Discover more from Margins by Ranjan Roy and Can Duruk
Amazon Private Labels and Antitrust
Is Amazon just a grocery store?
Ranjan here. This week I’ll be digging into the topic of whether Amazon’s private label businesses are monopolies
Techfolk: Amazon is a transformative company that has revolutionized everything from the way we shop, read, and even build business infrastructure. Jeff Bezos is a visionary that can fix the ailing American healthcare system and even take us to Mars.
Techfolk, when talking about antitrust: Amazon is just like any other grocery store.
While June certainly hasn't been rough for Amazon's stock price, it's been rough for the company on the antitrust front. There was a major ProPublica investigation on Amazon giving preferential treatment to its private-label products, and then the EU launched an investigation into Amazon's use of third party-seller data.
I kept seeing the response that Amazon promoting its own private-label goods is the exact same thing that retailers (especially grocery stores) have done for decades.
While most Margins' readers are most likely familiar with the business practice of private labels, as a quick reminder, private label goods are products created directly by the retailer. They’re either clearly tied to the parent brand (AmazonBasics) or an entirely new brand creation (like Target's 42 private label brands). Retailers can capture the entire margin so they’re usually able to underprice competing products. It’s why that bootleg-looking Walgreens acetaminophen is always a bit cheaper than brand-name Tylenol.
Share or revenue from private-label goods is pretty eye-catching for some retailers:
Office Depot: 20%
Dollar General: 20%
So every retailer creates and promotes its own private labels? Why is it then a big deal that Amazon does?
First thing - The Curse of Bigness and Goliath were two books that profoundly shaped my thinking on the topic of antitrust. As regular readers are aware, the concentration of big tech power is central to a lot of our writing.
But on the nitty-gritty of antitrust policy, I'm certainly not an expert. As writing Margins pieces is about trying to better educate myself on various topics, let's try to better understand the whole "is Amazon's private label business an antitrust problem" debate.
PERCENTAGE OF THE RETAIL MARKET
The first thing that seems to come up in any Amazon antitrust discussion is the question of their market dominance. This December 2019 Benedict Evans piece analyzing Amazon's market share usually gets cited. He came up with Amazon owning 35% of US ecommerce, and about 6% of the overall addressable US retail market.
First, this is all pre-COVID, where we almost have to assume Amazon has increased market penetration across most sectors. More importantly, it addresses the question too broadly. It's far more relevant to understand Amazon’s power per product category:
A study conducted by Stephen Kraus at Digital Insights last year found that Amazon averages a 74 percent share of digital transactions in the US in product categories that are sold through the platform.
In categories such as consumer electronics the study estimated that Amazon market share was about 80 to 85 percent. On the other side of the spectrum, Amazon captured only around 35 percent of the online market for women’s clothing. Within the products it actually distributes, Amazon at times controls over 90 percent of ecommerce sales.
For categories like consumer electronics, and others Amazon chooses to focus on, they are insanely dominant. For other stuff, not so much, so the average seems a bit more palatable.
This was Jeff Bezos's vision for The Everything Store - start with books, and build out power category by category. In ebooks, according to a Bloomberg estimate, Amazon controls 88.9% of all ebook unit sales. In terms of market dominance, for certain categories, it's unquestionable they are the only game in town.
Next, let's get more into the specifics of how Amazon's private-label businesses function. The first thing to note: Amazon has "first-party sales" (similar to any retailer) where it buys products from a manufacturer and has full control over the pricing, marketing and distribution. Then it has 3rd-Party Sales (the Amazon Marketplace) where it functions as a platform. Within the Amazon Marketplace, there is Fulfilled-by-Amazon (FBA), where Amazon holds the 3rd-party sellers inventory and manages all distribution, and then Fulfilled-by-Merchant, where the merchant handles everything and just uses Amazon.com as another sales channel.
The Marketplace is at the center of all of this. In his 2019 shareholder letter, Bezos exclaimed, “Third-party sellers are kicking our first-party butt. Badly." 3rd party sales (with FBM and FBA combined) accounted for 58% of gross sales across all of Amazon at the time. It's a booming business.
This is where things get interesting. Especially for its FBA business, you can imagine just how much data Amazon has on the products. They know how many items are sold. The conversion rates. The inventory turnover. They often know exactly the manufacturer, as they directly receive the products.
They know a new hit product before anyone else sees it coming. As they often have multiple sellers selling similar products, they would know more about the product category than each individual seller.
Basically, they have all the data to tell them what's a winner. They'll know this before the rest of the world has any idea.
NASA AND ADVERTISING BUDGETS
Hal Singer's piece takes this even a step further. Amazon's advertising business has been exploding (40% YoY growth, $14.1 billion in revenue in 2019). What does that mean for antitrust?
Amazon has a deep understanding of how much ad spend is required to sell a product. Anyone involved in ecommerce understands that online customer acquisition cost efficiencies are everything. Amazon not only understands which products are selling and how fast, they understand what products are selling organically versus being propped up with heavy ad spend. As Singer states:
Comparing this type of refined information to the data that brick-and-mortar retailers collect is like comparing the data NASA scientists use to that used by medieval astronomers.
I think Amazon, in its current state, is a highly problematic entity (note: I still shop on there, a good amount. That's how monopoly works.) But I will never pretend Amazon is not an absolute innovation monster. They are NASA Scientists and our antitrust discussion assumes medieval astronomers.
Amazon is just like a grocery store.
Except they can track every customer that walks in and follow their eyes to know exactly what they're looking at for how long. They can even read the minds of every one of those customers to see what they're interested in, and just how interested they are. They can aggregate and analyze that data in near real-time, and leverage years of aggregated data and machine learning to predict exactly what private labels they should be creating. Normally, I'm quite skeptical of tech companies touting "MACHINE LEARNING!" but these are the exact predictive problems it’s built to address. And, for many product categories, they're the only grocery store in the country.
Finally, Singer brought up one other key point: traditional private labels are built on replicating already established product categories. This means the original seller is already established and competitive. Amazon is they can pick off every burgeoning new product category. They can crush the small and medium-sized sellers innovating and inventing new product categories. They can wait and watch which products are picking up steam and are the most profitable, and simply attack these vulnerable sellers. Leave everyone else fighting for the lower-margin scraps and only attack the under-the-radar, juicy things. It’s why the big WSJ investigation focused on Car Trunk Organizers.
Seriously, how the hell did Amazon just come up with car trunk organizers as something to pursue?
Amazon Just Lies
Next, we get into the most elementary problem: Amazon just lies.
Amazon is not supposed to leverage individual-level, third-party seller data to inform their own product development (they are allowed to view "aggregate" category data, which is still incredibly valuable).
Amazon's Associate General Counsel testified to Congress last July, "We don’t use individual seller data directly to compete." Back in April, the WSJ published an investigation that found 20 former employees who directly said they did:
Some executives had access to data containing proprietary information that they used to research best selling items they might want to compete against, including on individual sellers on Amazon’s website. If access was restricted, managers sometimes would ask an Amazon business analyst to create reports featuring the information, according to former workers, including one who called the practice “going over the fence.” In other cases, supposedly aggregated data was derived exclusively or almost entirely from one seller, former employees said.
Of course they did!
Amazon is a hugely ambitious and aggressive machine, willing to work its workers to death both before and during the pandemic, and they get away with it. It can sell counterfeit opioids that kill people, weapons, and drugs, and then invoke Section 230 to shield itself from liability. And in true platform fashion, they've mastered the art of the tech apology:
“There are bad actors that attempt to evade our systems,” Amazon said of products in violation of its policies that appear on the site, adding that “should one ever slip through, we work quickly to take action on the seller and protect customers.”
It's almost comical to me that they would not be leveraging all this data when the only safeguard is their own internal controls. Haven’t we all learned by now?
Finally, one point that I don't see addressed in most of what I read on the topic is Amazon's ability to subsidize these private-label categories. At some point in an economics class, I had learned the term "dumping", which references selling a product at a loss to eliminate the competition.
Specific to Amazon, there is a famous 2009 case with Diapers.com:
Soon after, Quidsi noticed Amazon dropping prices up to 30 percent on diapers and other baby products. As an experiment, Quidsi executives manipulated their prices and then watched as Amazon’s website changed its prices accordingly. Amazon’s pricing bots—software that carefully monitors other companies’ prices and adjusts Amazon’s to match—were tracking Diapers.com.
Selling an individual product at a loss is one thing. But in 2020, Amazon can leverage its profit powerhouse of Amazon Web Services, or even it's pure profit Amazon Prime membership fees, to subsidize every new private label product.
Then, to compound things, you have the eternal handshake between Bezos and AMZN shareholders. AMZN shareholders don’t expect a profit. Just think about the additional financial leverage and competitive power that provides an organization.
I don’t think antitrust law addresses beneficial capital-raising abilities in any substantive way (please reply if I’m wrong). But specific to private labels, between the ability to subsidize with AWS profits, combined with a lack of profit expectation from shareholders, Amazon has so much more wiggle room on the margins side than any other retailer. Normal private labels underprice competitors by a little. For its private labels, Amazon could bleed money and show worse unit economics than even Uber, and still come out financially fine across the company.
Perhaps we could understand how these financial dynamics work together, but Amazon famously doesn’t break out any information:
Jeff Bezos’ Flywheel strategy is the stuff of organizational theory legend. But the one missing piece of the diagram is monopoly power; where sellers in certain product categories are dependent on your platform and you can leverage their seller data to create a virtuous feedback loop with your own private label business. They’ll stick around because they have to and the flywheel will spin faster and faster. Then add in all the high margin adjacent businesses that can direct even more energy into the retail flywheel.
Amazon is a one-of-a-kind business beast that’s brilliantly built up a powerful ecosystem of interests that all feed each other to increase the overall organization’s power. It’s why the same company that sells you toothpaste also tells you the weather and is your bookshelf and shows you NFL games and wins Oscars and is your doorbell and is your WiFi system and is your kid’s tablet and hosts your startup’s entire infrastructure and I’m probably missing a few things. Amazon is not a grocery store.
Note 1: The Amazon Coat
In one of my first Margins’ pieces, I wrote about the Amazon Coat. Remember it? I’m still partially convinced that somewhere within the Amazon organization a bet was made. “I bet you I can get a bunch of posh NYC women to wear an ugly, cheap coat made by a Chinese plastic furniture manufacturer and convince themselves it’s fashionable.” That’s just how powerful the flywheel is.
Disclaimer for anyone who owns the Orolay coat and is angry reading the above passage: My wife bought one too.
Note 2: Antitrust tests
Apparently, when talking about “dumping” or artificially lowering prices to kill competition, there is an official test in antitrust theory called the “profit sacrifice test”. This led me down a rabbit hole to discover there are also amazingly named tests like the “no-economic-sense-test”
Note 3: Show Your Work
The Markup had a great piece on Amazon selling dangerous or banned items. One of the things I love about the Markup is their "Show Your Work" section, where they provide full transparency on how they researched a piece.
For this one, they capture it so perfectly: "It wasn't that hard".