Amazon: Phenomenal businesses, but rich valuation means we’re taking some profits

Jan 30, 2019


Amazon is an amazing firm, with two phenomenal businesses inside of it: the online retail business that everyone knows (and most people love); and the Amazon Web Services business that has a lower profile with consumers but is more profitable and growing faster than the eCommerce business.

The challenge for investors is that the strength of these businesses is now very well-known by the investment community, with Amazon’s share price up more than 6x since the start of 2015. Amazon is now priced at more than 100x next year’s earnings, whereas the average for other tech titans Apple, Facebook and Google/Alphabet is less than 20x next year’s earnings.

We believe Amazon is likely to be a good (but not great) long-term investment from this point onward, despite its’ high current valuation. But investors need to be cautious: Amazon’s share price could drop materially in the short to medium term on the slightest bit of bad news.

We plan to scale back our investment in the firm, therefore, to hedge our bets. If Amazon’s share price continues to rise our small investment will continue to prosper; and if Amazon’s share price falters we expect that we will increase our investment substantially at the new lower valuation level.

Great business #1: Retail

It’s well known that Amazon is the leading eCommerce business in the United States and in many other OECD countries. There have been countless volumes written on Amazon’s success to date, with Brad Stone’s 2013 book The Everything Store (and note that link naturally goes to Amazon) among the best of them.

What’s less appreciated, however, is (i) the size of the opportunity that Amazon’s retail business still has in front of it, (ii) the magnitude of Amazon’s leadership and what that means for the potential profitability of Amazon’s eCommerce businesses, and (iii) how Amazon has backed up their leadership of the digital world with physical infrastructure that means their business’ economic moat is not only very wide, but also very deep.

Let’s dive in.

The size of the opportunity

Amazon’s eCommerce business has grown quickly since its’ inception in the late 1990s, and it continues to do so: North American eCommerce revenue grew 42% in the twelve months to June 2018.

It’s natural to assume that after twenty years of growth Amazon’s North American eCommerce business has reached maturity, and its growth rate will soon slow. We believe Amazon will not reach this saturation point any time soon, however, as the firm has consistently shown their ability to grow their potential market by re-shaping consumer perceptions about what we should be buying online: what started as books & CDs has grown into home furnishings, clothing, business supplies, and lots more.

To fully illustrate the scale of Amazon’s remaining opportunity, we need to jump into the data. The first point we should focus on is that after 20 years of strong growth online retail is still only 10% of total US retail sales. There’s lots of room left to grow this share as website technology improves and consumers become more accustomed to buying different categories of products online.

We expect online sales could easily be 25-30% of total retail sales in ten years’ time, as, for example, online sales are currently 23% of total retail in China and 19% in the UK. This leaves room in the market for Amazon to more than double their online retail business in the United States, even before we include the chances of Amazon growing their share of the online retail market, which they’re currently doing at a rapid pace.

In the past three years Amazon’s share of online has grown from 17% of the market to 24% (see chart), at least.

We say “at least”, because while these figures demonstrate how Amazon’s market share is growing, they probably understate Amazon’s true scale as they only count Amazon’s revenue from their third-party services business (which is reported in Amazon’s accounts), and not the retail value (Gross Merchandise Value, or GMV) that passes through their platform. If we adjust the data by assuming Amazon’s third-party revenue is only 15% of GMV, then we can see that ~50% of US eCommerce passes through

We expect Amazon will continue to grow their eCommerce business at a rapid pace in North America through both means that they have to date: online channels gaining share from offline channels, and Amazon gaining share within the online channel.

But that’s not enough for Jeff Bezos. Amazon is also going after brick & mortar retailers directly buy opening their own stores.

Amazon has entered grocery retail at scale with the purchase of Whole Foods, and it is developing both formats suited to convenience stores with amazon go, and a proposition for people who love to browse with amazon 4-star (that only sells products rated 4 stars and above on

And finally, there’s potential to grow the International business to the same level of success and market leadership as the US business. Amazon can continue to grow their international business through participating in online market growth, gaining share from weaker competitors, and eventually rolling out innovative retail formats. While it’s unlikely Amazon will ever replicate their success from the US in all their larger foreign markets, we expect they will be able to grow to profitability in Europe and other countries where amazon has a material presence, such as Singapore and Australia.

Amazon’s global retail business (North America & International) grew 25% in 2016, 33% in 2017 and should grow at a similar pace in 2018 (helped by the inclusion of Whole Foods). We’re confident that Amazon’s retail business can continue this rapid growth for many years to come.

We expect Amazon’s retail business to grow 25% in 2019, and then slow to 22%/20%/18% in the following years as the sheer scale of Amazon’s business weighs on Amazon’s percentage growth rates.

Amazon’s leadership and profitability implications

Amazon is not only growing quickly – it is miles bigger than its’ online peers. The second-largest US eCommerce platform is eBay, whose market share is 6.5%, which means that Amazon’s platform is ~8x the size of their nearest online competitor. This relative scale confers several advantages onto Amazon that have helped it grow its’ business even faster to date, and, we believe, will help Amazon improve its’ profitability over time.

First, most consumers choose their retailer for their low prices, which means the lowest cost (and usually largest scale) retailers typically have the most profitable businesses. Online retail is no different, where any investment Amazon makes can be amortized across 8x the sales base that eBay has, and even more for other retailers. Amazon’s scale in online retail helps fuel the firm’s flywheel (see diagram later), which ultimately sets Amazon up well for higher margins than their peers over the long run.

Second, Amazon’s leadership means that many consumers start (and finish) their search for products on Amazon, which helps Amazon grow revenue, and in some circumstances may make customers less price sensitive as they value the convenience of using Amazon as a one-stop-shop. Amazon’s prime offering, which builds loyalty through discounted shipping charges, further entrenches this behavior (and earns ~$10bn in revenue annually).

Amazon’s third advantage from market leadership is that the firm is now a gateway for many other firms to reach potential customers, so Amazon can extract a “toll” from these other firms for providing access to Amazon’s customers. Amazon has two main paths for earning this toll: through third-party fulfillment services and via their advertising business (where vendors can pay to promote their products higher up on the site’s rankings list).

These services businesses (third-party, subscription, and other retail) have been fantastic innovations for the firm as they’ve added another aspect to Amazon’s flywheel (see diagram later), and they’re also powering Amazon’s retail growth with a 51% YoY rise in Q2-18, and $47Bn of revenue in 2017.

In sum, there are many reasons to believe that Amazon’s North American retail business should be highly profitable, but to date it hasn’t been: segment operating margins have been <4% for the past seven years, well below the operating margins of other leading US retailers.

One could argue that these low margins are due to Amazon’s lack of scale relative to offline retailers. Wal-Mart has the highest margin of the food & general merchandise retailers, and their global revenue in 2017 was ~50% more than the ~$327bn GMV that passed through Amazon’s platform. Amazon’s margins are broadly in line with CostCo and Kroger, whose 2017 sales of $129bn and $123bn are more in line with Amazon’s (whose 2017 US sales were ~$100bn and GMV ~$200bn).

We disagree, however, and believe that Amazon’s low margins mask the retailer’s true profitability, as the profits from some business units are used to support earlier-stage businesses as part of Amazon’s quest for growth. We expect that many business units are earning operating margins of 5% or higher, and that overall margins could rise to the 8-10% range over the long term, more in line with the specialist retailers in the chart above.

For the record, at Cogent Alpha we’re completely fine with Amazon’s strategic approach of reinvesting profits into growth, as Amazon would not be the same firm it is today without such dedication to growth. We do expect, however, that retail margins will rise over time as more business units within the overall retail business reach maturity, and their profitability overwhelms Amazon’s investment in growth initiatives.

Amazon’s economic moat

Amazon has one further advantage from scale, and it may prove to be their biggest of all – the ability to develop their own logistics infrastructure. This infrastructure makes it much harder for upstart firms compete with Amazon.

Scale has long been a great defensive tool against small, innovative firms: while innovators may develop a great product, they couldn’t necessarily win market share from incumbents as their lack of scale meant they couldn’t deliver their product to customers for the same low price their peers could.

The advantages of scale have diminished in the digital economy, however, as software and other internet-enabled services can scale much more efficiently than physical products, as digital services have virtually no marginal cost or cost of distribution. This is great for firms that want to grow quickly, which many firms, including Amazon, have benefitted from. But it also removes some defensiveness for these businesses, as it makes it easier for their potential competitors to grow quickly to challenge them.

In a purely digital world the key formula for lasting success is to offer a fantastic service that locks customers in, potentially with network effects that means the more people use the service, the better it gets. Amazon has followed this playbook through having the widest selection, good prices, convenient delivery options, and programs such as Prime that encourage customer loyalty.

But Amazon’s gone further too. They’ve built the physical infrastructure required to deliver the products purchased through their platform in the most efficient manner, and therefore created a further competitive advantage of having the lowest delivery costs possible. This competitive advantage is very durable too, as it is built on physical assets, not digital ones. If peers wish to compete with Amazon directly on this front they would need to replicate the same physical infrastructure, which would take them years to build and cost more than $10bn. No other online retailer can afford such an investment, which means that Amazon’s leadership in online retail has a very wide and deep moat around it.

Amazon’s offline competitors, such as Wal-Mart, are better equipped to compete with Amazon’s logistics capabilities. They already have an efficient logistics infrastructure for moving products across the country, and the capital necessary to enhance this infrastructure for online retail as required.

But they’re constrained in other ways, as any move to promote online retail will be seen internally as competing against their legacy offline business. And despite the threat of online retail being well known for twenty years, very few (if any) primarily offline retailers have managed to successfully transition their mindset away from bricks and mortar to compete with the large scale, primarily online retailers. We expect this trend to continue.

Retail conclusions

The prospects for Amazon’s retail business remain very bright, despite twenty years of excellent growth to date. There is potential to more than double the size of the business, margins should grow due to many competitive advantages, and these advantages (and profitability) should endure thanks to Amazon’s foresight to invest in the physical assets to back up their digital advantages. The future of Amazon retail is very bright indeed.

Great business #2: Amazon Web Services

Amazon Web Services (AWS) is a newer business than Amazon’s retail business, but AWS may be more important than retail for Amazon’s long-term profits.

That may seem like a big call, given that 90% of Amazon’s revenue still comes from retail. But AWS is growing faster, at 46% annual revenue growth in the past year vs. retail’s 35% (including Whole Foods for the first time), and is much more profitable, with 26% operating margins vs. 1% for global retail (as North America’s +4% margins are offset by International’s -5%).

What is Amazon Web Services?

AWS is the world’s largest provider of cloud computing services, with ~34% market share (according to Synergy Research Group), which is more than Microsoft (~14%), IBM (~8%) and Google (~6%), combined.

“Cloud computing” is a term that is often referred to, but not necessarily explained well. We described Cogent Alpha’s view on cloud computing in our initiation article on Alphabet:

There are three general forms of cloud computing services: Software as a Service (SaaS) provides the software product directly to users who likely just log in to the software via a web browser; Platform as a Service (PaaS) allows for more customization by the client, who can build their own applications on top of the core service; and Infrastructure as a Service (IaaS) provides just the core computing components (such as server space) that the client must manage themselves. There’s a good description of the differences between these services here, and a great analogy using different ways of making pizza here.

Amazon largely invented the outsourced cloud computing business when they launched Amazon Web Services in 2006. AWS’ initial focus was on infrastructure: solving the problem that many corporations faced that they needed to have compute (mainly server) capacity for their potential peak demand, but it was expensive (and inefficient) to constantly own enough capacity to service those “potential peaks”.

The solution that Amazon provided was to allow corporations to rent capacity on an “as used” basis, and priced as such, in a similar manner to the way that utilities price their electricity supply. Amazon offered this service across many different building blocks of computing services, which Amazon called “primitives”. Amazon’s goal, as articulated by AWS’ CEO Andy Jassy in The Everything Store, was that they “tried to imagine a student in a dorm room who would have at his or her disposal the same infrastructure as the largest companies in the world”. That statement certainly encapsulated the demand for AWS’ services and helps explain how the business has since grown to generate $5.5Bn in annual operating income.

Why & how did Amazon create AWS?

AWS services can seem like a strange bedfellow for Amazon’s retail business, but it was borne out of traits that are uniquely Amazonian. Jeff Bezos deplores corporate bureaucracy, and so relatively early on he restructured the firm into “two-pizza teams” that are small enough so that when they’re working late they can be fed with…two pizzas (in practice this means ten people or fewer).

This “two-pizza” concept meant that many teams in the company work very independently, so it was natural that some teams become suppliers to others’ demands, which is what happened with AWS’ computing primitives. When legendary VC John Doerr questioned Jeff Bezos about AWS during a board meeting (“why would we go into this business”), Jeff Bezos allegedly replied “because we need it as well”.

This highlights the second unique amazon trait – they often launch businesses to external parties when they’re the first and best customer for that business. AWS is a good example, as is Fulfillment by Amazon (now third-party services), which let third-party sellers use Amazon’s logistics services in exchange for a fee. These strategic moves pushed more volume through AWS and Amazon’s fulfillment centers, lowering per-unit costs, and thereby speeding up the flywheel that powers the entire enterprise.

AWS’ prospects

AWS’ business today goes far beyond outsourced computing infrastructure. Amazon has extended beyond IaaS to PaaS, as they help clients to customize AWS’ capabilities for their own use. While this approach creates customer acquisition costs up front, it secures customers’ loyalty to AWS’ platforms as each firm builds more of their computing capabilities on AWS’ platforms.

The Cloud market is relatively young, as the transfer of corporate computing capabilities to the cloud is only ~30% complete, and it continues to grow rapidly. Gartner Research believes the cloud market will grow at >20% p.a. for at least the next three years, and primarily in IaaS where Amazon remains the largest market participant.

We expect that AWS will more than capture their share of this growth, and that AWS’ margins will increase as well as the business further benefits from economies of scale. AWS may be less well-known that Amazon’s retail business, but its’ prospects are at least as promising.

A word on AWS’ accounting

Most investment analysts agree with our view that AWS has fantastic prospects for growth. But there is much more controversy regarding the economic potential for the business.

The controversy stems from two sources: Jeff Bezos’ stated preference that investors focus on Free Cash Flow rather than accounting profits when analyzing Amazon; and Amazon’s use of capital & finance leases to acquire the computing assets needed to provide AWS’ services.

The challenge is that, under GAAP accounting methodology, the cost for the AWS computing capacity is allocated to “financing cash flow”, which means that it is not included within a normal definition of Amazon’s preferred “Free Cash Flow” measure of performance.

This makes a big difference to results: free cash flow for the past twelve months was $10.4Bn, whereas Free Cash Flow after adjusting for the leases was only $0.5Bn. And this has a huge impact on valuation. Under a standard definition for free cash flow amazon is valued at 90x cash flow; or at >1000x cash flow under the adjusted numbers.

At Cogent Alpha we prefer the second metric over the first as it better represents economic reality – but we also believe that it is not the best way to value the business (we prefer operating profits, for reasons we explain below). Amazon’s stock is expensive, but it is not THAT expensive.

Amazon’s Financial Prospects

Now that we’ve explained why we believe Amazon’s two businesses each have excellent prospects, let’s quantify the overall impact their growth is likely to have on Amazon’s financial results, before we turn to our assessment of prospective shareholder returns.

We expect Amazon will continue to grow revenue at >20% p.a. for the next three years, before settling to just below that benchmark by 2022. This revenue growth should be overwhelmingly driven by the North American retail business because that’s the biggest business today. We expect AWS will grows revenue faster, but it is starting from a much smaller revenue base today, so AWS’ growth has a much smaller impact on the firm-wide financials.

AWS’ growth has a much larger impact on Amazon’s overall profits, however, as its higher operating margins (AWS +26% vs. global retail’s +1%) more than offset its’ lower sales volumes. We expect Amazon’s profits will grow 37% p.a. over the next three years, as sales growth is amplified by slight margin increases in all business units.

In most instances this high level of sustained growth would be highly likely to produce excellent returns for shareholders, but much of this growth has already been bid into Amazon’s share price. We’ll assess prospective shareholder returns in the next section.

Note: Normally we would go beyond operating profits and assess cash flows and their growth to assess potential returns, as we described here. However, Amazon’s re-investment policy means that shareholder cash flows are very low right now, so growth rates mean little and it’s clearer to focus on profits.

Expected Shareholder Returns

Amazon is currently in a situation where it feels easier to predict shareholder returns over ten years, than it is for ten weeks or months.

The reason is that we’re very comfortable with Amazon’s businesses, and our view that their operating profits will grow at a very healthy rate over the long-term. But Amazon’s high valuation (chart below) means that shareholder returns over the short & medium term will be determined more by the whims of the market than the firm’s underlying prospects.

Note the Price/Cash Flow metric is generous to Amazon as operating cash flow includes the benefit of Amazon growing negative working capital in their retail business and doesn’t include the depreciation on Amazon’s AWS infrastructure, both of which inflate cash flows and reduce the P/CF ratio.

If Amazon stumbles at any point over the next few years, maybe after reporting a bad quarter’s results, or on the threat of government regulation, then its’ valuation could plummet. As investors in the stock – not the company – we need to be aware of that risk.

We believe the best way to approach as assessment of potential returns in these circumstances is to evaluate a range of scenarios, as we have done in the table below. Our “balanced” scenario is that we believe Amazon is likely to be a good but not great investment over the next three years, with prospective returns of 11% p.a. as Amazon’s 37% growth in operating profits/share is offset by valuation compression of ~18% p.a. This scenario assumes that Amazon’s valuation compresses from 75x operating profits to 40x profits over our holding period.

This table makes one point very clear to us: the biggest source of risk for Amazon’s shareholders is valuation, as valuation change can make the difference between -12% annual returns or +27% annual returns, or anywhere in between. And note that we believe these are all reasonable scenarios. Scenario (1) shows Amazon’s valuation reverting to being in line with other tech titans at ~20x EBIT (for a -12% annual return), and scenario 5 shows Amazon retain its high valuation as it has done for many years.


Amazon’s investors need to be comfortable with the fact that we invest in the stock, not the business. Even though Amazon is an excellent business, investors returns will largely be determined by the whims of the market not the firm’s underlying prospects, even if we intend to hold the investment for several years.

Therefore, we recommend that investors approach Amazon with caution. It is perfectly rational to sell the stock at its current valuation, despite the firm’s excellent prospects.

Our Cogent Alpha Portfolio has had a good run with Amazon, and the stock has more than doubled since we bought our position at ~$850/share in March 2017. We intend to reduce our holdings here – although we will maintain a smaller position – and re-deploy the capital into investment ideas where we have more confidence in the returns we’re likely to earn.


Disclosure: Warwick Simons owns shares in Amazon.