Losing My Amazon Religion

Benedict Evans asks a good question:

The growth curve is impressive enough, even before you notice that the scale is logarithmic. And to be sure, Amazon has had doubters from the beginning: few gave the company a chance when it started, and even fewer thought it would survive the bursting of the bubble. And yet, today Amazon is the king of e-commerce, at least in the United States, and the clear leader in cloud services, and now they are making a big push into digital content and devices.

I too was once a skeptic. I remember several years ago, after being invited for final interviews with Amazon, I decided that I wouldn’t take the job even if offered.1 Beyond the fact the retail-focused role wasn’t a great fit, I was also concerned that total compensation at Amazon, at least relative to other established tech companies, is heavily stock-based. At that time the stock was trading at a price-to-earnings ratio of nearly 90, which on the surface seemed unsustainable, and it didn’t seem worth the risk.

As of today, the stock price has doubled.

Since that time I’ve come to appreciate what an incredible business Amazon has built, as well as the size of the opportunity. Just about a year ago, when every one was freaking out about Amazon’s earnings (this week’s angst is nothing new), I wrote about Amazon’s Dominant Strategy:

Jeff Bezos’ critical insight when he founded Amazon was that the Internet allowed a retailer to have both (effectively) infinite selection AND lower prices (because you didn’t need to maintain a limited-in-size-yet-expensive-due-to-location retail space). In other words, Amazon was founded on the premise of there being a dominant strategy: better selection AND better prices – the exact same as Sears.

And, just like Sears, Amazon has added convenience. No, they haven’t opened retail stores; instead they created the amazing Amazon Prime.

In a happy coincidence, the same day I posted my piece the aforementioned Benedict Evans posted his own defense of Amazon:

Amazon is constantly creating new business lines. When they start, like any new business, they’re loss making. But they don’t ‘flip a switch’ to get to profitability – they just grow and execute, like any other business…

To put this another way, Amazon is LOTS of different startup ecommerce businesses on one platform. All the profits from the ones that work are spent on new, loss-making ones.

This approach makes sense of the compensation scheme I was nervous about: small autonomous teams have a lot of agency over their own results, even as they all work for a mutually shared outcome, and each team has a part to play. Older groups like books and media are the cash cows, funneling profits to growth engines like clothing or shoes or auto or any of the myriad of businesses under Amazon’s roof, all of them focused on the massive e-commerce opportunity (e-commerce is still only 6% of United States’ retail), and each doing their part to deepen the moat that is Amazon’s scale, logistics network, and multi-sided marketplace of customers, suppliers, and third-party merchants.

My confidence had been further bolstered by the approach Amazon had taken with Kindle: while their own devices had created the market, Amazon was quick to have a Kindle app immediately available on all platforms. Clearly they understood that e-commerce – and its digital equivalents – was a service business that needed to be in front of as many people as possible; to be overly focused on their own devices would be a mistake.

I could even see the point of the Kindle Fire tablet; at that point the cheapest iPad was $500, and Android-based tablets were even more expensive. Here came an alternative for half the price, and even in version one the seamless integration of media you owned, were subscribed to, or could purchase was brilliant. Amazon’s investments into digital media were understandable: their e-commerce business was originally built on books, CDs, and DVDs, but all of those were going away in favor of digital alternatives. Worryingly, Amazon was barred from selling said alternatives on iOS devices, so it made sense to offer an iOS alternative with their media stores fully integrated.

It’s on this point, though, that the Amazon narrative starts to break down, at least for me. See, while Amazon’s revenues keep going upwards, their costs do as well – as, indeed, they always have. Those costs, though, are increasingly not about e-commerce, but rather two areas in particular: devices and video.

Amazon’s financials are famously opaque, and the investments the company is making in streaming video rights, original programming, and their devices are spread around between Cost of Sales, Fulfillment, and Technology and Content. Complicating matters is that spending for Amazon Web Services falls in the latter bucket as well. Still, in 2013 BTIG estimated that Amazon would spend $500 million that year on video, and FierceOnlineVideo has put this year’s expenditures at nearly $1 billion. Last quarter’s 10-Q noted (emphasis mine):

The increase in cost of sales in absolute dollars in Q2 2014 and for the six months ended June 30, 2014, compared to the comparable prior year periods, is primarily due to increased product, digital media content, and shipping costs resulting from increased sales, as well as from expansion of digital offerings.

And, in the earnings call, Amazon’s CFO said spending on original content in particular would keep going up:

So video content, for example, we’re ramping up the spend from Q2 to Q3 significantly. And so it’ll be also a significant growth year-over-year. Keep in mind we have two types of content. We have licensed content. We also have original content. A lot of you have probably seen a lot of the announcements that we’ve green-lighted a number of pilots. We’re going to be in heavy production in those series that have been green-lit during Q3. We’ve also announced a number of pilots that will be in production on. And so that original content, it’s a portion of our total content, will be over $100 million in Q3.

It’s this focus on original and exclusive content – and devices that deliver it – that concerns me, and not because it’s expensive. Rather, what exactly does this have to do with e-commerce?

Best I can tell, Amazon’s story goes something like this:

  • Amazon gets or creates exclusive content at considerable cost
  • Customers are attracted to said exclusive content and thus sign up for Amazon Prime
  • Because customers are members of Amazon Prime, they start spending significantly more on e-commerce

Oh, and since mean ol’ Apple won’t allow Amazon’s digital content to be sold on iOS, Amazon will make devices to better sell said digital content. Which will ultimately accrue to e-commerce. And, profit!

I suppose this makes a certain kind of sense, but it reeks of what a former manager of mine calls a “double bank shot.” Amazon seems to be arguing that through this rather convoluted chain of events, all of which carry significant challenges and risks that are outside Amazon’s expertise (content creation, ecosystem development, etc.), they will be better placed to increase e-commerce’s share of retail.

Here’s my question: why not spend all that money – and time and executive attention – on simply growing e-commerce? Instead of pushing for the Prime Rube Goldberg machine, how about simply advertising Prime? And instead of pursuing a separate ecosystem, with all of the challenges and incentive risk that implies, why not focus on both building better apps and on creating partnerships with Apple in particular (who certainly has no intention of competing in e-commerce; Google is obviously much more of a competitor)?2

Moreover, I’m concerned about the internal incentives that Amazon is creating for itself. Amazon is increasingly competing with its suppliers, particularly in the digital space, and I just noted that potential partners like Apple are instead rivals. More concerning is the effect of devices on the company’s overall strategy. In the Fire phone introduction, Bezos was very clear that any device needed differentiation. His answer was Dynamic Perspective, but when that doesn’t work – spoiler: it doesn’t – the easy fallback is to differentiate on services. This will be particularly tempting given that Amazon is clearly looking to make a profit on each device sale. I can’t overstate what a terrible idea this is; I hate the Fire phone not because it seems to be a terrible product, but rather because I see its very existence, at least in the current incarnation, as actively harmful to Amazon’s core business, which needs to be great on all devices.

So what exactly is going on? Why is Amazon building vertical devices that don’t fit a horizontal company? Why are they pursuing convoluted double-bank-shot strategies that are extremely expensive and high risk? All of these are much more pressing questions than why Amazon is or isn’t making a paper profit.

  • The first possible explanation is that Amazon’s core e-commerce business is in fact very threatened by the shift to mobile, and they feel they have no choice but to build their own platform with its own lock-in.

    The biggest problem with mobile is that, according to Michael Mace, while PC shoppers convert at about a 3% rate, mobile shoppers are a mere 1%. That’s a massive drop-off. Moreover, the rise of apps as a primary channel makes vertical and brand-specific retailers just as accessible and visible as Amazon (in contrast, few people go directly to specific web sites).

    In addition, the sort of shopping experience that Amazon is particularly strong in – extensive research, reviews, etc. – simply doesn’t work as well on mobile. What does work well are things like flash sales or direct marketing pitches. These new marketing and conversion channels threaten to break into Amazon’s share of wallet: you may not necessarily have bought an item purchased in a flash sale from Amazon, but you do have that much less to spend for the rest of the month.

  • At the other extreme, it’s possible that Bezos simply wants to rule the world, at least when it comes to buying and selling anything that can be bought or sold. It’s certainly hard to doubt the guy!

    Still, the Fire phone in particular gives me pause. Beyond the incentive issues I noted above, rumor has it that Bezos was very involved in the Fire phone’s development process, and that he fashions himself as a product guy. The way-too-long introductory keynote certainly hinted at a certain grandiosity about the phone and its development process.

    The problem, though, is that the phone is simply not good. Bezos is clearly an operational and organizational genius, but there is nothing in Amazon’s history to suggest he is a product person.3

These are two rather unappealing possibilities from an investor perspective: a rational response to a mortal threat, or an irrational belief Amazon can seize a seemingly non-existent opportunity. And still the question remains: what’s wrong with simply focusing on the massive e-commerce opportunity? Again, I don’t mind that Amazon doesn’t make profits; I love the way they are constantly building new businesses from scratch. But why can’t those new businesses leverage Amazon’s strengths instead of accentuating its weaknesses?

Or, perhaps there remains truths I still do not fully understand. I was wrong about Amazon in 2010, but my error was one of depth; once I took the time to understand the company, I became a believer. This time though, feels different: I want to give the benefit of the doubt, but I don’t believe in double bank shots or blind faith.

(Check back in 2018 when I write about how I was wrong).

Update May 2015: I was wrong, and it only took me a year: see The AWS IPO

  1. I try to make decisions before I have to 

  2. To be clear, this ship has long since sailed; I’m referring to what Amazon could have done instead 

  3. An earlier version suggested that no one at the company was a product person; that is obviously not true and was unfair. What I meant to say is that Amazon’s specialty is not finished physical products; rather, they are a services company that improves iteratively. There is nothing wrong with this (it’s something that Apple, for example, is terrible at)