Updates #10

Amazon's 3P marketplace pricing story takes a different turn

Not held.

---

A quick note for potential subscriboooors - from July 1 I plan on increasing the subscription price of new subscribers for the blog (existing subscribers will not be affected). If you enjoy the previews please take advantage of the free article or the 7-day free trial.

---

---

Having recently sold the last of my Amazon shares I thought it would be a propitious moment to touch on how the story has been changing with respect to Amazon’s best business: retail.

I expertly bought my first Amazon shares at the start of 2022, just before a near 50% decline in the share price. Even though that was less than 4 years ago, Amazon stands at a different juncture now than it does now. My thesis was not complicated: retail pricing up, cloud margins normalised, capex down, and eventual capital return. I was partially right… for a little while at least.

I think Amazon’s North American retail operation is one of the wonders of the modern (business) world. The fact that Bezos, and the line up of retail lieutenants who assisted him over the years, were able to daisy chain their way from a measly book retailer to probably the most dominant online retailer in the world is nothing short of miraculous. It is truly one of the great entrepreneurial achievements in history.

The correct lens, however, through which to understand the modern Amazon is provided by Ben Thompson:

Amazon is building a lot of businesses that look like AWS: taxes on major industries that work to everyone’s benefit. The reason, though, is that AWS is a lot like Amazon itself.

Stratechery, The Amazon Tax (2016)

The Amazon that we know and love today builds infrastructure and then commercialises it “as-a-service”. It builds data centres, and rents them out. It buys GPU’s and then rents them out. It’s built a pipeline directly from your phone, computer, or Alexa to your front door. Unsurprisingly, that pipeline is also available for rent.

The irony, of course, is that Amazon started life as a physical retailer, and over time backed their way into a purportedly more efficient 3P marketplace business. One of the longest running tropes amongst certain investor groups is that Amazon’s traditional 1P retail operation is of decidedly low quality (‘…and you get retail for free’). I was a proponent of this! In the light of my better angels, I think that take requires more nuanced analysis.

1P retail is the underlying draw of Amazon to customers, and its ultimate aggregator of demand. In Bezos’s own words, the promise of Amazon to its customers is lower prices, wider selection, and convenience. Indeed it is difficult to think of a more enduring cocktail of consumer desires. In the early days this was (in)famously applied to books: Amazon’s centralised inventory meant that they were able to offer the widest selection of books to customers. At a certain scale their buying power pushed prices down. It was only the convenience part that would be answered when 3rd party shipping partners found they were unable to keep up with demand.

The book parable could be applied to almost every SKU imaginable. In time Amazon’s purchasing power could be used to source unfathomably large quantities of every consumer product known to man. The internet forum meant that it was not only economic but expedient to offer as a wide a selection of products as possible. From the 2013 holiday season they began nailing down the convenience part, mostly out of necessity.

Unfortunately, the 1P business does not have an exciting income statement. To meet its promise to consumers, and indeed grow, the company needed to suppress its profit margins on an already low margin endeavour. The cashflow statement was another issue entirely. The nature of payments in this regard - the customer pays first, and then Amazon delivers the product - engendered one of the most extreme growth curves in modern business. Amazon might have had lousy margins but it operated on negative working capital, and every new order funded the one behind it. The stuff dreams are made of.

Beginning in the early days of Amazon’s existence, but turbocharged during 2013 and then in the 2020-2022 buildout (blowout?), was the labyrinthine construction of Amazon’s back-end logistics infrastructure. While this might have started with a move to solve the last mile problem that UPS and USPS struggle with, it’s culminated in more than 110 NA large scale fulfilment centres, and a further 490+ logistics facilities. This is to say nothing of its last mile delivery fleet, air fleet, airport infrastructure, trucking and shipping capabilities.

Certainly the 3P marketplace business has existed for a long time. Amazon first opened this to the second hand goods category in late 2000. Its efficacy, however, to both merchant and Amazon would not be fully proved until the advent of Amazon Prime (2005) and Fulfilled by Amazon (FBA) in 2006. These are the foundations of Amazon’s rock solid market power over their merchants (3P sellers on Amazon): The vast consumer surplus supplied by the 1P business aggregated incomprehensible amounts of consumer demand, Amazon’s physical infrastructure allowed them to deliver convenience benefits to Prime subscribers (sometimes same day delivery), and the access to these were gated by FBA.

Starting in 2018, and ending in early 2024(ish), Amazon took an enormous amount of pricing out of the 3P marketplace - and its related businesses. This was probably done more out of necessity than anything else. Amazon’s enormous infrastructure and cloud build out, as well as its breath-taking step up in headcount, necessitated materially higher levels of monetisation. Pricing really couldn’t come out of 1P, or AWS for that matter - although cost savings could. Some subscription services experienced a nominal price increase, but Amazon really ripped the pricing lever in a byzantine series of service charges to merchants.

As additional context, Amazon will allow merchants to sell through the marketplace either via FBA or by fulfilling their own orders. Merchant fulfilment is a virtual non-starter as it precludes merchants from actually being discovered in the first place. By definition it also precludes them from being Prime eligible - which means their products can’t be delivered with the lightening fast delivery times that Prime members expect. This runs counter to the virtuous cycle I described above. When a merchant instead uses FBA they run a gauntlet of various charges. This begins with selling plans (fixed annual fee or a fixed dollar fee per sale), but there are also referral fees (a percentage of the cost of the product), fulfilment fees (which include itemised costs for shipping, picking, packing, and storage), certain media products incur additional fees, and advertising charges are a de facto charge on 3P sellers who will need to advertise to access shoppers.

Selling plans experienced significant one-off upticks in pricing during the Pandemic but have been fixed since early 2024. Referral fees have also been ameliorated amongst a range of products. Traditionally referral fees were around the ~15% mark. In light of some of the issues we’ll get to later, some fee lines have been reduced to as low as 8%, while higher margin products incur referral fees up to 45%. Again, these have largely been frozen. In late 2023 and early 2024 fulfilment fees experienced a significant change. These changes were controversial because they segmented inbound and outbound service charges, and made estimating fees ex ante incredibly difficult.

A short summary of fulfilment fees is warranted to bring colour to the point I’m trying to make. These begin on a per unit shipped basis which covers customer service and most warehousing related services. These are broken down on the size of the product:

• Small standard (up to 4 oz): $2.50

• 15 oz: $3.21

• 1.5 lb: $5.13

• 2 lb: $5.37

• 3 lb: $6.04

• Over 3 lb (up to 20 lb): Add ~$0.32 per pound above 3 lb

Products priced under $10 receive a discount of about $0.77 off the standard fulfillment fee.

A new line of fees was applied to those products which are oversized:

• Small oversize (Large Bulky): $8.84 + $0.38 per lb above the first pound

• Extra-large: Base fee of ~$25 or more, plus higher per-pound surcharges for very heavy items

To top this off, apparel items incur higher fulfilment fees than non-apparel of the same size.

Next in line are the various categories of storage fees.