retail searches, however, it just sorts more ads these days than organic results.

Who gets to do the sorting for you?
There are two lucrative marketplaces on the internet – attention marketplaces and retail marketplaces. The biggest leading indicator for the success of a marketplace is whether or not they do the sorting for you.

If Facebook gets to sort the information you see through the course of a day (and, for a majority of the world’s adults, they do), the attention marketplace is theirs. And, if Google gets to sort the results of your retail searches, then they win big as well.

Understanding this truth about sorting explains a lot about the state of the big technology players today. Facebook – including Instagram, Whatsapp, et al – sorts the information the world sees every day. But, they’re working hard to invest in video because they don’t want Netflix and YouTube to sort what you watch.

The New York Times homepage used to be the most prestigious display of sorting two decades ago. It isn’t anymore. Old world publishers are, thus, falling by the wayside. The smarter ones are realizing that the only place they can do the sorting of information for their users is email. So, they’re doubling down on email and building subscription businesses on that (quaint?) piece of technology.

Disney, 21st Century Fox, and the like don’t want Netflix to sort the shows you want to watch. They’re scrambling. So are large telcos who’re finding ways to get in on this action.

Google, on the other hand, is targeting a different kind of foe. Google makes a lot of money off retail searches. But, increasingly, people are starting their retail searches on Amazon. Worse still, they’re able to do so by just talking to Alexa. From a monetization standpoint, that is an existential threat for Google. Google’s $550M in Chinese retailer is just the tip of the iceberg for Google’s marketplace ambitions. Expect plenty more to follow.

One thing is for certain – In all these moves and the many more to come in the future, sorting will be omnipresent.

(Also posted elsewhere as part of the “Notes by Ada” Project)

The future of automation and work

Michael Osborne, a Professor at Oxford, has a great 45 slide deck on the future of automation and work. Here are a few of the notes I’m thinking about –

“If a typical person can do a task with less than one second of thought, we can probably automate it using AI either now or the near future.” | Andrew Ng

Quantity of available work: Well defined portions of retail, service jobs, accounting, auditing, and logistics – some of the highest employed jobs – are automatable. These jobs involve physical and motor skills. On the other hand, jobs that require more cognitive skills are not. However, it is the former that accounts of high employment numbers.

Quality of available work: So, what about the fact that large sections of the population moved away from challenging, industrial work to become iOS developers, Yoga instructors, and so on? To that, Osborne has a powerful stat – it took ~60 years for the English revolution to improve the life of workers. 

The 60 year stat is powerful and important. On the one hand, it signals the importance of walking into this machine learning driven workplace paradigm with caution. We don’t need sentient AI to disrupt our lives – just take the machine learning techniques available to us today and apply it universally.

Second, there isn’t a reason we won’t be able to find a way forward and adjust to the new reality. The question is – how can we work together to do it faster?

Xerox Parc

In 1969, the Chief Scientist at Xerox – then a rich and famous photocopier company – suggested that they ought to invest in a second research center away from headquarters. They chose Palo Alto and created Xerox Parc.

In the next decade, the scientists at Xerox Parc incubated, among other things, the first graphical user interface controlled by a mouse, graphics display, and text editor.

Of course, Xerox never ended up commercializing any of this as their executives never saw the potential of these inventions. So, many of the best researchers left to places like Apple Computer and ended up inventing the future.

Xerox Parc was a one off in many ways. It was and will likely remain among the most productive research facilities in the history of mankind (Bell Labs, perhaps, would be its only real competition). But, the retelling of this famous story doesn’t fail to inspire humility.  The executives at Xerox weren’t stupid – they were just not incentivized to recognize the potential of these breakthrough ideas. That could easily be you or me.

It also never fails to remind me of the importance of investing a small but consistent portion of resources – be it at work, in our financial investments, or in life – in research/exploratory projects with no immediate pay off.  Many of these don’t seem to pay off for the longest time.

Until they do.

Don’t rush to be embarrassed by the first version of your product

The intent of good quotes is lost over time. So, they are often misunderstood and misused because they are applied out of context. Reid Hoffman’s quote – “If you are not embarrassed by the first version of your product, you’ve launched too late.” – is a great example of loss of intent.

I’ve seen this quote used as an excuse to justify a crappy v1/first version product. I haven’t heard Reid talk about this in person – but, I’m fairly certain that that wasn’t the intent.

There are two good reasons to be embarrassed about v1 (in hindsight). The first is the most common – you didn’t know better and/or couldn’t do better with the tools available. The first website I put together looked horrendous. I didn’t understand the basics of web design and it was also built on an early version of Adobe Dreamweaver. Now, however, I have slightly better design skills and, more importantly, have access to amazing tools. Thanks to the likes of Bootstrap and services like WordPress, it is very easy to build a good website.

The second is the result of prioritizing one killer use case/risky assumption for your product and ignoring everything else. You may still be embarrassed by the first version – but, you’ll still have served that basic user/customer need.

Source: Unknown – thank you to whoever made this.

The truth is that you’ll be embarrassed by nearly everything you ship. Over time, your skills will improve, the tools will get more sophisticated, and your understanding of the user/customer need will get better. So, you don’t have to work too hard to cut a few corners now to ship something you’ll be embarrassed by. Time will take care of that. The key, instead, is not to knowingly do something you will regret.

So, the two questions I’d suggest asking are –

  • Is what we are shipping helping us learn what we want to learn while providing value to the user?
  • Is this our best effort based on what we know/have access to now?

If the answers to both are yes, ship away. Even if you are eventually embarrassed about what you ship, this approach will make sure you will not have any regrets.

How we were sold tobacco, bacon and the ideal of thin women

Edward Bernays is one of the most influential persons in the 20th century. He is considered the father of “Public Relations” and changed how we think of mass marketing and advertising at scale. And, yet, it is likely you’ve never heard of him.

Despite his enduring impact on the world, there are many reasons for this lack of popularity. However, chief among them is a reluctance among the folks in his industry to talk about his work. So, you don’t hear Marketing professors or advertising executives mention him or his work. Not doing so denies some fascinating lessons that might shape how we think about the attention economy.

Edward Bernays and Propaganda
Edward Bernays was an Austrian American whose family moved to the United States in the 1890s. He spent the early part of his career as a Medical Editor and Press Agent. In both these roles, he showcased an ability to take strong positions on certain causes and successfully solicit support from the public — among them elite figures like the Rockefellers and the Roosevelts.

After the US entered World War I, he was recruited by the US Government’s “Commission on Public Information” to build support for the war domestically. Since a large portion of Americans had just fled Europe, this didn’t make much sense. But, Bernays coined a phrase — “Make the World Safe for Democracy” that became the meme President Woodrow Wilson needed. It gave the senseless war a higher purpose. And, Bernays began referring to his work as “psychological warfare.”

Bernays also added significant artillery to his propoganda techniques. He did this by incorporating the lessons from a then-infamous psychologist uncle who’d published work about how individuals are driven by unconscious needs, desires, and fears. Sigmund Freud, until then, was a relative unknown as he had been scorned by European society. But, his nephew, Edward Bernays, made him and his work famous in the United States and ensured he attained fame and prestige. Bernays applied his uncle’s insights to great effect by manipulating public opinion through mass media. As he became the world’s foremost expert in propaganda, he realized it was as powerful a tool in peacetime as it was during war.

So, after the war, he moved to New York and decided to counsel companies in propoganda. However, since the word propoganda was controversial and since its alternate “advertising” was too mundane, he decided to rename it “Public Relations.”

Why tobacco and bacon are great for you
Sadly, Bernays’ counsel was sold to anyone who cared to pay him well for it. And, chief among his clients were the tobacco companies and the pork industry. In his work with them, he demonstrated his skills as a master campaign strategist.

For example, he staged the “Torches of Freedom” event during the 1929 Easter Day Parade as a means of conflating smoking and women’s rights. Tens of millions of women threw off their shackles to claim their right to smoke in public.

His media strategy involved persuading women to smoke cigarettes instead of eating. He began by promoting the ideal of thin women by using photographers and artists in newspapers and magazines to promote their “special beauty.” He, then, had medical authorities promote the cigarettes over sweets.

Bernays also pioneered the covert use of third parties. For instance, he conviced a doctor to write to 5,000 physicians asking them to confirm that they’d recommend heavy breakfasts. 4,500 physicians wrote back and agreed. He arranged for these findings to be published in every newspaper across the country while stating that “bacon and eggs should be a central part of breakfast.” Sales of bacon went up.


The Engineering of Consent
Bernays called his brand of mass manipulation the “engineering of consent.” He worked with every major political power during his day to help provide the tools to non-coercive control of the mind. In 1928, he crystallized some of his lessons in his book “Propoganda.” Here’s a passage that describes his thought process about the importance of his work in society —

“The conscious and intelligent manipulation of the organized habits and opinions of the masses is an important element in democratic society. Those who manipulate this unseen mechanism of society constitute an invisible government which is the true ruling power of our country. …We are governed, our minds are molded, our tastes formed, our ideas suggested, largely by men we have never heard of. This is a logical result of the way in which our democratic society is organized. Vast numbers of human beings must cooperate in this manner if they are to live together as a smoothly functioning society. …In almost every act of our daily lives, whether in the sphere of politics or business, in our social conduct or our ethical thinking, we are dominated by the relatively small number of persons…who understand the mental processes and social patterns of the masses. It is they who pull the wires which control the public mind.”

Enid Blyton and Bernay’s popularity
When I was growing up in India, books from “Enid Blyton” (a British children’s authors) were recommended reading for all kids. I was a huge Enid Blyton fan myself. One of the enduring memories I have of her storytelling is her focus on food and her insistence — via various characters in the book — that “bacon and eggs was the best breakfast in the world.” As I grew up and learnt more about bacon, I couldn’t understand why she said that.

Now I do.

Edward Bernays held the masses in contempt. That’s why we don’t know much about Bernays. He simply didn’t care about popularity in the eyes of those who he held in contempt. As is evident from the Enid Blyton story, the impact of his techniques on society are undeniable. Every marketing and PR campaign since has used his techniques to shape our minds. We study his work in every case on mass marketing — without ever referring to him.

The incredible jump in the proportion of British workers who voted “leave” in the 2 months before the EU referendum would not have been possible if it wasn’t for Bernays techniques.


3 notes to ponder 

(1) Many column inches in the past year have been devoted to the role ads and social media have played in the tumultous political climate in the past couple of years. Here’s what scares me — a large proportion of the population is responding by consuming and discussing events on private messaging tools like Whatsapp. The Reuters Insititute confirms this trend among younger Americans. If you think polarized social news feeds can be propoganda carriers, you haven’t experienced the power of Whatsapp in spreading lies (see this and this).

(2) How would we go about learning marketing and public relations if we studied the life and work of Edward Bernays? I understand why professors and executives don’t want to talk about Bernays. Discussing his beliefs and techniques can seem akin to touting the power of the dark arts. But, every useful tool has its dark sides. And, the founding story of the PR industry is a great example of that. It is not an example we should avoid. Instead, it is a story we must learn from. It’ll make us all better marketers and, perhaps, better human beings.

(3) I was one among many who was surprised by the size of the impact of social media tools on the global politics. In retrospect, there were many warning signs. But, hindsight is always 20:20. That said, I don’t think I’d have been anywhere as surprised if I’d read the Edward Bernays story. The story of the propoganda maestro from a hundred years ago, it turns out, is very relevant today.

History doesn’t repeat itself — but it rhymes. It is why any attempt to understand the present and predict the future is futile if it isn’t preceded by an understanding of history.

Links for additional reading

Invisible asymptotes

Eugene Wei, who used to run Product at Hulu and Flipboard, had a fantastic post out on Invisible Asymptotes. An invisible asymptote is the ceiling of the growth curve if we proceeded down the certain path.

For example, Amazon’s invisible asymptote (and that of most e-commerce businesses) in the early days was shipping. People hated shipping fees and bought considerably more once they were on Amazon Prime. While such insights are obvious in retrospect, these asymptotes aren’t easy to identify. And, to that end, he offers two thought provoking insights.

The first is that customers are excellent at telling us what they don’t want or don’t like. Product managers spend a lot of time optimizing their funnels and learning more about who reaches the bottom. This is great in the early days as survival depends on strong product market fit with one group. However, as a company grows, we identify our invisibly asymptotes by understanding who falls out at the top of the funnel. That’s how we expand our offering.

The second is about around how he finds successful people to be much more conscious of their own personal asymptotes at a much earlier age than others. Somebody he knew determined in grade school that that she’d never be a world-class tennis player or pianist. Another knew a year into a job that he wouldn’t be the best programmer at his company and so he switched over into management; he rose to become CEO.

His final two paragraphs brings both these takeaways together beautifully –

By discovering their own limitations early, they are also quicker to discover vectors on which they’re personally unbounded. Product development will always be a multi-dimensional problem, often frustratingly so, but the value of reducing that dimensionality often costs so little that it should be more widely employed.

This isn’t to say a person needs to aspire to be the best at everything they do. I’m at peace with the fact that I’ll likely always be a middling cook, that I won’t win the Tour de France, and that I’m destined to be behind a camera and not in front of it. When it comes to business, however, and surviving in the ruthless Hobbesian jungle, where much more is winner-take-all than it once was, the idea that you can be whatever you want to be, or build whatever you want to build, is a sure path to a short, unhappy existence.