I think generalizations are tough in either direction. I'm sure plenty of businesses required many smart decisions and great execution. Others not so much. I know plenty of very large and insanely profitable asset management businesses that essentially made a few good decisions early on (indistinguishable from luck given small sample size) and milked the track record ever since. Just one example.
I don't know enough to really analyze the Amazon situation, but I heartily agree that economists know staggeringly little about how businesses run. It was shocking to me, going to grad school after working for a decade, how blithely professional economists waved off any considerations about how the markets they claimed to understand actually function. Working in supply chain makes it even worse, as very few people have a good grasp of the problems there, and probably near zero academic economists.
All in all, economists have a strange lack of interest in how businesses operate, it seems to me.
The investment firm I founded has made over 1400 small investments in incredibly early stage startups (2-3 rounds before the traditional VCs that most people think of when they hear the term "VC").
We then go to an incredible amount of trouble to track what happens at these companies, though even that effort results in only very partial information.
Based on our observations, I think your conclusion is correct in that the idea of an inevitable niche is wrong. But I think your analogy to Go, and typical games of skill in general, is also wrong.
There are so many random events that affect the success of a startup that, while a skilled entrepreneur will do better than an unskilled one on average, the achievement of generational "unicorn" status like Amazon is mostly random. If Jeff Bezos started Amazon in 100 universes, this is very likely the only one where it becomes the dominant force it is today. (Though Amazon might be successful in the ordinary sense of making early investors 10X+ returns in a large proportion of universes).
Imagine if you played Go and fairly frequently, a bunch of pieces would randomly flip. Then occasionally the very rules on what cause them to flip changed. That volatility will make it hard for even the most skilled player to consistently win in a dominant fashion.
you are explaining why the winner who builds a company in a new space must be consistently fierce, brilliant, exceptional.
Scott's argument is that as technology develops, many inventions are made by multiple people around the same time (Kevin Kelly argues something similar). So selling things over the internet is one of those class of ideas. Kevin Kelly argues that the brilliant entrpenaur can advance things maybe 5 years earlier than would've happened anyway. So for example without Steve Jobs launching the iPhone in 2007, we would've gotten some kind of rectangular touch screen phone in 2012.
my point here is your post doesn't disagree with Scott's, you are talking about related but different things. I think the Kevin Kelly frame here is how they should tie together. How many years can brilliant execution pull in a technology? In that sense, yes, Scott is correct and we'd have ecommernce in 2022 even without Besos. If maybe on a 5 year time delay, and slightly different set of companies providing it. For example AWS likely would've happened differently and by not the same company that pioneered ecommerce.
Search is another things which became inevitable. There were a half dozen internet search companies in the late '90s: Yahoo, AltaVista, others I can't remember. They all worked, they all let you get stuff done. Then Google came along with a _much_ better implementation and cleaned everyone's clocks.
So I'd say it was inevitable _someone_ would fill that niche. What's not clear is _how well_ that niche would be filled. It's not inevitable we'd have products which work as well as iPhones, Google, and Amazon. That's part of why Tesla is notable: Teslas work much better in some ways than products produced by the incumbent companies.
Similarly, if Amazon didn't come around, I'm quite sure Walmart would have eventually produced an e-store. It just probably wouldn't have worked as well as Amazon. TBH, what amazes me about Amazon was that they were able to successfully duplicate and surpass Walmart's overwhelming logistical head start.
yes. Kling is tacitly making this point as well. The business landscape and detail implementations details are highly contingent on which new companies and leaders succeed, and would differ quite a bit. I could easily see an alternate world in mobile phones where there was a single monopoly instead of the Andriod/iOS duopoly we have, and that world would be locked into a worse off system. But we'd still have rectangular touch screens slabs for our phones in that alt world.
Regarding your point about academics not truly understanding the difficulty of building a successful business, an analogy to a construct for understanding war is helpful.
War is understood to be fought on three basic levels:
1. the strategic, which is the high-level aims and goals;
2. the tactical, which is the actual fighting; and
3. the operational: which are the means by which the strategic and tactical levels connect. That is, the logistics—the ability to produce tanks or aircraft and deliver them to the front line.
Academics generally understand the strategic level of what business is attempting to achieve. However, many are deficient in their understanding of the tactical and operational demands of that achievement.
Exactly my thoughts. There is the old saying "Amateurs study tactics; professionals study logistics", and having worked in supply chain/logistics for well over a decade, I can tell you that many professional businessmen need to study it a lot more. Academic economists need to study it, period. They haven't even surpassed the amateur mark.
Arnold's excellent and far too little understood or discussed point about 75% optimal vs 65% optimal is easier to visualize with Othello than with the more uncertain poker, but in both the fact of a "small" difference becomes more important over multiple decisions.
Scott does seem weak on real business, as do most economic professors.
Very often, with two people "of equal skill/ talent", the one which puts in more total efficient hours of practice/ analysis will do better. Which is why workaholics are more often successful - and most successful small business that grow were founded by a workaholic.
Consistently making optimal decisions is really hard. Especially for lower IQ folk.
Successful religions & "morality" are especially good at helping the lower IQ folk make fewer and less sub-optimal decisions, more often.
The "do what you love" truth is that, if you're a workaholic "loving" the job you're doing, it's not such a bad work-life balance. I don't think this is so true for most middle managers nor gov't workers tho.
-- “taxing billionaires a lot” and “taxing billionaires not at all” are at least two different unfair failure modes with their own advantages and disadvantages from a desert point of view.
I'm now moving towards a lot more taxing of successful billionaires to pay for a stable society, like Ukraine is NOT.
I'm also thinking of capping the value of IP - patents and copyrights both, so as to increase consumer surplus without much disincentive to create. Does Rowling really write more Harry Potter to get a second billion, where she wouldn't if she paid more tax on it? Or Disney with Mickey Mouse? It's the gov't which supports the IP monopoly, so I'm increasingly OK with those who get IP benefits paying more for gov't.
I'm also in favor of less direct gov't spending, so the tax burden is lower, something Trump half did, with great tax cut & jobs results, but is not credited with.
Walmart's Internet retail service is competitive with Amazon. I use both. They both can do same-day delivery, prices are generally about the same. Amazon is a little better, their website is a little nicer, their selection is a little broader, but Walmart is very close.
Amazon has branched into other businesses: AWS is a giant one. I was going to write how I was excited about Amazon getting into health care, with amazon.care, but within the past few days, Amazon officially announced they are shutting that down, so I guess that was one of their failed attempts.
On the one hand, you're right: Amazon as we know it was not inevitable. I would have thought it much more likely that someone would have partnered with Walmart, Sears, Target, and Costco to come up with something more like DoorDash. At least that would have been my (probably failed and long forgotten) approach.
But what was likely inevitable was that thousands of people would have noticed the need and tried to address it. By definition, _someone_ would have guessed right more than the others and produced the company we all use and love. It might have been Amazon, it might have been Walmart, or it might have been my ur-DoorDash. In hindsight, it might look obvious why the winner won. That would not have been clear at the time, just like it wasn't at all obvious iPhones would win over Nokia and Blackberry.
Written (and I'm referring to Lenin) like someone who never actually managed a business.
I'm pretty smart and have worked in companies for 35+ years. I really, really doubt I could run a successful company. I don't have the vision, organizational ability, leadership, or any of the other talents a founder or CEO needs.
He has a tiny bit of a point: once a stable business is running, perhaps it can survive on inertia and rote. But no business is stable for long. Societies are just too chaotic. I think that lack of understanding the chaos and dynamism of societies is why Lenin and economy planning fail.
The main reason that Amazon had to clear so many hurdles was because the market was so competitive that even slight inefficiencies would be punished. To say a market is competitive is to say that there are other companies offering similar goods or services that were ready to replace Amazon. In 1892, the world-champion mail order company (Sears Roebuck) was successful despite being very primitive by our standards; in the 21st century Sears offers a much better retail experience but is only marginally competitive in the retail market.
Your view is rather like saying that since becoming an NBA superstar is incredibly difficult, if Kobe Bryant had become a dentist we wouldn't have any basketball games.
Reading Scott's piece after yours, I think there are two more points you could have made:
1: Not only did Amazon find a niche, but they out competed other companies within that niche over time. Then they expanded to other niches, and kept out competing. Scott's "intuition pump" runs the opposite way he imagines.
2: Scott implicitly assumes that no one has a right to their property, and thus has to justify everything they own in order to be allowed to keep it. If we do not make that assumption, the burden of proof falls to prove that someone acquired their property unjustly before we are justified in redistribution. That is not an easy burden to bear in Amazon's case (though one could, no one has to my knowledge.)
When I was working in the optimization/quantum computing area a while back I talked to a person in the logistics consulting business. We got to talking about Amazon; they said Amazon succeeded by doing better/different math than anyone else. Everyone else in the industry was trying minimize costs by building a few giant distribution centers across the country. Bezos/Amazon looked at it differently and decided they could deliver items much faster, the same day in many cases if they put 300+ distribution centers all over the country. This off course meant tackling and solving a host of very complex problems.
Anyway, that was the view of some working in the field.
I think a steelmanned version of Alexander's argument can still go through. Suppose Jeff Bezos had to do two things to build Amazon:
1. Make a bunch of great business decisions and work hard to implement them in a way that only .001% of the population could or would have done.
2. Among that .001% of the population, be the one in the right place at the right time.
(1) gives him due credit for the extreme difficulty of what he did. But (2) means that there are a few thousand other Americans who could have done what he did had they been in the right place at the right time. Scott's central argument is that if Bezos hadn't been there, one of those other few thousand would have stepped in and made something Amazon-like instead.
So the question is whether Bezos is "only" a one-in-100K talent or is more like a one-in-100M talent. I honestly don't know, but the former isn't intuitively implausible given what an outlier a one-in-100K talent already is.
To start with an online bookstore and end up with Amazon Web Services--that is not a predetermined path waiting for someone to find it. The combination of business decisions that created Amazon was unique to Bezos. Yes, someone other great business leader could have built a dominant online retailer, but it would be shaped differently and might not be worth nearly as much.
Many great companies have ended up in very different places from where they started. It's a mistake to gape at the unique path Amazon took. MANY odd-seeming routes would have ended up at something much like Amazon Web Services.
Unfortunately, this is all anecdata, and many people will offer opposing anecdotes of equal strength (concerning Meg Whitman, for example). I think the right attitude to take here is one of uncertainty and humility. There's no good way to figure out how representative Arnold's experiences were. So we should be pretty uncertain about how deserving business leaders are--which is not to say we should be uncertain about the benefits of markets to society as a whole.
I think generalizations are tough in either direction. I'm sure plenty of businesses required many smart decisions and great execution. Others not so much. I know plenty of very large and insanely profitable asset management businesses that essentially made a few good decisions early on (indistinguishable from luck given small sample size) and milked the track record ever since. Just one example.
I don't know enough to really analyze the Amazon situation, but I heartily agree that economists know staggeringly little about how businesses run. It was shocking to me, going to grad school after working for a decade, how blithely professional economists waved off any considerations about how the markets they claimed to understand actually function. Working in supply chain makes it even worse, as very few people have a good grasp of the problems there, and probably near zero academic economists.
All in all, economists have a strange lack of interest in how businesses operate, it seems to me.
It's extremely easy to see everything in the rear view mirror as inevitable.
The investment firm I founded has made over 1400 small investments in incredibly early stage startups (2-3 rounds before the traditional VCs that most people think of when they hear the term "VC").
We then go to an incredible amount of trouble to track what happens at these companies, though even that effort results in only very partial information.
Based on our observations, I think your conclusion is correct in that the idea of an inevitable niche is wrong. But I think your analogy to Go, and typical games of skill in general, is also wrong.
There are so many random events that affect the success of a startup that, while a skilled entrepreneur will do better than an unskilled one on average, the achievement of generational "unicorn" status like Amazon is mostly random. If Jeff Bezos started Amazon in 100 universes, this is very likely the only one where it becomes the dominant force it is today. (Though Amazon might be successful in the ordinary sense of making early investors 10X+ returns in a large proportion of universes).
Imagine if you played Go and fairly frequently, a bunch of pieces would randomly flip. Then occasionally the very rules on what cause them to flip changed. That volatility will make it hard for even the most skilled player to consistently win in a dominant fashion.
you are explaining why the winner who builds a company in a new space must be consistently fierce, brilliant, exceptional.
Scott's argument is that as technology develops, many inventions are made by multiple people around the same time (Kevin Kelly argues something similar). So selling things over the internet is one of those class of ideas. Kevin Kelly argues that the brilliant entrpenaur can advance things maybe 5 years earlier than would've happened anyway. So for example without Steve Jobs launching the iPhone in 2007, we would've gotten some kind of rectangular touch screen phone in 2012.
my point here is your post doesn't disagree with Scott's, you are talking about related but different things. I think the Kevin Kelly frame here is how they should tie together. How many years can brilliant execution pull in a technology? In that sense, yes, Scott is correct and we'd have ecommernce in 2022 even without Besos. If maybe on a 5 year time delay, and slightly different set of companies providing it. For example AWS likely would've happened differently and by not the same company that pioneered ecommerce.
Search is another things which became inevitable. There were a half dozen internet search companies in the late '90s: Yahoo, AltaVista, others I can't remember. They all worked, they all let you get stuff done. Then Google came along with a _much_ better implementation and cleaned everyone's clocks.
So I'd say it was inevitable _someone_ would fill that niche. What's not clear is _how well_ that niche would be filled. It's not inevitable we'd have products which work as well as iPhones, Google, and Amazon. That's part of why Tesla is notable: Teslas work much better in some ways than products produced by the incumbent companies.
Similarly, if Amazon didn't come around, I'm quite sure Walmart would have eventually produced an e-store. It just probably wouldn't have worked as well as Amazon. TBH, what amazes me about Amazon was that they were able to successfully duplicate and surpass Walmart's overwhelming logistical head start.
yes. Kling is tacitly making this point as well. The business landscape and detail implementations details are highly contingent on which new companies and leaders succeed, and would differ quite a bit. I could easily see an alternate world in mobile phones where there was a single monopoly instead of the Andriod/iOS duopoly we have, and that world would be locked into a worse off system. But we'd still have rectangular touch screens slabs for our phones in that alt world.
Regarding your point about academics not truly understanding the difficulty of building a successful business, an analogy to a construct for understanding war is helpful.
War is understood to be fought on three basic levels:
1. the strategic, which is the high-level aims and goals;
2. the tactical, which is the actual fighting; and
3. the operational: which are the means by which the strategic and tactical levels connect. That is, the logistics—the ability to produce tanks or aircraft and deliver them to the front line.
Academics generally understand the strategic level of what business is attempting to achieve. However, many are deficient in their understanding of the tactical and operational demands of that achievement.
Exactly my thoughts. There is the old saying "Amateurs study tactics; professionals study logistics", and having worked in supply chain/logistics for well over a decade, I can tell you that many professional businessmen need to study it a lot more. Academic economists need to study it, period. They haven't even surpassed the amateur mark.
Arnold's excellent and far too little understood or discussed point about 75% optimal vs 65% optimal is easier to visualize with Othello than with the more uncertain poker, but in both the fact of a "small" difference becomes more important over multiple decisions.
Scott does seem weak on real business, as do most economic professors.
Very often, with two people "of equal skill/ talent", the one which puts in more total efficient hours of practice/ analysis will do better. Which is why workaholics are more often successful - and most successful small business that grow were founded by a workaholic.
Consistently making optimal decisions is really hard. Especially for lower IQ folk.
Successful religions & "morality" are especially good at helping the lower IQ folk make fewer and less sub-optimal decisions, more often.
The "do what you love" truth is that, if you're a workaholic "loving" the job you're doing, it's not such a bad work-life balance. I don't think this is so true for most middle managers nor gov't workers tho.
Scott ends on taxation:
-- “taxing billionaires a lot” and “taxing billionaires not at all” are at least two different unfair failure modes with their own advantages and disadvantages from a desert point of view.
I'm now moving towards a lot more taxing of successful billionaires to pay for a stable society, like Ukraine is NOT.
I'm also thinking of capping the value of IP - patents and copyrights both, so as to increase consumer surplus without much disincentive to create. Does Rowling really write more Harry Potter to get a second billion, where she wouldn't if she paid more tax on it? Or Disney with Mickey Mouse? It's the gov't which supports the IP monopoly, so I'm increasingly OK with those who get IP benefits paying more for gov't.
I'm also in favor of less direct gov't spending, so the tax burden is lower, something Trump half did, with great tax cut & jobs results, but is not credited with.
as I like to say...If you have never signed the FRONT of a paycheck, you'll never truly understand business...much less how to run one.
Walmart's Internet retail service is competitive with Amazon. I use both. They both can do same-day delivery, prices are generally about the same. Amazon is a little better, their website is a little nicer, their selection is a little broader, but Walmart is very close.
Amazon has branched into other businesses: AWS is a giant one. I was going to write how I was excited about Amazon getting into health care, with amazon.care, but within the past few days, Amazon officially announced they are shutting that down, so I guess that was one of their failed attempts.
On the one hand, you're right: Amazon as we know it was not inevitable. I would have thought it much more likely that someone would have partnered with Walmart, Sears, Target, and Costco to come up with something more like DoorDash. At least that would have been my (probably failed and long forgotten) approach.
But what was likely inevitable was that thousands of people would have noticed the need and tried to address it. By definition, _someone_ would have guessed right more than the others and produced the company we all use and love. It might have been Amazon, it might have been Walmart, or it might have been my ur-DoorDash. In hindsight, it might look obvious why the winner won. That would not have been clear at the time, just like it wasn't at all obvious iPhones would win over Nokia and Blackberry.
Lenin wrote that business is easy. Any clerk can do it. But no one would say that about a football coach or a popular singer.
Written (and I'm referring to Lenin) like someone who never actually managed a business.
I'm pretty smart and have worked in companies for 35+ years. I really, really doubt I could run a successful company. I don't have the vision, organizational ability, leadership, or any of the other talents a founder or CEO needs.
He has a tiny bit of a point: once a stable business is running, perhaps it can survive on inertia and rote. But no business is stable for long. Societies are just too chaotic. I think that lack of understanding the chaos and dynamism of societies is why Lenin and economy planning fail.
The main reason that Amazon had to clear so many hurdles was because the market was so competitive that even slight inefficiencies would be punished. To say a market is competitive is to say that there are other companies offering similar goods or services that were ready to replace Amazon. In 1892, the world-champion mail order company (Sears Roebuck) was successful despite being very primitive by our standards; in the 21st century Sears offers a much better retail experience but is only marginally competitive in the retail market.
Your view is rather like saying that since becoming an NBA superstar is incredibly difficult, if Kobe Bryant had become a dentist we wouldn't have any basketball games.
Reading Scott's piece after yours, I think there are two more points you could have made:
1: Not only did Amazon find a niche, but they out competed other companies within that niche over time. Then they expanded to other niches, and kept out competing. Scott's "intuition pump" runs the opposite way he imagines.
2: Scott implicitly assumes that no one has a right to their property, and thus has to justify everything they own in order to be allowed to keep it. If we do not make that assumption, the burden of proof falls to prove that someone acquired their property unjustly before we are justified in redistribution. That is not an easy burden to bear in Amazon's case (though one could, no one has to my knowledge.)
Link to the full essay here https://dochammer.substack.com/p/scott-alexander-is-wrong-about-business
When I was working in the optimization/quantum computing area a while back I talked to a person in the logistics consulting business. We got to talking about Amazon; they said Amazon succeeded by doing better/different math than anyone else. Everyone else in the industry was trying minimize costs by building a few giant distribution centers across the country. Bezos/Amazon looked at it differently and decided they could deliver items much faster, the same day in many cases if they put 300+ distribution centers all over the country. This off course meant tackling and solving a host of very complex problems.
Anyway, that was the view of some working in the field.
I think a steelmanned version of Alexander's argument can still go through. Suppose Jeff Bezos had to do two things to build Amazon:
1. Make a bunch of great business decisions and work hard to implement them in a way that only .001% of the population could or would have done.
2. Among that .001% of the population, be the one in the right place at the right time.
(1) gives him due credit for the extreme difficulty of what he did. But (2) means that there are a few thousand other Americans who could have done what he did had they been in the right place at the right time. Scott's central argument is that if Bezos hadn't been there, one of those other few thousand would have stepped in and made something Amazon-like instead.
So the question is whether Bezos is "only" a one-in-100K talent or is more like a one-in-100M talent. I honestly don't know, but the former isn't intuitively implausible given what an outlier a one-in-100K talent already is.
To start with an online bookstore and end up with Amazon Web Services--that is not a predetermined path waiting for someone to find it. The combination of business decisions that created Amazon was unique to Bezos. Yes, someone other great business leader could have built a dominant online retailer, but it would be shaped differently and might not be worth nearly as much.
Many great companies have ended up in very different places from where they started. It's a mistake to gape at the unique path Amazon took. MANY odd-seeming routes would have ended up at something much like Amazon Web Services.
Unfortunately, this is all anecdata, and many people will offer opposing anecdotes of equal strength (concerning Meg Whitman, for example). I think the right attitude to take here is one of uncertainty and humility. There's no good way to figure out how representative Arnold's experiences were. So we should be pretty uncertain about how deserving business leaders are--which is not to say we should be uncertain about the benefits of markets to society as a whole.