What does the future of e-commerce look like? Today in the US, it seems to be largely dominated by the likes of Amazon, which owns nearly half of the market. But will this always be the case? And if not, what would the disruption of a juggernaut like Amazon look like? In this article, we explore these questions and make the argument that the future of e-commerce is likely to become significantly less concentrated than it is today, and that Amazon’s business will look, perhaps surprisingly, a lot more like that of an airline than that of a platform monopoly.
What Are The Principal Components Amazon’s E-Commerce Business?
To understand Amazon’s e-commerce business, we need to first break it up into components. This “unbundling” will allow us to then reason about each component individually in order to forecast what the future looks like for Amazon’s e-commerce platform as a whole. Amazon has many businesses, but we think that the core of its e-commerce business can be distilled into three principal components:
- A “logistics” business. This consists broadly of a network of warehouses, a large workforce to package products for shipping, relationships with shippers that allows it to fulfill orders that its users place on its marketplace platform, and software to manage all of these operations. Notably, Amazon uses this logistics business to fulfill not only inventory that it owns but also to fulfill inventory it holds on behalf of third-party sellers on its marketplace platform through a program called “fulfilled by Amazon” or FBA. At a high level, everything Amazon does “behind the scenes” to deliver a product to someone who ordered it, whether that product came from its own inventory or the inventory of a third-party seller, can be considered as part of this logistics component for the purposes of our analysis.
- A “marketplace platform” business. This consists of the website amazon.com, which maintains a two-sided marketplace with a large user base that places orders on one side and sellers on the other side who list, advertise and sell products on the platform. Importantly, it is worth noting that Amazon itself is one of many merchants who “plug in” to the seller side of the platform, and that some third-party sellers use Amazon’s logistics engine to fulfill their orders by first shipping their products in bulk to an Amazon warehouse, while other sellers fulfill orders themselves directly and just use the platform to list their products. In the case where a third-party merchant is not using Amazon’s logistics business for fulfillment, the marketplace platform looks more like eBay, effectively acting as an “aggregator” of merchant listings. Indeed, the power of Amazon as a platform stems in large part from the fact that it is not the only merchant on its platform but rather an aggregator of third-party merchants (as a reference, “aggregation theory” refers to the tendency of an internet platform to “aggregate” players in a market and commoditize them, as Amazon’s marketplace platform has effectively done to third-party merchants). Notably, we think this marketplace platform component should be considered as distinct from Amazon’s logistics business when analyzing Amazon. Amazon’s logistics business “plugs in” to its marketplace platform, but it could easily exist as a stand-alone entity. For example, there is no fundamental reason why Amazon couldn’t offer fulfillment services (i.e. FBA) to merchants on eBay’s platform or advertise products that it sells directly itself on eBay’s platform as well. In fact, third-party sellers have been known to use Amazon’s logistics to fulfill items ordered from them on eBay through a practice known as “drop-shipping.” As such, it should be clear that Amazon’s logistics can be leveraged by third-party sellers independent of its marketplace platform (e.g. a seller who lists on eBay but fulfills using FBA), and that, on the flip-side, Amazon’s marketplace platform can be used by a third-party seller without making use of Amazon’s logistics business (e.g. a seller who lists items on Amazon but does not use FBA). Thus it should be clear that the two businesses can be considered separable for the purposes of our analysis. Of course, there are strategic reasons why Amazon may prefer to “tightly couple” its logistics business with its marketplace platform, but we will discuss these later as well as consider why this may cease being the case in the longer term.
- A “loyalty program” known as Amazon Prime. Perhaps the most unique component of Amazon’s e-commerce business is its Amazon Prime membership, which it offers as a paid benefit to buyers on its platform. The program initially started as a loyalty program constructed around a strategy to increase economies of scale on the logistics side of the business. Customers who signed up for Amazon Prime and paid an up-front fee would be entitled to free and unlimited two-day shipping, as well as other perks such as free refunds. While there were likely many components to the strategy of offering such a loyalty program, we believe one major reason why the company enacted it was to gain an “economy of scale” around its logistics business. Simply put, by offering a loyalty program like Prime at a loss initially, Amazon thought it could significantly increase long-run order volume. This would allow it to eventually exploit economies of scale in its logistics business, lower shipping costs, and thus eventually result not only in the loyalty program being sustainable but also in a competitive advantage that would be hard for others to replicate in the form of a loyal buyer-base and a logistics business with larger economies of scale than any other e-commerce operator. Notably, the strategy was a money-loser for a long time, which also fits the general model for bootstrapping a two-sided marketplace, whereby operators initially subsidize one side of the marketplace (in this case the buyer side) in order to attract the other side of the marketplace (in this case third-party sellers). Today, Amazon’s loyalty program offers much more than fast/free shipping and free refunds, encompassing streaming videos (Amazon Prime Video), grocery delivery (Amazon Fresh), and more. But ultimately, in spite of the fact that Amazon Prime’s benefits are somewhat broad today, we will argue in a later section that it is not much different than loyalty programs offered by airlines. In particular, we will argue that this program could easily exist independent of Amazon’s marketplace platform, similar to how airline rewards are tied to a particular airline rather than to the platforms which aggregate ticket booking (such as Hipmunk, Expedia, etc…). Moreover, we think there is reason to believe this “unbundling” of Amazon’s loyalty program from its marketplace platform could happen in the long run.
Notably, in enumerating the three components above, we’ve only concerned ourselves specifically with Amazon’s e-commerce business. Of course, Amazon is a conglomerate, operating other relatively independent businesses such as Amazon Web Services and Amazon Echo, but we think considering these businesses as distinct and focusing solely on e-commerce is the best way to draw meaningful conclusions about that part of the business.
How Does Amazon’s Business Compare to the Airline Industry?
Interestingly, when we break Amazon’s business into the three components we enumerated earlier, namely a logistics business, a marketplace platform, and a loyalty program, we begin to see marked similarities between Amazon’s business and that of the airline industry. The key difference today is that Amazon manages to “vertically integrate” these three components in a way that the airline industry either unable to pursue or forbidden from pursuing due to regulation. In other words, whereas Amazon owns all three components of its business, the airline industry has different owners for their equivalent of the logistics business and the marketplace platform. But nevertheless, we think it is quite instructive to enumerate how the three components that make up Amazon’s e-commerce business analogize to the airline industry.
To understand how the components of the airline industry map to Amazon, it’s easiest to consider what you do when you go to book a flight. First, you start at a ticket aggregator such as Expedia, Hipmunk, or Google Flights. This piece maps to Amazon’s “marketplace platform” because it aggregates buyers of tickets on one side and airlines providing “A to B” flights on the other side (the equivalent of aggregating merchants on Amazon’s marketplace). Following this analogy, airlines like American Airlines, United, etc… are responsible for the logistics of actually getting you where you need to go in a timely manner, making these companies analogous to the logistics portion of Amazon’s business, which mainly concerns getting packages where they need to go (as opposed to people). Moreover, similar to how Amazon’s marketplace platform tends to aggregate and commoditize third-party sellers, ticket aggregators tend to do the same to the airlines who list flights on their platforms, the difference being that ticket aggregators do not themselves operate airlines in the way Amazon operates as a merchant on its own marketplace platform. Given this tendency of an aggregator to commoditize those who list on it, and because airlines, as primarily logistics businesses, benefit from economies of scale in terms of being able to run full planes as often as possible, airlines typically try and bundle loyalty programs into their offerings in the form of “frequent flyer” programs or other schemes (similar to how Amazon Prime was initially intended to bootstrap an economy of scale for its logistics business). Examples of this include AAdvantage and United Mileage Plus, which offer benefits beyond simply cheaper tickets such as upgrades, lounge access, and other perks, not unlike Amazon Prime’s expansion beyond offering merely fulfillment benefits. To summarize in bulleted form:
- Amazon’s logistics business = a traditional airline (e.g. American Airlines, United, etc…).
- Amazon’s marketplace platform = a ticket aggregator (e.g. Expedia, Hipmunk, etc…).
- Amazon Prime = frequent flyer programs (e.g. AAdvantage, United MileagePlus, etc…).
So, if the two industries are so analogous, a natural next question becomes: Why was Amazon able to vertically integrate its components so effectively while the airline industry wasn’t? While we don’t think we have a certain answer one way or another, we suspect regulation preventing consolidation in the airline industry is the main reason why that industry can’t vertically integrate the way Amazon has. To understand this, let us consider one potential path to increased profit for the airline industry in the absence of regulation:
- First, the logistics side of the business, consisting of the major airlines, would consolidate through mergers and acquisitions. Due to continued increasing returns on scale, we don’t see a reason why the airline industry would not consolidate its logistics business to the extent Amazon has, or more, in the absence of regulation, resulting in one airline owning over 50% of the market. This being said, in spite of recent airline deregulation increasing M&A activity, one airline would likely still not be legally allowed to control over 50% of domestic flights in the way Amazon controls roughly 50% of e-commerce. Why is this distinction important? Because there is a big difference between a market in which there is one major player and a lot of small fish, and one in which there are four major players who are not allowed to coordinate as is the case with the airline industry. Specifically, a situation with multiple equally-large competitors typically results in an unstable equilibrium in the prisoner’s dilemma game while a situation with a single juggernaut may not. This makes a situation with multiple equally-large competitors, as is the case with the airline industry today, a much more competitive situation that can easily thwart the further vertical integration steps we cover below.
- If one major airline controls the logistics side of the business, it could then use its leverage to create its own ticket aggregator where one could always expect to find the vast majority of flights. Not only that, but it could competitively exclude all of its flights from competing aggregators, thus extending its economy of scale in logistics to build a two-sided network around its own ticket aggregator. In doing this, it could not only “cut out the middle-man” by disintermediating existing aggregators like Expedia and Hipmunk, but it could also then effectively force any remaining unconsolidated airlines onto its platform by virtue of it already having exclusive access to the vast majority of flights (namely its own flights). Such a move would almost directly mirror Amazon’s extension of its marketplace from serving only its own listings to allowing third-party sellers on-board. Again, this type of coordination is not only much more difficult in the absence of a clearly-dominant market leader, but it is also important to note that the consolidation of a two-sided ticket aggregator with a largely-dominant airline is something we expect would be untenable from a regulatory standpoint (though we don’t know this for sure).
- Finally, although this may not be strictly necessary, the use of a loyalty program associated with the now-dominant “airline + ticket aggregator conglomerate” would help make it very difficult for other already-disadvantaged ticket aggregators to compete with the juggernaut because the buyers of tickets become relatively locked-in to the dominant aggregator. This would be analogous to Amazon’s implementation of Prime, which served not only to help bootstrap its initial logistics dominance but now also serves to maintain that dominance. Notably, if airlines could successfully consolidate the logistics side of the business, it does not seem terribly necessary to also offer any form of loyalty program because their control of the dominant aggregator seems stable even in the absence of such a program– but we could be underestimating the extent to which it is important to continuously fend off small competitors in the market.
Although not necessarily inevitable, we hope the above helps to paint a picture of how vertical integration in the airline industry, were it feasible from a regulatory standpoint, would be a highly likely consequence of the industry actors’ natural tendency to pursue economies of scale to maximize profit.
How Does Amazon Get Unbundled?
After walking through the hypothetical picture of vertical integration in the airline industry in the absence of regulation, Amazon’s current circumstances seem quite stable, having vertically integrated its now-dominant logistics business with its now-dominant marketplace/aggregator business. Given that, does this mean that the future of marketplace platforms is just Amazon all the way down? Actually, we don’t think so. Rather, we think it highly likely that, in the future, regulators will seek at minimum to “unbundle” Amazon’s logistics business from its marketplace platform and, less likely but still possible, to seek to break up the logistics side of its business to create multiple dominant players as is the case in the airline industry today (possibly both of these).
How likely is it that the government will eventually impose regulation on Amazon? We think a good way to answer this is simply by asking: Is there a benefit to the public in unbundling Amazon’s business?
To answer this, let’s consider what would happen if the government were to force Amazon to choose between divesting itself of its logistics business or divesting itself of its marketplace platform, effectively forcing these two pieces to “unbundle” and be operated by different corporate entities. Let’s call these two new companies “Amazon Logistics” and “Amazon Marketplace.” Perhaps the biggest impact of such a change would be that the now-standalone “Amazon Logistics” business would see little downside to offering products on other marketplace platforms such as eBay or even Amazon’s former competitor jet.com, compared to today where the strategic value to Amazon of offering products exclusively on amazon.com is quite high. This seems like it would significantly decrease the margin that the operators of such marketplace platforms, including the now-separate “Amazon Marketplace” entity, would be able to extract from their users, since products offered by the now-separate “Amazon Logistics” company would be available on any platform where buyers wish to shop on (including, potentially, via a direct Google/Bing search). Moreover, assuming “Amazon Logistics” were allowed to stay intact as opposed to being further broken up, then that entity would still be able to reap the benefits of its economy of scale in logistics, even if it ceases to be attached to the platforms on which it lists its products. Amazon Logistics could even be allowed to maintain its “Amazon Prime” loyalty program, and as long as it agrees not to exclusively list its inventory on one particular platform it seems like society as a whole would be much better off without impacting the bottom line of Amazon Logistics too heavily. Indeed, as long as Amazon Logistics maintains its quality of service, it should expect to sell a similar amount of inventory.
Of course, what would obviously be lost in the separation of Amazon Logistics from Amazon Marketplace would be the profitability, and ultimately the competitiveness, of Amazon Marketplace and, more broadly, of aggregator platforms in general. Indeed, without the ability to exclusively offer Amazon Logistics inventory, the market power of an aggregator like Amazon Marketplace is significantly diminished, as is the rent it can extract by charging fees to third-party merchants on its platform. Again, to use the airline industry as an analogue, consider the relatively weaker market power of companies like Expedia or Hipmunk, who largely offer the same inventory on their sites. However, although the profitability of Amazon Marketplace would certainly decrease, it seems highly unlikely that it would not be viable as a stand-alone business. Moreover, as far as value to society is concerned, a drop in the profitability of Amazon Marketplace results in almost a direct increase in surplus to buyers and sellers on the platform, and a decrease in “dead-weight loss” caused by the rent extracted by Amazon Marketplace (and indeed other aggregators as well). Not only that, but because Amazon Logistics would likely be willing to post listings on any aggregator, regardless of how dominant that aggregator is, it seems like the relative drop in Amazon Marketplace’s market power would be complimented by a rise in up-start marketplace platforms that try to improve on the user experience for particular verticals (think like what Airbnb did to the housing section on Craigslist). Imagine custom niche marketplace platforms tailored to particular types of items that can abstract away fulfillment by relying on a now-decoupled Amazon Logistics backend to deliver items. And contrast this with the current UI on amazon.com, which has not seen a major redesign in years.
Given all of the above, it seems the value to society of splitting the Amazon Logistics component from the Amazon Marketplace component would likely create much value for society not only by redistributing the deadweight-loss extracted by Amazon Marketplace’s current market position to buyers and sellers, but also by encouraging further innovation on the marketplace side of the equation by increasing the incentive for small up-starts to craft highly-tailored marketplaces for particular verticals. Given the value to both buyers and sellers of separating Amazon Marketplace from Amazon Logistics seems quite high, we think it will be seriously considered by government in the future.
Now, although we have discussed separating Amazon Logistics from Amazon Marketplace, we think it is still worth considering whether it is worth it from a societal value perspective to further break up Amazon Logistics into smaller pieces. The problem with doing something like this is that breaking Amazon Logistics up would cause it to lose the economies of scale associated with operating as a single entity. In fact, it is not even clear how one would divide such an entity without significantly damaging the efficiency of the machine as a whole. As just one example, forcing a break-up of Amazon Logistics into “regional” components would cause efficiency to decrease when transporting items across long distances. Moreover, the gain to society of such a break-up seems like it would be limited to the differential between what Amazon Logistics can charge through its economy of scale and what an individual seller would charge to fulfill an item without this economy of scale, which ultimately does not seem like a large amount of value to go after in the first place. This seems to be due in part to the fact that Amazon Logistics would likely still rely on commoditized “shipping rails” that are either competitive like FedEx vs UPS, or operate as a public utility like USPS. Perhaps if Amazon Logistics found a way to completely monopolize the ability to ship an item from one place to another, the way airlines can sometimes monopolize particular routes, then there would be more cause for concern. But this seems unlikely to happen, even in the long run. As such, all things considered, it seems like the juice from breaking up Amazon Logistics is not really worth the squeeze.
The Long Run: An Open Marketplace Platform
Of course, it is far from a guarantee that government will decide to do something like unbundle Amazon’s business. Indeed, it would be a drastic step the likes of which we haven’t seen since the breakup of the Bell telephone monopoly in the early 80’s. That being said, hopefully we have made a decent case that splitting the “Amazon Logistics” component from the “Amazon Marketplace” component is something that would likely increase societal surplus, while still allowing both businesses to earn a decent profit as stand-alone entities. In this section, with the understanding that it is not obvious such a breakup will eventually happen, we elaborate on one potential outcome if such a decoupling were to occur.
Until now, all stand-alone marketplace platforms such as eBay or Craigslist have been centralized entities owned and operated by private, for-profit corporations. One would expect such entities to perform well given their profit-motive gives them a decent incentive to continue to provide benefits to their buyers and sellers. Indeed, it is difficult to imagine a nonprofit marketplace platform being able to bootstrap itself in today’s market, even if the eventual unbundling of Amazon were to occur. But what if there were a way to create a new type of marketplace that is not a non-profit but simultaneously not owned by any individual entity? What we envision in the long run is effectively a marketplace platform whose code is completely open-source, meaning a global community of developers can modify it as opposed to a small group of hierarchically-governed engineers at a company. We think such a platform would likely be able to improve and innovate at a faster pace than something that is privately-owned, in the same way Linux or Android was able to eventually become competitive with privately-developed software for server-side architecture and mobile phones respectively (indeed, would Android have been able to compete with iOS were it not open-source?).
Although it is still early days, an open-source marketplace platform is what we are working on with the Ultranet, the first blockchain marketplace. And because it is fully open-source, we are already seeing the power of the community help us in improving it. Imagine a marketplace where all the listings are effectively in a “public database,” and where anybody can write a new “ranking algorithm plugin” that you have the option of using instead of the default. Imagine if, like with spam filters that developed on top of email as an open protocol, an open marketplace platform like the Ultranet can allow for the rapid collaborative improvement in the user experience of the platform, improvement in the ranking of listings, the creation of UI’s for specific niches, and much more all on top of an open-source base. Put another way, imagine if the “Amazon Marketplace” component of Amazon’s business could be something that is developed collaboratively and openly in the long run. That is what we’re trying to develop with the Ultranet. And as of right now, the first iteration of the Ultranet is fully-launched, downloadable on the website, and functions as the only marketplace where orders are end-to-end encrypted (initially attacking privacy as an underserved niche). The Ultranet also has its own cryptocurrency called Ultra, listing later this month on the LAToken and p2pb2b exchanges, that we designed to help solve the platform chicken and egg problem.
Of course, it is impossible to know for sure what the future will look like. After all, who knows, with the rapid pace of technological progress, maybe someone will invent a matter transporter tomorrow and render most of the arguments in this post moot. Nevertheless, we don’t think one can go wrong forecasting the best ways to maximize societal value given today’s constraints, and we hope our exercise in doing that here was at least enjoyable for you.