Amazon

Amazon keeps popping up on a value screen I run, but I’m having a hard time seeing that value in Amazon. My contention is that Amazon is really five different businesses: an online retail behemoth, a subscription based entertainment service, a logistics company, AWS, and advertising services. The way that Amazon reports its segments is regionally (North America and International) as well as breaking out AWS. The retail, subscription, and logistics services are all reported collectively but really should be broken out separately within each region to get a grasp of the value of its parts.

Let’s start with AWS, which many cite as a sufficient reason to invest in Amazon. It looks like a great business on paper. Sales were $62.2 billion last year accounting for just over 13% of net sales. Operating profits for AWS were $18.5 bil after operating costs of $43.7 bil, resulting in a not too shabby operating margin of 30% and accounting for a whopping 79% of Amazon’s operating profits! Of course we don’t really know what net margins would be if Amazon spun out AWS because a lot of corporate costs are consolidated.

After AWS, things get complicated. North American sales account for 59.5% of all sales, with International accounting for just over 27%. Now they do break out net sales by online stores, physical stores, 3rd party sellers, subscription services, advertising services, and other, however they don’t break it out by region. If we consolidate non-AWS revenues we get $553.4 bil in sales of which $222.1 bil is through the online stores and $17.1 bil is in physical stores. Another $103.4 bil in sales is generated through 3rd party transactions. This allows us to isolate $31.8 bil in subscription services and another $31.2 bil in advertising services. What is missing from this breakout is the logistics business, as well as the entertainment revenues and costs, which are being folded into the revenues and costs of the subscription services. While we can estimate that Amazon Prime delivery and Amazon Prime video are generating $32 bil in revenue, we have no idea what the costs of these services are to Amazon. What we do know is that the operating income from all of the non-AWS businesses is $7.4 bil.

On the cost side of the equation Amazon breaks out costs of sales at $272.3 bil, fulfilment costs (warehouse and logistics) at $75.1 bil, technology and content at $56.1 bil, marketing at $32.6 bil, and GA at $8.8 bil. I think it is safe to put cost of sales and fulfilment costs with consolidated retail sales. The tech and content costs include AWS, which we know was $43.7 bil, so about $12.4 bil of tech and content costs need to be assigned somewhere else? Amazon Prime Video? Same with advertising costs. A share of ad costs are attributable to AWS for sales commissions, but the distribution of these costs across the various businesses is a mystery.

The bottom line is it would appear that AWS is the beating heart of Amazon’s profitability. With operating margins of 30%, AWS accounted for almost all of Amazon’s operating profits. Excluding AWS, Amazon earned just over $6.3 bil in operating income on sales of $407.6 bil, for an op margin of 1.6%. The retail business–consisting of online, 3rd party, and physical stores–would appear to be a money loser if you include the cost of sales and fulfilments costs (never mind including a share of marketing and tech costs as well). We have no idea what the margins might be for a stand alone Prime Video streaming service, but I would guess there is some money being made there given the loss leading economics of the retail business.

Given the byzantine operational structure of Amazon, I don’t see how anyone can value it in a meaningful way.

PP

18 Likes
Sales	 Costs 	         Op Margins	% Op Margins
AWS	                           $62,202.00 	 $43,670.00 	 $18,532.00 	29.8%
Consolidated non-AWS	           $407,620.00 	 $401,273.00 	 $6,347.00 	1.6%
Consolidated Retail	           $342,516.00 	 $272,344.00 	 $(4,939.00)	
Fulfilment Costs		                 $75,111.00 		
Prime	 $31,768.00 			
Tech and Content Costs net AWS	   $12,382.00 		
Advertising	                   $31,160.00 	 $32,551.00 		
Other	                           $2,176.00 	 $8,885.00 		
				
	                           $469,822.00 	 $444,943.00 	 $24,879.00 	5.3%

The retail, subscription, and logistics services are all reported collectively but really should be broken out separately within each region to get a grasp of the value of its parts.

But why is it important to grasp the value of its parts? The parts fit together synergistically. They’re not independent business lines. The enterprise is worth the value of its expected future discounted free cash flows, with the important proviso that, just like Berkshire, Amazon plows a big hunk of its profit back into business investments that generally pay off handsomely.

Don’t take my word for it. Read the M* report, for example.

Also, you don’t mention the substantial advert revenues, which are almost all gravy on top.

I don’t own Amazon at the moment. (I sold just before the most recent Q report to safeguard a good profit in my Roth account.) But I’m looking to get back in.

1 Like

Given the byzantine operational structure of Amazon, I don’t see how anyone can value it in a meaningful way.

Amazon is an interesting Rorschach test for value investors. The headline P/E is always elevated. Dow Theory’s quantitative Quadrix evaluation system gives it a value score of 9 on a scale of 1-100.

Seven years ago, after it appeared and disappeared from Tom Gayner’s equity portfolio at Markel, I got a chance to ask him about it. Gayner is a Warren Buffett disciple and emulator who identifies as a value investor, but he is considerably younger and has exhibited more willingness to resist traditional quantitative value analysis when it comes to evaluating digital disruptors.

He said he had taken a small stake in Amazon because he thought it was an important company, then availed himself of a quick profit on a short-term surge because he didn’t really understand it.

Gayner said the headline P/E was mostly an accounting illusion. If Amazon expensed its many investments over years, the way more traditional companies did with plant and equipment, profitability would look better. The fact that it often didn’t produced, among other effects, vanishingly small tax obligations.

Today, Amazon is a 3% position in Markel’s equity portfolio, seventh-largest, behind Berkshire, Brookfield, Alphabet, Home Depot, Diageo and Deere.

The synergies among its various businesses, as you note, are virtually impossible to quantify. The streaming service might look profitable because of all the Prime subscription revenue, but that’s a function of the logistics business. Nobody I know subscribes to Prime for the streaming service. And yet, it is the new home of Thursday Night Football, making it indispensable for the legions of U.S. NFL addicts.

Similarly, the loss-leading online retail business is indispensable to millions because of the logistics business. It is as if they operate on a different planet when it comes to speed, reliability and ease of returns.

I recently moved and ordered something from Amazon I thought would suit my new digs. It turned out it did not. It arrived two days after ordering. That same day, I took it to a nearby Amazon fulfillment center, scanned the QR return code and put it in the locker that popped open. In the 10 minutes it took me to return home, the return had been credited and a refund issued.

About the same time, I ordered something from Staples, which has an above-average logistics operation, particularly around speed of delivery. I neglected to update the delivery address. Delivery was refused at my old address and the item returned to Staples. It took me two weeks, interacting with a variety of online chat assistants and call center employees, to get a refund issued.

This gathering indispensability across a variety of verticals, backed by the largest cloud computing network in the world, seems to me much harder to establish than monetize. It is not easily susceptible to quantitative analysis, but not everything is. There is a qualitative aspect to investing as well.

Gayner’s view that Amazon is an important company has only been confirmed in the years since. If you look back 20 years or so, Amazon was discussed in value forums the way SaaS and Tesla are now. In the digital age, value is a more complex idea than ever. Gayner is an example of a value investor with an accounting degree who is willing to incorporate qualitative factors into his analysis. In the case of Amazon, it’s worked out pretty well.

25 Likes

My contention is that Amazon is really five different businesses: an online retail behemoth, a subscription based entertainment service, a logistics company, AWS, and advertising services.

You can cleanly separate AWS, but can you?

Most of the features of AWS are built for Amazon. Let that sink for a minute. AWS has over 240+ cloud services. Outside of GOV and DISA cloud services, all other services were primarily created for Amazon.com. This starts with sharing compute, storage, network, database, ability to scale, all of analytics, even their AI/ML (Hello Alexa, for those who remember Alexa was termed as money down the drain), all of their automation, requiring to scale Amazon.com website, content delivery developed first for Amazon.com, then for Netflix. When Larry said Amazon.com runs on Oracle Database, Jezz Bezos demanded to get out of Oracle in 18 months and AWS, Amazon did it. By then Amazon’s database is one of the largest Oracle database implementation. At that time Oracle had 35 years of RDBMS product development, selling, testing, supporting, experience, code, IP, yada, yada…

AWS came up with a family of Database offering to replace Oracle. For those in Software product development can comprehend what an achievement it is/ was. Today, you can run AWS as a separate company, but still the engine of innovation at AWS is powered by Amazon.com requirement.

The point I was trying to make is Amazon is a unique company where the sum of the parts are bigger than the parts.

Prime, is an eco-system by itself, Prime is an annuity engine, that subscription revenues falls directly to the bottom line, likewise, all of the ad has a very high profitability and it is $30 B+ growing business. Remember prime subscribers get video and audio for free, they pay for the free delivery. So if Amazon introduces Ad’s for prime members, you are not going to see attrition, rather, Amazon will offer some content ad free and some supported by ad, and over time increase ad supported content. Today logistics is bigger than FedEx, imagine that. Recently I posted about One Medical acquisition, just one stroke one Medical has access to 60 million Amazon subscribers. This optionality and ability to organically build massive companies within, fund it internally is what you are seeing the valuation.

Berkshire generates cash, invests in equities, and buys operating companies. Amazon generates cash, invests to build organically.

10 Likes

Berkshire generates cash, invests in equities, and buys operating companies. Amazon generates cash, invests to build organically.

Really good post, Kingran. Long ago I was a skeptic of Amazon as a business, but I’ve since come to be amazed by its ongoing growth and reach. I don’t own any yet, but I do want to own it.

As Charlie commented during an interview within the last few years, Amazon has a LONG runway ahead of it. Bezos is an incredible visionary and manager.

Bezos is an incredible visionary and manager.

FYI: Bezos has left Amazon.

tecmo

“Bezos is an incredible visionary and manager.”

FYI: Bezos has left Amazon.

Thanks, tecmo. I see he is no longer CEO, but he is still Executive Chair, Amazon’s largest shareholder, and apparently still in close touch with CEO Andy Jassy and involved in major strategic decision-making.

Recent story: https://www.businessinsider.com/jeff-bezos-still-focused-on-…

So, Bezos’ leadership doesn’t have the importance it once did, but he is still part of the company and an important asset.

1 Like

Bezos’ leadership…

His leadership principle is enshrined into Amazon. When amazon hires, it is these leadership principles that carry far higher weight than the ‘technical’ or ‘functional’ knowledge related to the job.

https://www.amazon.jobs/en/principles

The day-one mindset is still very much alive.

https://aws.amazon.com/executive-insights/content/how-amazon…

Bezos has left the company a framework, and it will guide him beyond his day-to-day involvement.

But why is it important to grasp the value of its parts? The parts fit together synergistically. They’re not independent business lines. The enterprise is worth the value of its expected future discounted free cash flows, with the important proviso that, just like Berkshire, Amazon plows a big hunk of its profit back into business investments that generally pay off handsomely.

Also, you don’t mention the substantial advert revenues, which are almost all gravy on top.

I believe it’s the advertising revenues which are the starkest proof. The margin on that is likely astronomical. They don’t have to employ a lot of workers running around a warehouse, or even going door to door (like radio, TV, newspaper). They open up the gates (algorithmically speaking) and people give them money for placement. (There are similarities in other businesses: vendor support advertising in traditional media, slotting allowances in retail but those are labor and time intensive getting it all put together between manufacturers, ad agencies, media, planners, retailers.)

And yet without the enormous “store”, Amazon’s advertising revenue would be zero. Who would advertise on a website with no traffic? Who would pay a premium to reach people shopping for “a hammer” if there were no hammer shopper eyeballs? Traditional media has to produce a product, usually expensively: music on radio, TV shows on television, newspapers for print, etc. Amazon needs to produce nothing extra - (or everything, depending on your point of view) to garner millions in ad revenue.

So how is that expense allocated? I wouldn’t be surprised to see margins of 90%, even greater for the “ad business”, and yet that’s not fair either. How to value? Impossible, really. It’s a bolt-on business with zero value to someone without a web store, and pure value to Amazon. AWS is a bit less so, but not much less. How do you pull logistics out without admitting it’s one of the legs the company stands on?

This is a business where the sum of the parts likely doesn’t come close to figuring out what each part is worth. What’s a steering wheel worth? A lot more if the car is going 60 than if it is at rest. Some things are holistic, I think Amazon fits that category nicely.

24 Likes

Another value of Amazon not reflected in financial reports is the scalable business model. They have been steadily extending to new businesses with existing web and logistics platform. They just entered healthcare by acquiring One Medical. Why wouldn’t I use their service if they can deliver medicine in one day at lower price?

And yet without the enormous “store”, Amazon’s advertising revenue would be zero. Who would advertise on a website with no traffic?

It’s a good point, and applies to most of the moving parts within Amazon.
The connections among streaming, Prime, shipping costs and low-profitability retailing are inextricable.

But I would argue that AWS is quite different.
There is no massive integral joining with the rest of the business.
Sure, I’m sure they use some of the same tech support and infrastructure, but then so do all of the customers of AWS.
AWS could easily be a separate business.
By extension, there is nothing wrong with a sum of parts approach to Amazon which considered “AWS” and “everything else”.

That doesn’t mean it’s easy.
The trick is trying to slice the rent and depreciation and labour expenses roughly in the right ratio.

Speaking of parts which could be summed—
One interesting question is the future of the ex-North-America retailing business.
I gather it has always lost money.
As a customer I’m a fan, but it seems their solution to that problem is to gradually scale back on the direct selling and become more of a pure marketplace.
They need to get their vendor reliability as high as it is at Ebay.
There is much more of a case for tightly bound synergy between PrimeUS and RetailUS than there is between RetailEurope and RetailUS.

Jim

6 Likes

They need to get their vendor reliability as high as it is at eBay

Appreciate what everyone’s said here, and agree with most.

But one thing that seems perennially under-addressed in Amazon discussions is that sometime a few years back they lost control of their storefront, their recc system, their vendors.

And for a business that has been perpetually proactive, their sluggish response here has been striking. The shady Chinese-knockoff stuff is, if anything, worse year-over-year.

A warning that they’re getting complacent in what’s supposed to be a core competency? Or a fluke? Or are the best of the young executives all jostling to work in a shinier division?

–sutton

11 Likes

Mihir A. Desai points out some of Amazon’s unique advantages in his book “How Finance Works”
Amazon actually manages its inventory, receivables and payables in such a way that inds up with what’s called a negative working capital cycle, or negative cash conversion cycle. . . .in 2014 Amazon averaged fourty-six days of inventory, and it collected from its customers after twenty-one days on average . . . The icing on the cake is that Amazon due to its market dominance , can exert a large amount of power over its suppliers to make them wait before getting paid, and it averaged ninety-one days to pay suppliers.
Desai uses Amazon as an example many times throughout the book.
I personally have a love/hate relationship with Amazon. Love their customer reviews but recognize there are often fake reviews mixed in. The only place you can quickly get many items, it’s actually easier than getting in the car and going to the stores only 5 miles away. I get more than half the items I order in 1 day. The bad they don’t control their vendors, there are many counterfeit, substandard items in the mix.

RAM

1 Like

…sometime a few years back they lost control of their storefront, their recc system, their vendors.
And for a business that has been perpetually proactive, their sluggish response here has been striking. The shady Chinese-knockoff stuff is, if anything, worse year-over-year.
A warning that they’re getting complacent in what’s supposed to be a core competency? Or a fluke?
Or are the best of the young executives all jostling to work in a shinier division?

It is a bit of a mystery.
But the neglect of the basic retail process is very much in keeping with the age-old mystery of why their search engine is so bad.
I understand why in recent years the results are sorted almost entirely based on which vendor paid the most, not what matches your query best.
A bit annoying, but at least they have a good monetary reason for that.
But a basic functioning search ability eludes them, and always has–since long before they were focusing on the ad and placement dollars.
For example, prioritizing things that match the largest number of your keywords, allowing a quoted phrase.
I find one has to use Google to find something at Amazon so often that there is little point starting at Amazon even if you know that’s where you want to buy.
As I say, a mystery.

Jim

16 Likes

I find one has to use Google to find something at Amazon so often that there is little point starting at Amazon even if you know that’s where you want to buy.
As I say, a mystery.

This is true of so many products. I use google rather than built in search when using
Windows
PCs
Microsoft Office including Word, Excel, PowerPoint, OneNote
Mathworks products including Matlab, Simulink

and I’m sure a host of smaller apps I use less often without thinking about.

I almost wonder if Google could build a more directed product around this phenomenon, and be integrated into apps that want better search than they can build.

R: