How Amazon Reshapes Small Business to Serve Itself

Moira Weigel


Paris Marx is joined by Moira Weigel to discuss the third-party sellers who supply many of the goods sold through Amazon, how the company’s policy decisions reshape small businesses to act like mini-Amazons, and what that means for regulatory responses.


Moira Weigel is an Assistant Professor at Northeastern University, a Faculty Associate at the Berkman Klein Center at Harvard Law School, and a founding editor of Logic Magazine. Her most recent book is Voices from the Valley: Tech Workers Talk about What They Do--And How They Do It, co-edited with Ben Tarnoff. Follow Moira on Twitter at @moiragweigel.

Support the show

Venture capitalists aren’t funding critical analysis of the tech industry — that’s why the show relies on listener support.

Become a supporter on Patreon to ensure the show can keep promoting critical tech perspectives. That will also get you access to the Discord chat, a shoutout on the show, some stickers, and more!



Paris Marx: Moira, welcome back to Tech Won’t Save Us!

Moira Weigel: Thanks so much for having me again!

PM: Very excited to chat with you again. Always great to get your insights on the tech industry and all these things anyway. But you have this new report out looking at Amazon, and in particular, its relationship to third party sellers and how its platform, how its policies shape the types of businesses that take place on this platform. I think this is really important as we try to understand the impacts that Amazon is having on the world, how it’s shaping the economy more broadly. I want to dig into all these things with you.

I want to start by asking you a broader question. When we talk about Amazon, especially on this podcast, we often focus on delivery workers and warehouse workers and what that relationship to Amazon tells us about the company. In the report, as I said, you explore the relationship to third party sellers. Why do you think that’s an important way to look at Amazon?

MW: Well, I should say loud and clear that it’s extremely important to focus on workers in the fulfillment centers and delivery and logistics workers as well. I certainly intend my approach to be a complement to the excellent work by various people doing that and in the spirit of being a good feminist scholar, my colleagues at Data and Society Aiha Nguyen and Eve Zelickson, put out a fabulous report on how the Ring doorbell cameras that Amazon sells are used to monitor delivery workers, and the racialized implications of that, thinking about that as a form of racialized labor surveillance. There’s a new book out that I’m very excited to read by a scholar Sarrah Kassem on organizing at Amazon. I believe it’s a comparative study based in a couple countries in Europe.

All that to say — lots of important things to look at with Amazon. I got to focusing on third party sellers through a circuitous path. So I hadn’t studied retail in depth before starting this research a couple of years ago, at this point. I first got interested in Amazon in 2018, or 2019. I did this collaborative project on Amazon book publishing with Ava Kofman and Francis Tseng, and we were looking at right-wing and white nationalist publishing on Amazon, which is a big thing. Our study or piece on that came out in April 2020, so we’d been working on it in months leading up to that. So we didn’t really look at COVID issues. But since COVID, Amazon has been a huge platform for people publishing false information about the disease and about vaccines.

All this to say, I got into looking at Amazon through an interest in something that seems very different, maybe, than what I’m doing now, which was conservative and right-wing media production. I don’t really like the term disinformation, but since that’s what people use, I’ll just say disinformation. In the course of doing that research, I got really fascinated by Amazon as a platform. It sounds silly, but, still to this day, there’s new work coming out on Amazon. But I think Google and Twitter and Facebook, and basically search and social media platforms, for the most part, have been much more extensively studied by scholars.

Then as I was saying, there’s been a lot of great work from a labor perspective on people who work in Amazon warehouses and logistics, but not that much study — and hardly any study I came to find — of Amazon as a platform and what it meant to ‘platformize’ retail. As with any project, my methods and research questions evolved over a period of time. When I first thought: Okay, I want to really understand how Amazon retail works because it doesn’t sound glamorous, I thin. We like to think of the cloud or we think of logistics, but retail is the core of Amazon’s business and always has been in a number of ways.

At first, I thought I would take an approach that I’ve taken in the past. In the book of oral histories that I did with Ben Tarnoff, “Voices from the Valley” and then these anonymous interviews that I did over the years for Logic. I thought: Okay, great. I’ll find people who work in retail at Amazon; I’ll find software engineers, or folks inside HQ in Seattle, basically, who will talk to me about how this platform works from the inside. I did a few interviews like that, we published a long interview I did with an AWS enginee in Logic a couple years ago, which made people angry in a way that delighted me at the time. Because I was like: he’s supposed to make you angry; he’s supposed to be a jerk.
Anyway, I very quickly realized that this method wasn’t going to get me what I wanted to understand. That was for a few reasons. First, of course, Amazon’s full-time employees or NDA’d. It’s hard to find ones who can talk to me. It’d be very hard to get a random sample of ones who could talk to me and do academic research on them. But more importantly, engineers have a very specific type of knowledge of a platform like Amazon. I interviewed one or two folks who’d worked on one really particular widget, like the AI system that tells the Amazon algorithm which search results to display in which country based on that country’s product safety laws, which might be different, or something like that.

Engineers I spoke to were really brilliant about explaining whatever widget they’d happen to work on. But the nature of Amazon’s third party marketplace and its retail business is so global, and so sprawling, that I started to think that I needed to talk to the people who actually use it and navigate it to get a more global perspective. This, of course, was politically appealing to me to think: How can we tell a story of this global platform that is not about a small handful of mostly white and perhaps East Asian, mostly male engineers in Seattle, but rather about the single mom buying stuff in Targets in North Carolina to resell on Amazon and the hustler from the country who’s moved to Shenzhen white-labeling goods like tomahawks, or whatever. That’s another story.

How can we tell a story in which these folks are the protagonists of Amazon’s rise, as they really are in a lot of meaningful ways? But the third party sellers, the fees they pay to use Amazon, have been among its largest sources of revenue, if not its largest source of revenue for years. They provide the products that stock the catalog that makes The Everything Store. Amazon keeps them mostly invisible — that’s by design. But in any case, I thought: How do I tell the story of this global platform in which these folks are the protagonists and that recognizes that in many ways, that single mom who has to figure out several times a week in North Carolina whether flip flops or lawn flamingos that she’s able to source from Target and Best Buy will sell for more on Amazon than they’re selling for there. Or the hustler in Shenzhen, who’s white-labeling ammo bags and tomahawks for the US market.

Those actors really, in many ways have a much more comprehensive view of the global system that Amazon is, then does the engineer who’s building the one piece of software to say: Okay, this baby product will show up in Turkey, but not in Germany because they have different product laws, or whatever it might be. I became very drawn to these third party figures. Third party sellers sell more than half of everything bought through Fees that they pay to use Amazon’s physical and virtual infrastructure, have in many years Amazon’s highest source of revenue for years, considerably higher than AWS.

AWS is only just beginning to catch up and they go up all the time. The work they do sourcing and finding and advertising products and the risk that they bear that work and that risk bearing has really made Amazon the global behemoth that it is. They’re almost invisible the way Amazon is designed. It’s designed to be invisible in many ways. So I was really interested in their perspective as a way to try to tell the story of this global platform, not quite from below, but from in the middle of it or from the edges. I think they have a perspective that lets us refract an account of how Amazon has grown and developed.

PM: I think it’s really fascinating to hear what you described with that because there was a report recently from Marketplace Pulse that Amazon’s fees have now exceeded 50% of the average price of what is being sold on the platform. That means that it’s taking a significant chunk of the sales. As you say, when we think about Amazon we think about Jeff Bezos, maybe we think about Andy Jassy and some of these other executives and engineers in places like Seattle. We don’t think about these third party sellers who are supplying a lot of what actually goes on to the platform, what’s actually sold through the platform. It’s designed in such a way often to abstract when you’re buying from third party sellers.

I think that these are important things to understand. I want to explore various periods of how this has evolved, how the platform has evolved, and how that has changed how things work for the sellers, and how they have seen things change over that time. But first, I wanted to ask you about the particular names for these periods because these are names that come from the sellers, and they termed: the “Old times,” the “Wild West,” the “Jungle,” and “The Rocket Ship.” As I said, I want to dig into each one separately. But before we do that, I want you to speak to us first about those terms, and what they say about sellers perceptions about Amazon itself as this major global platform.

MW: When I set out into this research project, like anyone at the beginning of a research project, I didn’t know exactly what I was going to find. I found quickly that the world of e-commerce, to me anyways, so complex and so bewildering. I think not just to me; a lot of people talk about this. But it took me a really long time, relatively speaking, to just understand what was going on and to learn all this language that participants in the marketplace use to describe different features of it and different aspects of their experience. Over the course of conducting dozens and dozens of in depth interviews and lots of more informal conversations with people in this world.

It started to strike me that I was hearing, among the English speakers, a couple of phrases again and again. As you just said, these phrases were the “Old Times,” which was a phrase I heard from the relatively small number of people I met who had been active on marketplace before or around 2010. This phrase, the “Wild West,” which folks use to describe the era from about 2010 to the mid-2010s. I could say a lot more about that. I know the Wild West is a phrase that gets thrown around a lot in the tech industry. Then most strikingly, to me, this phrase, the “Jungle” that multiple people used to refer to the era that we have been in since about the mid-2010s to the present.

With a small footnote, I’d say that COVID sort of created its own era, and sometimes people call that the “Rocket Ship.” Although, now it seems maybe the rocket ship is crashing or coming to ground or something. I could say more about each of these phrases and periods and also what led me to take them seriously, instead of inwardly roll my eyes at the suggestion of a tech Wild West. But just to make a metapoint — over time, as I noticed people referring to these periods, and I noticed the fact that widely dispersed individuals who don’t have very much in common with each other — demographically or regionally or in terms of what they were selling, or anything like that — all seemed to share a sense of when these turning points or breaking points between the eras had happened. I came to believe that the fact that there are these abrupt turning points is itself a salient finding.

That everyone, again, from the single mom in North Carolina and to the very successful furniture business owner in the Midwest, a family man — sort of pater familias — to this family business to consultants in Shenzhen who tell Chinese merchants how to spin up Amazon businesses. The fact that they all, or many of them, agreed when these turning points happen, pointed to the fact these changes in Amazon’s policy and also the algorithmic implementation of its policies, created these really abrupt breaks in the livelihoods of the people on the platform. I came to think that that itself was a key finding and pointed to something important about the specific dependency and vulnerability that third party sellers experience, even when they do also sometimes have great opportunities. That ambivalence is is also part of what interests me about their position.

Maybe I just should say as a metapoint — I was working through the writing up of all this data that I gathered, I had come to think that these metaphors were important and that the periodization that they reflected, were important. I was talking about it a bunch with my editor on this report and friend, Patrick Davison asked me a question that really stuck with me and we were trying to tease out the metaphor. We were like: If it’s the Wild West or if it’s the “Jungle?” What did the sellers mean by that? Like, where is it? Patrick asked me a question that at the time, briefly, stumped me. He’s like: Well, if it’s the Wild West, is Amazon the land? Is it the land grab? Or is Amazon, like the United States, sending people out to grab the land?

I thought about that, going through the interview transcripts, and came to think it was actually both — that people I spoke to were imagining Amazon’s marketplace itself, which was growing rapidly in terms of its customer base as a land rush, or land grab, where they wanted to seize their opportunity. But their use of these metaphors also reflected this sense that, to them, Amazon was like a crown or a state that was sending them out to conquer new territory on its behalf.

I came to think that the fact that it was both was, actually, not a mistake, but pointed to something quite interesting about the ambivalence of the seller position where they’re both trying to seize opportunities that are opened up by this global — one might say, monopolistic — retail platform. They are also simultaneously, highly vulnerable to the opaque and non-democratic governing power of Amazon itself. Again, I could say a lot more about the colonial extractive imaginaries of these phrases, but I thought that that contradiction, actually pointed to something interesting about the seller position.

PM: Absolutely. At the same time, they see the opportunity of Amazon also being one where, especially if they were selling a particular product, things can take off really quickly,. There’s a an opportunity there that you might not have in a traditional economy or something like that, that comes with entering into one of these platforms. I want to talk about some of these periods. The first one being the “Old Times,” which you define as being pre-2010. What defines Amazon in this period and how do third party sellers start getting access to this platform and growing on this platform in this time?

MW: Amazon is founded in the mid 90s and Bezos has this idea of wanting to become somehow the intermediary for e-commerce. I always tell this anecdote, but I love to tell it as someone who writes books because it keeps us humble. He settled on books as a product category because the distributors are concentrated in the Pacific Northwest, three of the four national distributors are; books are mostly about the same size; there’s a limited number of boxes you have to buy to ship them and they mostly aren’t gonna break in transit. Of course, he was interested in being in the Pacific Northwest because at that time, if he was HQ’d in Washington, they could not charge sales tax to most of the people they ship things to.

Anyway, Bezos starts out in the 90s, coming from D. E. Shaw with this vision of becoming an intermediary for online commerce. At first they start with books and start a more traditional, so called first party business. In commerce, in US law, a first party seller is someone who sells directly to a customer. If I wanted to sell my hat that I’m holding here to you, Paris, we could say I’m the first party, and you’re the second party to the transaction, you buy it directly from me. When I hand the hat to you, and you give me money, it becomes your property and then you can do with it what you like. A third party in commerce refers to a situation where business is transacted and the entity who parents buys from, in this example, never actually holds title to the product.

I had my hat and then I handed it to my charming young child who is going to go sell it because she’s cuter than I am and more appealing to buy a hat from and Paris bought it from her. You could say that I’m the third party. Paris, you buy the hat from her, but actually, she never own the hat and she has to get the money right back to me. This is a very clumsy embarrassment. Just to make concrete what it means to be a third party seller is that you source goods that you sell through Amazon, without Amazon ever buying them from you, and what it means for a consumer — and mabye this is a neater way to put it is that — consumers mostly buy through Amazon, they buy from a third party seller through Amazon, rather than buying from Amazon. People do also buy from Amazon, which is part of what makes it confusing. I often say Amazon grafted an eBay onto a Walmart while making it look like a Walmart online.

Anyway, so Bezos had this idea in the 90s of wanting to be an e-commerce intermediary and starts by selling books first party. Amazon sells it themselves and then by around 2000 start opening up what they call the third party marketplace, starting with used books and CDs and then other product categories letting third parties sell through them. This begins the era that my interviewees called the “Old Times.” And in that era, Amazon was effectively an online catalog for them. How it worked is that Amazon would recruit merchants to sell through it, or maybe merchants would find Amazon on their own. This meant that when a customer went to Amazon and placed an order, Amazon would email — or sometimes I even heard fax — the name and address of that customer to the merchant and then it was the merchant’s problem, or the merchant would then go about sending those goods to the customer on their own.

In this early era, it’s basically just like an online catalog, where customers pay a cut of sales for their goods to appear in Amazon search results. In the late aughts, an important change starts to happen. In late 2006 Amazon introduced its Prime subscription program for customers, which meant that for a flat fee every year, customers could get two day delivery on goods from Amazon. This basically loses money. You can argue about what makes or loses money with an Amazon. But this is basically a loss leader to try to get more and more people to sign up and therefore to buy more and more from Amazon. Shortly thereafter, Amazon decides to open its warehousing and logistics services to third party sellers.

They also begin a redesign of the software that they use to run the warehouses, which they had initially basically just copied from Walmart, and they hired a lot of logistics executives from Walmart early on. But after starting Prime and trying to rev up the growth of their retail division, they both redesign the logistic services and open them up to third parties, which means that the third parties paying to use those services, subsidize their redesign and subsidize Prime. Now, to a good capitalist, this all sounds great. This is efficiencies, you get these users to supplement the design — everyone wins. I do have that haunting sense sometimes that someone reads this and is like: Oh, this is great.

Anyway, that begins this new era and particularly in 2010, when they finished redesigning the fulfillment system, that opens up this new era that the people I interviewed described as the Wild West. Why they described it that way is that it’s an era of quick expansion and lawlessness or lack of rules. In practice, that redesign, combined with the 2008 financial crisis, which pushed a lot of people into needing it. They lost their jobs and they needed a new job, or they needed to supplement their income to pay off debts. It led a lot of people to seek gig work and non-job work. A lot of scholars of the gig economy have talked about this, how that same era brought a lot of people to work for Lyft, or Uber and various other gig work companies.

PM: Just to interject there — I feel like when we talk about the impact of 2008 and of the financial crisis, we do talk a lot about the impact that that had on creating the entire gig work economy. You had all these people who lose their job, who go out trying to find other ways of making income. Even if they’re able to find a job, again, it’s not making the same money that they often would have had in the jobs that they had before, lower paid jobs. So they need to find some way to make up for that and so that’s often associated with Uber and Lyft and some of these food delivery apps. But we don’t often think about Amazon in the same way. It’s not framed in the same way with Amazon, even though Amazon has all these third party sellers and is taking advantage of it through that route.

Also has its own gig services as well that it created, whether it’s through the delivery side of things. Mechanical Turk, of course, benefited from this and we forget that that’s even part of Amazon sometimes. There’s this whole piece of it, you said in the “Old Times,” you wrote in the report that sometimes recruiting the sellers didn’t always work out. This whole program didn’t always work for the sellers. But as Amazon revamped everything for the Wild West, as the recession happens and pushes a lot more people in this direction of looking for other ways to make money, that significantly benefits Amazon. There are a few business models that really emerge in the Wild West era. Can you talk to us a bit about that and how these businesses are formed and shaped by the platform that Amazon creates?

MW: For sure. To step back one pace from that question, then I’ll answer the question. When I first started this research, I had the perhaps naive expectation — I’ve lost all touch with what’s a normal amount to know about Amazon or retail — bt the maybe naive expectation that when I looked for Amazon sellers who I would be looking for, or the people I’d be talking to, it’d be people who had some background in retail who had a brick and mortar store and then used Amazon to “get online.” I put quotes around that because I, increasingly, think that as with every aspect of our social world and lives, the online/offline distinction is not all that useful.

Anyway, I thought I was going to meet the mom-and-pop at a corner store, who then thought: Wouldn’t it be nice if we sold our widgets online? Of course, this is not what selling on Amazon actually looks like. I was startled, initially, and then realized it made perfect sense when I discovered that almost every successful Amazon business I’ve ever encountered — and I’ve spent hundreds of hours speaking to people in Amazon world — almost all of them are Amazon native, with business plans designed specifically for Amazon. I mentioned this moment too, because a striking thing about those old timers who I interviewed, the folks who’d been doing it for a long time, with one exception, everyone I can think of got recruited by Amazon, started trying to sell through Amazon in that era, it failed. It didn’t work to take their offline business onto Amazon, and then they went away and came back with an Amazon specific business. The reason that I emphasize this is that, again, we have such a political religion around small business in the United States, and this idea of the independent entrepreneur, and so on.

As I’m sure your listeners are used to thinking about this belief in entrepreneurship has trickled into other areas where folks also think Uber driving is also sold as a form of entrepreneurship or working for Instacart. All kinds of things are sold as entrepreneurship. What’s striking about the Amazon sellers is they really are entrepreneurs in a certain sense: they register as small businesses; they have all these sourcing relations. But then in another way, Amazon has really gigified entrepreneurship in the sense that I’d say their entrepreneurial autonomy is very significantly shaped, constrained, if not undermined by Amazon’s metrics and Amazon’s algorithm or Amazon’s market power. The one popular business model, and the one most closely associated with the Wild West is a business model of reselling, or what’s called retail arbitrage. Retail arbitrage or online arbitrage refers to trying to source goods, places that aren’t Amazon and sell them for more money through Amazon.

Amazon has a seller app that was released in, I want to say, 2013 or 2014, and it has a barcode scanner that will facilitate doing this. It lets you use their app to walk through Target or Best Buy and scan barcodes and see what the current price of those goods on Amazon are. The other way that Amazon benefits from the post-2008 crisis is the so called “retail apocalypse” and this collapse of big box stores that had been built and overbuilt because of tax incentives. And other reasons throughout the United States from the 1990s-onward in this big box Walmart-era. Amazon inspires the rise of this new business model and many people who got involved, particularly in that Wild West era, started out by signing up for an Amazon seller account, which is not very much money. Right now it costs $39.99 a month to sign up for the bare bones account, and then going out to see if they could find stuff at their local Target or Best Buy or wherever to resell on Amazon.

Fascinatingly, to me, the people I spoke to who did this, ssome of whom went on to do pretty huge volumes, reselling $50,000 or $70,000 worth of stuff a day. It was so interesting, because they interpret it in these really gendered ways, even though they use different gendered language to describe exactly the same thing. They’re communities of people who call themselves coupon queens, or moms who often were good at shopping from shopping for their kids who who got into this. Also people who refer to themselves, usually men, as Amazon nomads, who would talk about driving around the country and driving to different places to source goods. If any listeners remember, that guy at the beginning of the pandemic, who’d bought 17,000 bottles of hand sanitizer and was trying to price gouge people for them on Amazon, who I feel like was the most hated guy on the internet for two days at the beginning of the pandemic. He was actually a pretty well-known member of this community of male Amazon nomads and doing something the definitely is antisocial and wrong, but it’s exactly what Amazon has encouraged the arbitragers to do, for years. I loved his story because it’s a great glimpse of the pathology of normal market mechanisms.

Retail arbitrage was this model that thrived in the early 2010s. I should say Amazon calls these fulfillment services that they sell to sellers, Fulfillment by Amazon or FBA. Before FBA, if you wanted to sell throughout Amazon, you need to have your own logistics outfit, and warehouse and all this stuff. Now all you needed to do is be able to find stuff and ship it into whatever fulfillment center they told you to ship it to. It greatly reduced the overhead for getting into the marketplace. That is a business model that thrived in the early 2010s. It’s gotten a lot harder for reasons that I’d be happy to expand on. It gradually has been supplanted by private labels, or Amazon brands, which I’d be happy to expand on. But I’ll hit pause there for a minute.

PM: I appreciate you outlining all that. I think it’s fascinating to connect back to that early story from early the pandemic of what was happening to illustrate how this all works and a particularly extreme case of it.

MW: No one, but me would remember this — Matt Colvin is that guy’s name — and I’ve met people who were friends with him who used to work with him doing arbitrage. In the New York Times story and in the court case that eventually followed, he was working with his brother. In all the pictures of him that were taken from the New York Times, he was wearing this t-shirt that said “family business.” I just loved it. Because it’s like this ethos of the small business and the family business used by Amazon. I could say more about that, but to justify blatantly sociopathic activity, shows how powerful I think that small business, family business imaginary is.

PM: It’s just wild. It’s a wild story, and to see how these things interact with Amazon and benefit Amazon for this to be able to go on and continue. But you mentioned the private brands as an extension of this, as the evolution of the mode as retail arbitrage become something that works less and less on the platform. All of a sudden developing your own brand, relabeling, often, Chinese goods to sell into the US and global markets, is the way that you make a go of it through Amazon. Can you talk to us about that element of this?

MW: Retail arbitrage gets harder and harder to do over the course of the 2010s for a few reasons. There’s natural price competition driving a race to the bottom, which a capitalist, or someone with faith in markets, would say: That’s good. That’s what arbitrage is supposed to do, it creates a market with a single price; that’s a good price. The other reason is that Amazon starts to disincentivize it and creates these policies that make it — I could go into the details, but they’re sort of boring — but in terms of the paperwork that they’ll accept for resale goods, they make it extremely risky to do it because there’s a chance that they’ll say: Oh, you have bought $20,000 worth of lawn flamingos and we’re just not going to accept it.

The other big reason is this rise of brands, or what some people describe, as the pivot to brands. Like everything, there are a few different factors that go into that. One is Amazon developing its ad tools and wanting sellers to use its digital advertising tools. It starts that division in 2012 and then rapidly expands it over the 2010s. As I’m sure listeners of this podcast know, it’s now the third biggest digital advertising company after Google and Facebook. You also then in this era, have the rise of Alibaba in China and Alibaba’s sourcing platform AliExpress, which they launched in 2010, I believe. Other sites like, that resell cheap Chinese goods in bulk to customers in the West.

Platformization and disintermediation are happening in the Chinese economy, too, in ways that make competing on price harder if that’s all you’re competing on, which is what arbitragers are basically doing. Then finally, you have Amazon’s opening in China, which is a big piece of the story. If there’s anything I cover in this report, and I’m planning to continue to do in my ongoing research, it is to tell the story of how much Amazon’s growth is built in cooperation with the Chinese state and on the direct backs or with the direct participation of Chinese small business. So Amazon opens in China in 2015 and that makes it harder to arbitrage because even if you’re going to Best Buy in North Carolina, it’s not easy to compete with factories offering the same goods. This introduces a whole bunch of other issues and problems.

This interacts with this rise of brands and you get a lot of Chinese merchants and factories, directly developing their own brands through Amazon for the US marketplace. That’s the origin. There was a magazine story about John Herman a month or two ago that I saw a lot of people sharing, but it’s the origin of all these Amazon “brands.” I’m trying to write an academic paper about this. A brand in an age of digital monopoly is basically just like a position in the search results. It’s not necessarily a household name, but all these brands that are like HMXQ, that if you go to try to buy shoelaces or earbuds or whatever on Amazon, you’ll be familiar with if you click through. That moment in the mid-2010s when Amazon opens and starts very actively recruiting manufacturers and merchants from China is the beginning of the rise of that Amazon native brand.

PM: That’s the period that you term as the “Jungle,” when this opens up and things really expand. Just to touch on some of the points you mentioned there about China. I remember, I used to think of AliExpress as like one kind of e-commerce marketplace. It was the place you could go to get access to all these low-cost Chinese goods that are provided by all these companies, but you weren’t really expecting to buy something from Alibaba directly. Whereas Amazon seemed like something separate from that where you were going on to Amazon, you were buying from Amazon, or a small number of approved sellers or something. It was a very different experience. Whereas it feels like in the years that have passed Amazon has moved much, much closer to what AliExpress was doing, as it has opened to China, as it has allowed more of these sellers onto the platform.

Just hearing what you say there about how you’ve been looking into how much Amazon’s growth, especially since the mid 2010s, has relied on China and cooperation with the Chinese state and the investments and policies that it put in place to enable this growth. Like it just seems fascinating to me that we have these discussions now around: Oh, my god, TikTok is in North America, we need to stop it. Or Apple has supply chains in China that maybe it should be pulling back on, but we don’t usually hear Amazon treated in the same terms, even though it’s so reliant on Chinese goods and that relationship with the Chinese state.

MW: I like to cite this figure because Amazon gets very upset when people cite this figure. But Amazon actually loves to bring out this figure that by volume, 49% of third party sales in the United States through Amazon, come directly from China to the US consumer. This is not counting the additional huge percentage of things that are sourced in China by American, or North American first party sellers. Amazon gets very upset when you say that number because Amazon loves when it is asked about being a monopoly, or when Bezos has called before Congress to say: We’re just helping; we’re not like Walmart; we’re not a big business, we’re just helping small business flourish and over 60% of our sales are from small businesses.

I feel it’s tricky to talk about this because we’re living in this moment of really stupid, in my view, great power antagonism between the US and China, really horrible anti-Chinese and anti-Asian racism in the United States. I say this, not to say unnecessarily that it’s a bad thing that Chinese merchants are selling directly to the US through Amazon, but rather to just point out the disingenuous or hypocrisy of Bezos going before Congress saying: America is a great country that support small businesses and entrepreneurs and over half of what we sell is from small businesses. It just should be added as a footnote that about half of that is directly from China, not to count the other vast percentage of stuff that’s made in China.

Amazon’s retail growth, like the growth of Walmart, in the 1980s and 1990s is deeply, deeply dependent on Chinese manufacturing, and more specifically, on China’s push in the 2010s to support micro-entrepreneurship and what they call overseas e-commerce — kuajing dianshang — and Xi Jinping and more specifically Li Keqiang, the premier, really pushed that and put all these resources into supporting different — I should say it’s at a provincial level that this is mostly administered. But different provinces put a lot of material resources into getting Chinese merchants onto Amazon, to sell to the US. So, Amazon’s rise is just deeply entangled with Chinese capitalism and the Chinese state in a way that I agree. I think it’s rare to hear about it.

PM: I should also say, I completely endorse what you say about this rivalry and conflict being ridiculous, but I just think it’s important to note how Amazon escapes from these narratives that are positioned and thrown at some other companies. If people do want to know a bit more about what China was doing with micro-entrepreneurship, of course, Xiaowei Wang was on the show back in October 2020 to talk a bit about that and their book, “Blockchain Chicken Farm,” which is fascinating.

MW: And to throw out one other name just because it’s publishing this month. I have a good friend called Lin Jiang, who teaches at the University of New Hampshire. She has a new book out this month, March, called the “The Labor of Reinvention” that is about the Chinese platform economy in the wake of the 2008 crisis. Lin is also wonderful and people should read her book.

PM: If we look at the phase of COVID growth with Amazon, as you say, in the report this is referred to as the “Rocket Ship” where things really take off in a significant way for those first couple years of the pandemic in particular. Now we’re seeing a bit more of a pullback on that, a realization that is not reflective of what things are going to look like, forever. Can you talk to us a bit about that period. Also, what it means for some of these sellers that these trends that were expected, that the business was planning for Amazon was going to build out all these additional fulfillment centers that they’re pulling back on a bit now. What it means for that not to have really taken hold in the way that some expected.

MW: I began this research, as I said earlier, a little bit before and then at the beginning of the pandemic, and part of why I was drawn to it because it seemed like a lot of change was happening in e-commerce world — both because of the trade war and the pandemic at that moment. There was this strong narrative that people probably remember just as consumers or citizens at the beginning of the pandemic, that it was going to drive a permanent shift from shopping in person to shopping online. People would say things like: Oh, e-commerce has just been accelerated by five years, 10 years or 15 years. Again, one of these tech determinist narratives in a certain way, I think it’s problematic in so many ways because it’s like: What is e-commerce anyway? I could say a lot more things to complicate that.

Basically, there was this narrative at the beginning of the pandemic, when people were quarantining and a lot more folks were shopping online that that was going to cause a permanent shift in American shopping habits. Amazon’s retail did grow a lot for two years after the start of the pandemic and Amazon stock prices went up and up. There was also this whole other field, which I’m doing research on now and I’d be happy to expand on a little bit. But there emerged, right after the pandemic, this whole area of private equity or venture capital firms called the aggregators that were founded in order to buy portfolios of Amazon native brands, on the theory — which has turned out not to be quite as easy as it seemed to those Wall Street guys when they got into it — but on the theory that they could use their superior data driven insights and scale to make these brands more profitable.

When I first started this research there was all this anticipation in general in the culture around e-commerce, all this hype, people selling their brands for huge multiples to these aggregators in the sense that there was new money suddenly to be made. Another thing I just want to mention is that probably few of us think about this now, but at the beginning of COVID, Amazon made this decision that they weren’t going to fulfill non-essential goods, however, they deemed what that was. This was a huge thing for third party sellers because it meant that, suddenly, they couldn’t send in any new inventory to warehouses.

The inventory that was in warehouses was held hostage and couldn’t be delivered. For some of them, all of them have these hard earned search rank results and metrics that are dependent, many of them, on fulfillment rates and so it would hurt their rankings a lot. Some of them have loans from Amazon, whose interest rates and repayment are tied to their sales. So their loans went into default. So anyway, this thing had happened at the beginning of the pandemic, that was a major wake-up call or trauma. Every seller I spoke to said: Look, we understand; it was a crisis; it was important to prioritize the right things. But I think this led a lot of sellers to appreciate, in a newly acute way, the risks of being so dependent on Amazon.

I should say that of the people I interviewed, almost everyone did at least almost 90% of their business through Amazon. I talked to one or two people, where it was only 60% or 70%, but most Amazon sellers, are hugely dependent for their livelihood on Amazon sales. Anyway, the beginning of the pandemic was this interesting time where there was a sense of a great deal of possibility, but also precarity, or worry, that Amazon could just turn off tomorrow and then your whole livelihood, and your company, could be held hostage or disappear.

I think that combination of outward expansion and exploration and so on, but also shakiness or riskiness, were captured by this metaphor of the “Rocket Ship” that people used. Even though again, a rocket ship is a tech industry cliche. I think the only one of these phrases that wasn’t a cliche I’d heard a lot before was the “Jungle,” which was fascinating to me. Over time, I came to think it, one, referred to the fact that marketplace was getting denser and more dangerous and you needed more advisors and guides to it and so on. I think also reflected the rise of an influencer culture of Amazon influencers trying to brand themselves with Amazon’s own branding because it’s like the “Jungle” — it’s like the Amazon.

But anyway, COVID was this paradoxical time where on the one hand, sellers saw a lot of opportunity and possibility, perhaps, to take money — and sometimes a lot of money — off the table by selling to these aggregators. On the other hand, were very aware of some of these external risks. Again, whether it’s a trade war with China, which is its own whole thing. I remember talking to people at the beginning when Trump had threatened to ban WeChat, who were just like: Oh my god, how am I going to communicate with my suppliers if they ban WeChat? Then also, of course, COVID itself. If we think of Amazon as a huge prediction machine — that’s basically what these machine learning powered systems are. The beginning of COVID was like what happens when the unforeseen happens to the prediction machine and this place runs very efficiently, not much safety or slack in the system.

PM: I do think it’s interesting to think about the effects of that and the narrative that we had, and how that affected a whole load of things around Amazon, and surely, the third party sellers were not free from that.

MW: Another thing to maybe say, just really quickly, at some point, if not here — I said earlier that sellers are very controlled by Amazon metrics and algorithms. One way that this manifests can be in pressures to scale, and to scale very rapidly if there’s demand for your product. There’s this metric called the IPI or Inventory Performance Index, which is this mysterious metric, but has a lot of impact on where people turn up in search results, and so on, or has to do with how many units you move. When I said earlier that Amazon simultaneously holds out this promise of independence and autonomy and entrepreneurship, but in many ways undermines it to the extent of its control.

I’d say that, in some cases, and especially early in the pandemic, sellers, small to medium sized businesses saw not only opportunities to scale, but also sometimes felt real pressures to scale to compete and to take on large amounts of debt to get more inventory or whatever might be. Again, that could then cause whole new headaches with the supply chain crisis starting in 2021-2022. But I think that Amazon, it’s not only that you can scale, but Amazon places on sellers, very distinct pressures and imperatives to scale, at their own risk.

PM: You describe how these businesses have to act like their own mini Amazon’s because of the whole system and structure that Amazon sets up for them to participate in. I think that touches on the broader question that I wanted to ask, which is, after doing this research, and after looking in depth into how Amazon’s platform, how Amazon’s marketplace functions — and the incentives that it creates for third party sellers — how should we understand Amazon’s larger business?

One thing I want to make a note of is that as the company takes over half of an average sale in the various fees and charges and things that it accumulates, you also note that a lot of these third party sellers are benefiting from public policies and public supports and all this stuff. Indirectly, that is benefiting Amazon as well, though, that’s not often considered as a direct subsidy to Amazon. Broad question is, how should we understand Amazon’s business understanding its relationship to third party sellers?

MW: You’re asking a great question. I can get weird and cosmic in a way that is probably interesting to know about what actually is Amazon. It’s actually be pretty hard to describe to an alien, or a future anthropologist or something — what Amazon actually is. I have two things to say this question. And maybe you will then have further interesting thoughts that will help me be clear. I think that when I say retail is at the core of the business, I tend to think of the third party sellers as an engine that subsidize and make possible the growth of new products and offerings. We could say this about the retail business at large. Cloud is developed to help serve the needs of the retail business and then sold to other people.

Some of the specific services that are developed for cloud are based on the needs of third party sellers, as well as the retail business. The advertising business is also developed for third party sellers and funded by third party sellers who pay for it. So within this larger logic of what we might think of as risk shift or this relatively familiar, probably your listeners familiar, narrative about platformization. That what platformization does is shift work and risk on to third parties while reaping profits or money from their labor and risk-taking and so on. Amazon applies that logic first to retail and gets all kinds of products out of that, but then also tries applying it to other realms too.

We have the Amazon Flex program, which is a program to platformize delivery there. At least six or 12 months ago, I can’t remember when exactly, I was looking at the Flex app applications and there were these really sinister, classic, predatory inclusion-type glossy stories on the Amazon Flex page about how if you are Black Latinx, or Indigenous, you could apply for a loan from Amazon so that you could start your own business to hire other people, to be a mini logistics company for you on Flex. I think there’s this drive to platformize everything, and when we think of what that means platformization, probably the best way to think about it is using data driven and computer technologies to shift work and risk on to third parties.

What if Walmart could be Walmart, but instead of ever — Walmart does a lot of this on its own, let’s not spare any punches for Walmart. It’s pushed a lot of risk of its supply chains onto sweatshop workers throughout the world. But if you can think of how do you use platformization to turn Walmart or Target into a global flea market, where everyone in the world is trying to sell things, but under Amazon’s control, that’s how this logic of platformization works. How do you take the risk and work and shift it on to third parties to whom you own nothing. I should say, Amazon has these business agreements with its third party sellers in which they basically renounced any kind of rights over anything, I can tell a million stories about Amazon suspending businesses, or suspending products, and people losing millions of dollars literally, over that.

In the time it takes them to reach a human at Amazon who can explain what happens, Amazon can ban a business and just hold its money, never give them back their money. So one way that the third party sellers help us understand what Amazon is and how it works is that it’s this infrastructure for innovating new products that enable it to continue extracting capital while shifting work and risk onto third parties to whom it owes nothing. Incidentally, Amazon, strongly prioritizes customers in all sorts of ways, but Amazon also gets rid of product liability to consumers, in a lot of cases. For years, the Supreme Court had a ruling on this in 2018, and it changed, but from 2000 to 2018 Amazon was not responsible for any sales tax. It didn’t have to charge sales tax, it could have more competitive prices because it didn’t have to charge that.

Third parties illuminate the way that Amazon applies this logic of shifting risk and responsibility while maintaining control. I think it illustrates, again, to circle back to something that you said, we have this narrative of the new Cold War and the Great Firewall around China. It also illustrates the extent to which these two digital capitalism’s, or these two platform economies, are deeply deeply entangled. Amazon not only does this platformization thing, but does it at a global scale, and in ways that entangle the US and China, in particular.

I should have said earlier when I said that 49% of third party sales by volume are directly from China to US customers, that Amazon has pushed and pushed in recent years to try to get more sellers in Mexico and Latin America, and in India and Pakistan. It’s still the most recent data is like only 6% of third party sellers to the US are anywhere except North America and China. It’s all North America and China. So anyway, it shows how these logics of platformsization that maybe we’re used to thinking about with a firm like Uber, are applied to new domains and at a very global scale.

PM: The report is so insightful on these points as well. Just as you say about Mexico and Latin America, there was a story recently, I think it was in Rest of World, about how some of the economies down there are being transformed because of cheap Chinese goods from Shein and things like that coming in. I think the expectation, then that, they’re going to displace China from the goods that are coming into the United States or North America, or what have you, is probably not going to happen to the degree that they would want to be able to show that they’re becoming less reliant on China.

As you are describing the way that the company actually functions, by looking at the relationship to the third party sellers and the broader platform as a nation that’s going on there. One of the things that you note in the report as well is that some of the antitrust frames and arguments around Amazon don’t properly reflect what is actually happening there are based on what you’re observing. Can you talk to us a little bit about that and how you understanding Amazon in the way that you do in the report, how we actually respond to it with that understanding?

MW: In the past few years, we’ve had this really developed antitrust discourse, not just in the United States, but throughout the world. The EU, Britain, China have taken anti-monopoly actions, not against Amazon, in the case of China, but in the EU and Britain against Amazon, China against Alibaba. In the United States, too, there’s been a lot of momentum on that front. A common frame that comes up when talking about competition has to do with how Amazon competes unfairly, with small businesses that sell through it. If some small seller has been selling something successfully, Amazon can invent an Amazon Basics equivalent of that thing, they have access to data about the product and the supply chain, and so on.

One person I interviewed use this term “dragon boated” and I was like: What does that mean? And he’s like: Well, that’s when Amazon just goes to your factory in China and buys them out for the next three years. Then you go back to place an order, and they’re like: Sorry, we’re already making three years worth of that for Amazon. In antitrust discourse, because the ultimate goal is competition. We often hear about these cases of how Amazon eliminates competition by eliminating smaller businesses and small competitors. That is true. That’s absolutely true.

I want to pause and just think about how amazing it is, in a way, now that we know thanks to the Marketplace Pulse research that Amazon is taking 50% to 55% of every sale anyway. That it’s like how high the margins must be for Amazon to look at that and think: Oh, but we could get an even bigger cut by making it ourselves. Because I have to admit, there’s a part of me that just doesn’t understand why they do that. It’s so clearly anti-competitive and they get such a big cut without doing that.

PM: I feel like there was some reporting recently that they’re pulling back on that a bit, but maybe I’m wrong.

MW: I’m sure you are right, you’re probably right because to me, it’s always been completely irrational that they do it because it’s so clearly a violation. Anyway, sorry, to answer in a more concise format, as just a normal human reading news coverage — before I went deep into the rabbit hole of Amazon — as a normal human reading news coverage I had, in my mind, this idea that Amazon destroys small businesses because it puts small businesses out of business. I think that’s a narrative with deep roots in American culture going back, at least to Brandeis and these battles in the 20s, over national grocery store chains versus small businesses. It’s something that lefties a little closer to our age, I barely remember, you probably don’t remember, but it’s in the 90s these debates about Walmart and Kmart and the small local stores fighting them.

Basically, it’s this narrative of big versus small and what big does is that it eats up small or destroys small. I think, while there’s some truth to this, and as I said, Amazon does copy products from its third party sellers, and so on. Big versus small is not the best way to cognize the power that Amazon has and how Amazon uses its market power. In addition to eliminating small businesses through competition, Amazon elicits or calls into being all these new small businesses that are optimized for Amazon and do this important work for it. As I was saying earlier that I thought I was going to be meeting mom-and-pop’s who just, incidentally, got on Amazon, but actually, it was almost all people who had to make a business plan for Amazon.

Their businesses optimized to Amazon’s goals and prerogatives and are very susceptible to its control, utterly susceptible, or vulnerable to certain algorithmic governance or policy changes, like deciding: Oh, we’re gonna just let anyone with a Chinese IP address also sell retail goods through our marketplace, which is, again, we can have a complex discussion about it normatively, but that’s a great big uh-oh for all these retail merchants they brought on in the United States for years. I think that rather than thinking of it in terms of big versus small, I like to think about how it exercises its power. This is how the logic of platformization works.

Facebook likes to say it’s made information more democratic because now anyone can be a broadcaster and anyone can start a digital native news outfit for much less money than it would have cost to start a TV station or whatever with the same reach. And that’s true. That’s not bullshit — that’s true. It has democratized the production and dissemination of information in some ways, at the same time, it’s created the greatest concentrations, monopoly maybe, over information. The greatest concentrations of information and wealth in a single person or a single entity that’s ever existed in the history of the world.

These things are both true and this is platformization. The scholar Anne Helmond uses this phrase, but “they decentralized in order to recentralize power.” I think with Amazon, I clearly don’t have a snappy enough way to say this, but it’s not quite so much about big versus small as about how this huge corporation enrolls small actors, and ultimately growing its power and recentralizing its power. I don’t know how we build a moral imaginary around that that’s as powerful as the vision of big versus small when it’s Walmart versus the mom-and-pop. But I think that to cognize how a firm like Amazon works, we need to think in those terms.

PM: I think it really makes two important points for us. First of all, there are all these narratives around decentralization in relation to the tech industry, and thinking about how decentralization from a tech perspective then enables a recentralization, that is helping these companies is really important. Because that piece of it is often left out of these decentralization narratives, if they even happen properly anyway. Then the other piece of it is to say that it’s not just that you should be fighting for small businesses or whatever in relation to these large firms like Amazon, but if Amazon is reshaping the way that those small businesses work in the first place, to serve its own ends, regardless of what happens from an antitrust perspective, then how much further ahead are we really getting, you know, with antitrust measures in relation to Amazon?

MW: Perhaps, it’s also a way for those on the left — or those whose hearts don’t flutter at the thought of competition in the same way that some of the antitrust discourse does — maybe seeing it this way also helps us reframe the question a bit so that we’re saying not, how do we protect competition or a ideal of market competition? But rather: How do we ensure that the whole global economy isn’t drawn into this process of aggrandizing this platform and drawn into whatever Amazon’s KPIs are? We talked about how to succeed on Amazon, you need to be like a mini-Amazon, but maybe it lets us reframe that values conversation a little bit, so it’s not only about competition, but what goals do we want the world economy to be optimized for? Amazon’s goal is to keep as many things moving as fast as possible through Amazon. I don’t think that should probably be humanity’s goal.

PM: No, and I think that is a great place to leave this conversation as well. Moira it has been great to talk to you, as always. Highly recommend people go check out the report with even more great insights around Amazon and how its business actually works. Thanks so much for taking the time

MW: Thank you.