Amazon is the exception to nearly every rule in business. Rising from humble beginnings as a Seattle-based internet bookstore, Amazon has grown into a propulsive force in at least five different giant industries: retail, logistics, consumer technology, cloud computing, and most recently, media and entertainment. The company has had its share of missteps — the expensive Fire phone flop comes to mind — but is also rightly known for strokes of strategic genius that have put it ahead of competitors in promising new industries.
This was the case with the launch of cloud business AWS in the mid-2000s, and more recently the surprising consumer hit in the Echo device and its Alexa AI assistant. Today’s Amazon is far more than just an “everything store,” it’s a leader in consumer-facing AI and enterprise cloud services. And its insatiable appetite for new markets mean competitors must always be on guard against its next moves.
As the biggest online retailer in America, the company accounts for 5% of all retail spending in America, and the company has been publicly traded for two decades. While its market capitalization has swelled recently, so too have expectations. Wall Street banks like Morgan Stanley expect Amazon to continue growing at a rate that no company its size has ever done before: 16% average compound growth in sales through 2025. If Amazon were able to satisfy the lofty goals, it would be “the most aggressive expansion of a giant company in the history of modern business.”
Understanding the many-headed beast that is Amazon is no easy feat, especially because Amazon is far less transparent than its peers. As the Times has written, “It isn’t just secretive, the way Apple is, but in a deeper sense, Jeff Bezos’ e-commerce and cloud-storage giant is opaque. Amazon rarely explains either its near-term tactical aims or its long-term strategic vision. It values surprise.”
By all accounts, Amazon is just getting started in newer initiatives like cloud services, artificial intelligence, and logistics. Given Amazon’s enormous breadth, we won’t be covering every aspect of its business. But some of the main takeaways from our analysis include:
- Amazon’s latest raft of acquisitions could indicate more hunger: Amazon had a large uptick in M&A in Q1’17, buying Harvest.ai, a cybersecurity player; Do.com, a meeting productivity software; while also buying its way into a new geography with Souq.com, a Middle Eastern e-commerce site. This is out-of-character behavior given the company’s generally more conservative M&A history and could mean Amazon is shifting to a more proactive stance to fuel its AI and enterprise ambitions.
- Amazon’s next pillar is likely to be AI: In a letter to shareholders published in April 2017, Bezos wrote extensively about AI and machine learning as a focus of new company efforts to maintain relevance and its edge over the competition. Voice, virtual assistants, and natural language processing will continue to be a focus. But Amazon is also focused on AI-as-a-service and putting the basic tools of AI in the hands of its cloud computing and developer community. More than ever before, Amazon has aspirations to become a platform company.
- Amazon’s interest in GRAIL may foreshadow healthcare AI interest: Its investment into GRAIL was a vote of confidence that genomics, with its massive data and processing needs, will be a major area for computing. Amazon’s existing tools for big data and AI mean it’s well positioned to enter the healthcare arena, where many AI startups are already proliferating.
- Amazon’s corporate venture arm, the Alexa Fund, has nurtured the developer and hardware ecosystem around Alexa as a universal AI assistant: The Alexa platform offers SDKs which allow third-party developers to build skills for the AI assistant and other manufacturers of hardware to integrate the Alexa assistant into their products. The Alexa Fund’s investments also point to new interfaces — like gesture controls championed by Thalmic Labs — as well as hardware category possibilities, like robotic companions, such as those developed by Embodied.
- The company is also making more diversified investments into logistics, cloud apps, and media: Amazon’s recent forays into logistics and media foreshadow areas of new business interest. Amazon tends to invest mainly where it can make strategic partnerships. India-based Housejoy will help expand its reach in the region, and Twilio has partnerships with AWS.
- Secretive R&D skunk work Lab126 is behind Amazon’s recent consumer tech hits: The secretive Silicon Valley-based R&D lab is behind hardware hits like the Echo and Kindle. And although it was also where the ill-fated Fire phone was developed, it is an under-appreciated example of Amazon’s internal dedication to innovation.
- Next-generation logistics is a centerpiece of Amazon’s R&D: Nearly 80 of Amazon’s 2016 patents are focused on developing its logistics network, which is already far more than just a few years ago. Amazon also patents heavily for its cloud computing and cybersecurity efforts.
- Amazon has also raised its profile in consumer goods and physical retail: Amazon operates its own shoe lines and apparel brands, as well as consumer goods grouped under its Amazon Basics label. It has begun opening brick & mortar bookstores, and has launched its Amazon Go, check-out free, convenience store concept. It’s conceivable that Amazon Go could become a licensable white-label solution for retail tech. Certainly, Amazon is putting more pressure on traditional retail than ever.
Table of contents
- A little background on core Amazon
- Patent data analysis
- Amazon initiatives by sector:
- Final words
Background on core Amazon
Jeff Bezos, the company’s founder and long-time CEO, first hatched the idea for Amazon while working on Wall Street at the hedge fund and tech private equity group D. E. Shaw & Co. For a while, it was bootstrapped as an internet bookstore with Bezos’ money along with contributions from friends and family. In 1995, Bezos raised nearly $1M in small checks from 20+ local angels with a typical check size between $30k and $50k. Among those angels, Nick Hanauer, Eric Dillon, and Tom Alberg (of Madrona Venture Group) were brought on as company advisors. Finally, in 1996 Bezos sought outside investment from John Doerr of Kleiner Perkins Caufield & Byers. In Amazon’s only round before IPO, KPCB invested $8M at a $60M valuation for a 13% stake. In 1997, Amazon went public at a $381M valuation. Twenty years later, as of March 2017, its stock price is up 35,000%. Its market capitalization hovers around $419B.
Over the past two decades, the Seattle-based company built an e-commerce-centric business that now appears to be at an inflection point. In a recent interview Bezos said Amazon rests on 3 pillars: Amazon Prime, which offers membership e-commerce bundled with elite digital media products; Amazon Web Services, which leads the tech pack in cloud computing; and Marketplace, its third-party seller business. Bezos has mentioned there are several new pillars in the works, but added “ask him in 10 years,” on how they will pan out.
Amazon’s newer, possibly “pillarable” initiatives — including the Alexa platform for voice-enabled apps or “skills” — are aligned with the core e-commerce business and are already beginning to pay dividends by enabling even more frictionless commerce. Amazon’s customers can already order and purchase items directly through Alexa, and Prime members can access exclusive discounts and content through the platform.
In other words, success in the newer bets has cracked open new opportunities, but Amazon must defend and build on its new businesses. Amazon can only do so if it continues to innovate faster than rivals Apple, Google, Walmart, and Alibaba, among others. No one in big tech wants it to own e-commerce and the cloud like it does, and even its China-based e-commerce rivals are stealing from Amazon’s playbook and unrolling competing cloud services.
The other big question is if the famously customer-obsessed Amazon can successfully transition to becoming a conglomerate with toes in vastly divergent business models. Amazon’s newest pillar with AWS is a business-to-business product, as are many of its newest tools and services like voice and AI-as-a-Service.
Presently, Amazon is comprised of nearly 90 business entities held across the world, many of which are affiliates of its e-commerce and media businesses. Among these, Amazon has separate retail websites for the United States, the United Kingdom and Ireland, France, Canada, Germany, Italy, Spain, Netherlands, Australia, Brazil, Japan, China, India, and Mexico. However, some are subsidiaries — many of them the product of acquisitions — focused on specific categories, like Audible in audio books or Zappos in shoes.
To see where Amazon is adding human capital, we looked at the company’s open job postings. Here’s a graphic showing the proportions of open jobs by vertical and team.
Clearly, AWS is the biggest area Amazon is scaling up with more than 5600 jobs, which translates to about 33% of all the open listings. Fulfillment & Operations was the next largest hiring area, representing about 19% of open positions. Some surprising cohorts were the Alexa Team, with more than 890 jobs, making Alexa nearly 5% of all Amazon’s open positions. And the Amazon Devices team, which includes the recommendation algorithm team MAKO, also accounts for about 5% of job listings.
Amazon has earned a reputation as a conservative M&A player, but the tide may be turning. Doing 4 deals in Q1’17, the company’s acquisition tally matched a record quarterly high set in Q3’15, and outpaced activity in the previous two quarters. The sudden uptick in M&A is notable, given the company has rarely done more than 2 deals per quarter.
That said, Amazon’s appetite for M&A appears to ebb and flow. Looking at the acquisition history by year, activity clearly cratered during the dot-com bust, then rebounded through 2015, and has fallen again in the past two years. In 2016, Amazon’s M&A was on the downswing year-over-year, and according to the company’s annual 10K filing, Amazon only spent $103M on acquisitions, compared to $862M in 2014 and $690M in 2015.
Amazon’s lack of recent interest in high-flying, aka expensive, startups might be due to a culture of conservative investment. For example, Nat Burgess, a mergers-and-acquisitions specialist at TechStrat remarked that Amazon had a good business case for acquiring Twilio to strengthen AWS’ offerings, but likely balked when Twilio went public at a valuation that was 16x revenue. On Amazon’s general approach to M&A, Burgess also suggested that the company’s strategy hinges on fulfilling specific needs instead of wholesale buying their way into markets:
“Amazon is a conservative buyer. They think long term and they don’t get seduced by high-flying valuations….Amazon is unlikely to overpay for a high-flying, fully baked platform as the basis for the next dreamy business. They are more likely to fill gaps through smaller deals, which makes M&A less central to their strategy than it is to a company that expands to entirely new markets through acquisitions.”
Adding to that, it recently came to light that Amazon was mulling over a $10.7B acquisition of Whole Foods, which would have immediately invigorated its push into groceries. The deal “would turn Amazon into a grocery giant overnight and help it sideline Instacart,” but evidently Amazon shelved such a bold move, which would have been uncharacteristic for the company.
The data back up the idea that Amazon historically moves more cautiously in the M&A arena. Compared to its tech giant peers, Amazon is less acquisitive and did only 5 deals in 2016. Apple, for reference, has recently done between 8 and 14 M&A deals per year. Facebook‘s activity, while trending downward more recently, once hit back-to-back years of 14 deals. Google is certainly an outlier in M&A activity, since it is particularly active, but in 2014 it did 35 deals where Amazon only did 5.
What, exactly, does Amazon look for in a potential acquisition? Jeff Bezos himself outlined what constitutes a must-have “dreamy business” in a 2015 shareholder letter:
“A dreamy business offering has at least four characteristics. Customers love it, it can grow to very large size, it has strong returns on capital, and it’s durable in time – with the potential to endure for decades. When you find one of these, don’t just swipe right, get married.“
Some of Amazon’s largest acquisitions appear to fit those four criteria. In the company history, its largest deals were shoe retailer Zappos ($1.2B, 2009), e-sports streaming site Twitch ($970M, 2014), and warehouse robotics maker Kiva Systems ($775M, 2012). And, years later, these are still fast-growing, significant parts of the company. Much of the tech industry was skeptical when Amazon bought Twitch, but analyst Gene Munster expects the subsidiary to be worth $20B and generating $1B in revenue by 2020. Additionally, Kiva’s robots have helped cut operating expenses in fulfillment centers by 20%. While Zappos is still reeling from experiments with its org structure, it did manage to hit its 2015 profit goals, according to Fortune.
While Amazon’s biggest-ticket M&A deals helped establish itself in new markets and technologies, undoubtedly the recent theme has been about fortifying AWS offerings. A number of Amazon’s five 2016 acquisitions appear to reinforce its blossoming cloud services business: Italian startup NICE made software for technical computing, and Cloud9 IDE made a collaborative development platform. All of these help AWS cater to developers and become the go-to place for deploying code.
Amazon’s M&A spree in 2017 echoes the same desire to bolster AWS, along with a newer desire to expand overseas. The expansion effort was exemplified by the recent acquisition of Souq.com, the so-called “Amazon of the Middle East” that was bought for somewhere between $650M and $750M. The massive purchase will allow Amazon to expand its e-commerce footprint into Egypt, Saudi Arabia, and the UAE. The deal comes at a time when Amazon’s overseas efforts, particularly in Asia, are beset by intense competition.
There is well-funded competition like Flipkart in India (which was launched by two former Amazon employees in 2007) along with the China-based giant Alibaba, which has been acquiring in other markets including Lazada in Southeast Asia. Amazon’s competitors seem to be banding together to win certain markets, too. Flipkart recently raised $1.4B from Tencent, eBay, and Microsoft, which all compete with Amazon in varying ways.
But the remainder of the Q1’17 deals fall into the purview of AWS. Cybersecurity startup Harvest.ai was likely bought to strengthen cloud offerings. Digital video, which makes up a lot of the data stored on AWS, has also been a focus with content creation toolmaker Thinkbox Software (which will build on video editing assets like Biba Systems). And it’s been speculated that Amazon acquired enterprise meeting productivity tool Do.com to roll into AWS’ new Chime initiative, which is a videoconferencing suite for business.
Since 2010, Amazon has favored buying early, acquiring 10 Series A stage companies, followed by seed/angel stage companies:
It’s also worth noting that Amazon’s move in buying UAE-based Souq.com was a new area of geographic interest. Historically, almost all of the M&A deals have been to companies based in the US.
In summary, Amazon mainly acquires value-adds to its main business pillars Prime, AWS, and Marketplace. There are some blockbuster deals to “dreamy businesses” that have scaled up well, but more often Amazon uses small, practical purchases to develop. After all, this is the same company whose CEO drove a Honda and proudly made new employees fashion desks out of doors as “a symbol of frugality and a way of thinking.”
When it comes to corporate venture, Amazon is beginning to get more active. As a fledgling internet company in the late 1990s, Amazon lost hundreds of millions investing in now-infamous dot-com startup failures such as Drugstore.com, Pets.com, and Kozmo.com, among a host of others (it reportedly lost $60M on Kozmo alone). Narrowly evading death itself in the bust, the company would be licking its wounds for years and was decidedly inactive in investment until the late 2000s.
In more recent years, the focus has shifted to more forward-looking ventures across a number of industries spanning healthcare, voice, IoT, and communications platforms. The majority of these investments are thematically linked to its AWS ecosystem, which now encompasses voice, AI, development tools, and cloud computing.
In June 2015, Amazon committed $100M to founding its first stand-alone corporate venture capital unit called Alexa Fund, which specifically invests in voice technology and IoT technology to bolster the ecosystem for its Alexa Voice Service. The Alexa Fund is relatively small (Google Ventures, for reference, started with a $100M per year investment goal) and the fund was the 45th most active CVC in 2016, lagging far behind other tech giants’ CVC initiatives.
Here’s a look at Amazon’s investing frequency from both Amazon itself and the Alexa Fund:
Mirroring much of the venture ecosystem in 2016, Amazon pulled back on dealmaking and made only 3 bets as a company and 12 bets from its Alexa Fund. Both entities did fewer deals than they had the previous year. Amazon’s deal activity, excluding the Alexa Fund, remains on par with its numbers in the 1990s, back when the company was expanding as an internet bookstore.
Compare that level of activity to its tech competitors and Amazon falls somewhere in the middle. Google has made loads of bets, where Facebook and Apple hardly invest, opting instead to purchase companies outright or not invest at all.
With the Alexa Fund propelling Amazon’s investment effort, the company is showing a renewed interest in investing, and activity in the past two years is at its highest levels ever.
As Amazon makes its big foray into the AI world with its Alexa platform product, its corporate venture fund serves as a bellwether for its efforts to build the go-to platform for voice tech. The fund has existed for less than 2 years, and to-date (4/12/17) the fund has done 25 deals, 7 of which were announced with the fund’s inception.
The Alexa Fund mostly does early-stage deals (seed & Series A), but has participated in Series B rounds to consumer IoT heavyweights Ecobee (home automation), Thalmic Labs (gestural computer controls), and Owlet Baby Care (baby monitors), along with a Series C deal to connected doorbell maker Ring. In one way or another, all these startups play into the smart home category which is a big use case for Alexa, or into new human-computer interaction models. Here’s a visualization of the fund’s investment history:
The Alexa Fund has historically done just a deal or two per month, which, as previously mentioned, is well below the activity level of Google Ventures. Also, it is peculiar that there are no fund investments in 2017 so far. (Perhaps related is that in January 2017, Alexa Fund announced an all-new Alexa Accelerator in partnership with TechStars, one of the most active IoT accelerators.)
Alexa Fund investments: Nearly all Alexa Fund investments so far have a potential integration into Alexa’s smart home voice controls. These include Rachio (connected sprinkler system), TrackR (small items finder), Nucleus (connected intercom system), Petnet (smart pet feeder), Musaic (connected speakers), and Scout Security (security camera). Investing here likely offers more strategic value in bringing these products closer into the Alexa ecosystem than a chance at serious returns.
A more forward-looking move might be detectable in Alexa’s 2015 investment into Invoxia, which makes a magnetic kitchen device and was its first partner to use the Alexa Voice Service on third-party hardware (Lenovo and other hardware makers also now incorporate the Alexa platform). Other interesting bets have included Thalmic Labs whose Myo gesture tracking armband could add a new mode to control the Echo, along with DefinedCrowd, which supplies crowdsourced natural language processing (NLP) training data for 90% of world languages. (Presently, Alexa Voice Service only works in English.) Also interesting was the deal to Embodied, which is part of a new wave of social and educational robotic companions that are cropping up.
Alexa Fund’s typical deal partners include familiar names from among the most active IoT investors such as Felicis Ventures and Intel Capital. Interestingly, the fund has shared a number of deals with Canadian early-stage firm Relay Ventures.
Amazon corporate investments
Investments coming from Amazon’s corporate entity are relatively infrequent, which comes as somewhat of a surprise for a company that’s stated strategy is to “experiment patiently, accept failures, plant seeds, protect saplings, and double down when you see customer delight.” As we’ll later explore, however, Amazon is beginning to make more diverse bets.
Amazon’s already-sparse activity slowed in 2016. From 2015 to present, Amazon’s stock has grown at 7 times the rate it did from 2013 to 2015. It may be counterintuitive that a company with such a recent meteoric rise in stock price would slow down its corporate venture efforts. Evidenced by the chart below, investment peaked in Q3’15, with 3 deals done. Since then, investments have trailed down.
Despite the recent slowdown in activity, Amazon is putting capital behind a wider variety of industries. From 2011-2013, the company solely did deals to internet companies, whereas more recently, between 2014-2016, it has also invested in Media, Auto & Transport, and Mobile. This parallels some of Amazon’s broader new business initiatives: investing in freight forwarder Yodel Delivery Network expands its logistics network and strategic knowledge in the UK, and AWS’ developer community is bolstered with investment and partnership with now-public Twilio.
But perhaps the most intriguing new investment area for Amazon is in the Healthcare sector. The company recently invested in its first biotech startup with GRAIL, which focuses on genomics for cancer diagnosis. The deal marks interesting new territory for Amazon’s investing appetite, and because genomic sequencing requires intensive computing power, GRAIL and genomics research and applications in general could parlay nicely into Amazon’s existing AWS business. Not surprisingly, the homepage for AWS Health features gene sequencer Illumina, from which GRAIL spun out, as a customer success story.
Frequently, it seems that Amazon provides capital where it can also create a partnership. In 2015, it invested in cloud computing platform Acquia and India-based home services company HouseJoy. It now has partnerships with both. Twilio also has a partnership with AWS. And last year’s investment in Ionic Security also mentioned a “collaboration” with AWS to create data protection infrastructure for regulated industries.
Some final interesting data slices: the Investor Analytics tab from CB Insights database shows how Amazon’s corporate team heavily favors late-stage deals in the last 5 years. Most of the deals have fallen into the $10M – $25M range, and nearly one-third of deals have been at Series E+.
Interestingly, Amazon has co-invested frequently with the same investment syndicate that helped the company launch as a young startup. As mentioned, Amazon originally raised from Kleiner Perkins Caufield & Byers and angel Tom Alberg of Madrona Venture Group, so it’s worth noting that both firms and Amazon continued to do deals in the dot-com era.
Upfront Ventures, Andreessen Horowitz, and Windcrest Partners are top investors after Amazon makes bets.
Early on, Amazon’s zealous use of intellectual property sparked some controversy. One of the company’s early patents “Method and System for Placing a Purchase Order Via a Communication Network,” perhaps best known by its trademarked name 1-Click, was granted in 1999. The patent is still used today in Amazon’s online store, which, as the name implies, allows orders to be done in one click based on user data saved from previous orders. Notably, the 1-Click patent expires in 2017, and a number of e-commerce players, including Google, are already working on one-click browsers.
The 1-Click patent ended up being a central issue in Amazon’s early life, and according to Brad Stone’s book “The Everything Store,” Amazon was aggressive in enforcing its IP with competitors:
“Critics charged that the idea behind 1-Click was rudimentary and that its approval by the U.S. patent office was a symptom of lazy bureaucracy and a broken patent process. Bezos didn’t altogether disagree—intellectually he was an advocate for patent reform—but he was determined to exploit the status quo for any possible advantage. He sued Barnes & Noble for infringing on the patent in late 1999 and won a preliminary ruling that forced the bookseller to add an extra step to its checkout process. Amazon licensed the patent to Apple in 2000 for an undisclosed sum and tried to use it, ineffectively, to gain some leverage over a rising and worrisome rival that first showed up on Amazon’s radar in mid-1998: eBay.” [p. 77]
Since the dot-com era, Amazon’s patents have shifted and tracked Amazon’s new business priorities. In recent years, Amazon has built a trove of patents, which we’ll explore now in greater depth.
Note: This analysis comes with a few caveats, primarily that the patent filing process involves a significant time lag before the publishing of patent applications. This delay can range from several months to over two years. We also focused on Amazon proper for the purposes of this analysis, which would exclude patents absorbed through external acquisitions.
First, it’s worth noting that in recent years Amazon has put more resources toward intellectual property efforts. From a modest 248 patents filed in 2009, only several years later in 2013 the company filed 1100+ patents. Again, though, compared to Google’s patent efforts, Amazon is doing roughly one-third as many applications.
We also mined each year’s applications to tease out recurring keywords from the patent abstracts, using a significance weighting scheme to surface words and phrases. (Note that records prior to 2014 are likely complete, but analysis for the most recent years only includes applications published to date.)
The key phrases data illustrates Amazon’s diverse business priorities, albeit with some time lag. In the early 2010s, applications frequently used keywords like “electronic book” and “content device” which ostensibly would help dig a moat around Amazon’s Kindle efforts. Similarly, Amazon’s AWS business, which took off in the mid-2000s, offers virtualization services through Elastic Cloud Compute, or EC2. Evidently, securing IP around virtual machines is still a high priority: related phrases like “machine instance” were top patent keywords throughout the years, and “virtual machine” was the top key phrase in both 2010 and 2016.
While patents are still being released weekly, the ones that have rolled in from 2015 and 2016 indicate new interest in drones and cybersecurity, given the sudden prominence of keywords like “unmanned vehicles” and “cryptographic key.”
Aerial drones are a large part of Amazon’s strategy to expand its Prime Air logistics network, which Bezos announced in a 2013 episode of 60 Minutes. More recently, the company began making demo flights delivering sunscreen. With logistics and UAVs front-and-center in its patent portfolio, we isolated patents containing logistics-related keywords. This past year in 2016, for which we expect even more patents to keep surfacing, already has a record 78 logistics-related patents to date.
Also inside Amazon’s patent portfolio were some forward-looking patents that give a peek into the futuristic logistics network Amazon may one day engineer. Recently, an application for a patent came to light that suggests Amazon is trying to create a flying warehouse that would dispatch package-laden drones to the ground. Called an “Airborne Fulfillment Center” (AFC), the patent describes it as “an airship that remains at high altitude.”
In another patent, there are details about a drone mesh network that alerts all other drones about their surroundings.
All of these above Amazon patents were surfaced using our patent search engine.
The ‘river delta’: Where Amazon’s many businesses meet
Amazon operates in a wide range of businesses, but at the heart of the company is online shopping, and wielding technology to be the fastest, most convenient way to buy things. The best way to think about Amazon is as an amalgamation of many businesses with solid tech at its core. Ben Thompson on Stratechery wrote: “A more nuanced approach considers the fact that Amazon is not a monolithic operation, but rather a collection of businesses sharing resources, including a channel (Amazon.com), logistics, and a common technological foundation.” It might have its hand in every business imaginable, but undergirding it all is technological prowess.
While revenue has grown massively, Amazon has hardly turned a profit because it continually reinvests its cash into new businesses, including building new warehouses and beefing up AWS data centers. The philosophy and motivation behind Amazon’s bold reinvestment strategy is best articulated by CEO Jeff Bezos:
“I very frequently get the question: ‘What’s going to change in the next 10 years?’ And that is a very interesting question; it’s a very common one. I almost never get the question: ‘What’s not going to change in the next 10 years?’ And I submit to you that that second question is actually the more important of the two — because you can build a business strategy around the things that are stable in time. … [I]n our retail business, we know that customers want low prices, and I know that’s going to be true 10 years from now. They want fast delivery; they want vast selection. It’s impossible to imagine a future 10 years from now where a customer comes up and says, ‘Jeff I love Amazon; I just wish the prices were a little higher,’ [or] ‘I love Amazon; I just wish you’d deliver a little more slowly.’ Impossible. And so the effort we put into those things, spinning those things up, we know the energy we put into it today will still be paying off dividends for our customers 10 years from now. When you have something that you know is true, even over the long term, you can afford to put a lot of energy into it.”
Always thinking of how to build tomorrow today, Amazon is investing in and improving its core business, by developing new ones.
Moreover, in nearly all of its main categories Amazon’s position as a platform works in a data feedback loop: Amazon owns the richest data set on how consumers consume, how sellers sell, and (among the richest) in how developers develop. This, in turn, allows Amazon to optimize its online shopping experience, its logistics network, and developer environment (and even its voice AI), which all, in turn, make Amazon’s offerings even richer.
In short, many of Amazon’s businesses follow the classic network effect flywheel. And, as we’ll explore below, some of Amazon’s network effects are starting to collide.
A final theme that echoes through its dealings is the concept of diffusing its internal tools as products. The company began as the sole seller on Amazon.com, and eventually opened up its e-commerce platform (and logistics network) for 3rd parties to sell on. It opened up the computing infrastructure it spun up in-house, and sold the computing and storage tools through the now-dominant AWS. Now, it’s in a position to do the same in cutting-edge areas like machine learning, workerless retail tech, drone delivery, and voice computing. As these startup-dominated industries become more and more a reality of everyday commerce, Amazon yet again becomes the “internet tax” for doing modern business.
Core business in commerce and retail
Borrowing from Walmart’s “Everyday Low Price” playbook that was popularized in 1990s retail, Amazon delivered value through competitive pricing. After considering a list of 20 different items, Bezos originally settled on bookselling because of the markup and because no physical store could hold all the titles. Not having physical stores allowed Amazon to maintain a selection of over 1.1M book titles, and develop new customer friendly (and now-standard) e-commerce features like a personalized web page and book recommendation algorithms. By 1998, the company would offer music and DVDs. And nowadays, everything from cars to uranium ore can be bought online through Amazon.
Amazon’s dominance is still often linked to its pricing strategy. Since inception, the lower cost structure of having no stores allowed savings to be passed on to customers. And early on, it would use a web crawler to find competitor prices and undercut them.
For most of the past decade, any company that competed with Amazon was either acquired (Zappos, Diapers.com) or simply steamrolled. The company seemed relatively unaffected by the rise (and subsequent demise in some cases) of on-demand startups bringing goods via apps, although it did start its own version of these with Prime Now. As retailers are shuttering stores, Amazon’s e-commerce business is still growing. In 2016, Amazon accounted for 53% of the sales growth in online retail.
In 2000, Amazon gave outside companies the ability to sell on Amazon.com, and this program now accounts for 49% of the goods it ships. Behind the scenes, Amazon’s retail marketplace “is like this massive slowed-down stock exchange” where its 2 million registered merchants leverage algorithms to undercut competitors. Prices even for commodity goods spike and fall like those in a volatile exchange.
Increasingly, manufacturers are also going straight through Amazon and its Marketplace division to reach consumers, and Marketplace is now the company’s largest source of revenue after retail. Not coincidentally, in mid-April 2017 the company had more than 1200 open jobs listed for its seller services division alone.
And where Amazon can compete, it will also develop its own products as a competing supplier. About a year ago Amazon began selling over a dozen private-label goods for households. In addition to home essentials sold under its Amazon Basics brand, the company has private labels in apparel, CPG, and diapers. The strategy behind doing so is because Amazon can take advantage of higher profit margins: the company doesn’t need to spend much on marketing and brand development, and with its e-commerce data it can know which products will resonate with customers. But Amazon’s leverage as a seller and the platform owner makes for an awkward relationship. Amazon has the power to put its products higher in the search rankings. Resentful suppliers may want to leave the channel as a result but can’t afford to lose the distribution channel and are forced to compete with the store brand.
Amazon Prime — the membership program created to gratify more time-sensitive and less price-conscious customers — gives two-day shipping with an annual membership fee. Prime has also expanded far beyond free shipping, and it now includes Prime Video media streaming and other exclusive services to keep its subscription rates high.
After decades thriving as a storeless internet company, Amazon is also making its foray into brick-and-mortar retail on several fronts: Amazon Go, Amazon Fresh, and Amazon Books. Amazon Go, a vision for a store without employees, will employ RFID tech and computer vision to allow any Amazon Prime member to shop without a checkout process. (Amazon board member Tom Alberg is also a backer and board member of recently-IPO’d RFID company Impinj.) Amazon said the Go Store technology may be delayed due to kinks in the technology, but once perfected the retail tech could be packaged and sold as a platform for physical retailers. This would allow retail tech to be yet another way Amazon “dogfoods” its tools into a tax on commerce, and perhaps — collects more customer data.
Paradoxically, moving into brick-and-mortar allows Amazon to expand its reach with an offline presence. Some consumers simply prefer touching and seeing goods in person, especially apparel, which Amazon is now supplying with AmazonBasics. It’s counter-intuitive because stores, initially, were the enemy for Amazon and are limited by selection. But by having a last mile channel for books, groceries, and big ticket sales like furniture and appliances, it effectively sells what it wouldn’t already be selling online. AmazonFresh Pickup will be a car-side grocery pickup offering as the company eyes the massive grocery market. (Amazon reportedly passed on acquiring Whole Foods recently, as detailed above.) And the company is increasing its footprint with physical bookstores, which are just hitting major American cities.
Overseas, Amazon is said to be targeting India for grocery stores. India is also expected to become the world’s fastest-growing e-commerce market, and Amazon has said it’s investing $3B toward its India business. Its investment in home services company HouseJoy and insurance marketplace BankBazaar furthers its involvement in India’s tech ecosystem. In just a few years in India, it’s made contracts with major delivery services and is starting its own delivery service to augment these. But as it stands, only 35% of India’s population is connected to the internet. And Amazon faces tough competition in India from local players Flipkart and Snapdeal, with backing from Tencent and Alibaba respectively.
Transportation and logistics
Shipping items faster and cheaper has long been Amazon’s way of making the customer happy. As Amazon has grown to deliver hundreds of millions of packages per year, scaling up its fulfillment infrastructure has been the priority. Early on, Amazon realized its shipments were all unique combinations of goods, and that its warehouse techniques would actually be closer to manufacturing than shipping. It poached heavily from Walmart’s executives to grow its logistics network.
When Amazon bought warehouse robotics maker Kiva Systems in 2012, it might have been difficult to see the immediate value. But Amazon now has more than 45,000 robots in its warehouses, after adding about 15K each year. The acquisition of Kiva, according to Bloomberg, “set off an arms race among robot makers and shippers who scurried to keep up with the e-commerce giant” and its efficiencies.
Most recently, Amazon has been eager to grow its fulfillment center coverage. Right now, 44% of Americans live within 20 miles of an Amazon warehouse, compared to just 5% in 2015. The company is also looking to rely less on incumbent shipping networks and is taking steps to build its delivery network, which may eventually grow to a scale that could compete with Fedex and UPS.
Its efforts to achieve independence by land, air, and sea are bold: this year Amazon is finishing development of a $1.5B air freight center in Kentucky. The company is planning to lease 40 “Prime Air” cargo jets, and offering China-based sellers use of its delivery network. In addition, the company is acting as a freight forwarder by helping Chinese suppliers book space on ocean vessels.
Unmanned logistics and delivery remain a question mark with an uncertain timeline, despite the patents mentioned above. On the drones front, Amazon is heavily focused on an in-house effort to add autonomous aerial drones to its arsenal of delivery methods. It’s been several years since the announcement, and without much to show yet (except for isolated demos), a question going forward will be about how and if it can integrate this into its logistics network. FAA regulations have made domestic test flights more complicated. To avoid these complications, drone startups like Zipline International have tested in less regulated skies in Africa, where others simply focus on terrestrial drones.
AI & voice
Amazon’s work in AI is most outwardly visible with its Alexa Voice Service (AVS), which applies AI to natural language processing (NLP) to make a voice interface that’s powered by the cloud. A part of Amazon’s voice tech is built atop technology developed by a startup called Evi Technologies, which it acquired in 2013, and what would ensue was a consumer tech home run. Right around the time Amazon released its ill-fated Fire phone in late 2014, it also quietly released the Amazon Echo, a screenless cylindrical computer running its AVS cloud software. With NLP that appears to work better than the rest, the Echo has been a smash hit and shipped an estimated 5M units. Amazon’s app store for voice “skills” now tallies over 10,000. Alexa also unlocks a new source of revenue: vocal commerce. Alexa has skills so users can seamlessly order more Amazon goods, in addition to third party integrations like hailing cabs and ordering pizzas.
It’s estimated that Amazon is selling Alexa hardware at a 10% – 20% net loss (which totaled $300M in 2016 and is expected to run ~$600M in 2017), all in pursuit of being the dominant platform for NLP. But these might only be short-term expenses. Subsidizing its Alexa-powered hardware is a clear path to ascendancy within consumer and developer circles, and Amazon has conveyed it has no desire to make its bet solely on hardware.
Instead, the end goal is to be the cloud-based voice software powering everything from car dashboards to consumer wearables. As Don Morrill of the Alexa team has said, “You can run a skill [as a developer] for pennies a month. And so we find that’s the best way: we want to eliminate as much friction to Alexa as possible, and keeping it open and free is the way to adoption.” Amazon has thrown lots of resources behind this effort, including giving free AWS credits to Alexa and AI developers. Bezos has said they have more than 1000 people devoted to the Echo and Alexa ecosystem, and Amazon currently has 835 jobs listed for its Alexa team.
While it’s currently the preeminent voice platform, Amazon is up against tough competition. Google’s rival offering called Home powered by its own Google Assistant is a popular Echo competitor. Where Amazon might have the first-mover advantage, Google also benefits from its extensive AI research and from having 20% of its search engine queries being done with voice. Deepening Amazon’s moat in voice is important, especially if consumers perceive a quality edge in Apple’s Siri or Google’s assistant, or another competitor. The switching costs are low, and very quickly, Alexa could quickly lose luster and market share.
On the subject of commercial AI, Bezos has said: “All of the major tech companies will do this … Right now, bigger companies like Amazon have a bigger advantage especially because of the training data sets required to do this. You need a lot of data to do extraordinary things with the algorithms we have.”
Again, widely diffusing its AI tools seems to be Amazon’s strategy going forward.
Amazon’s AI strategy is about being everywhere, or more technically speaking, achieving scale by being the ubiquitous platform that developers use to access AI services.
Amazon recently began selling AI-as-a-Service with “Amazon AI” under its AWS banner. This product offers algorithm training proportional to the power and capacity needed. Similar to how server racks were enormously cost prohibitive for startups pre-AWS, the same is presently true for training machine learning algorithms. Startups are continually devising creative (and expensive) hacks to train their algorithms. Amazon AI‘s goal is to serve big and small-time developers who want AI without the upfront costs or hassle. Amazon AI unveiled offerings that will work like an API and allow any developer to access Lex (the NLP inside Alexa), Amazon Polly (speech synthesis), and Amazon Rekognition (image analysis). With a drone effort in the works, which involves vision for obstacle avoidance, it seems likely that Amazon will continue to move into the vision space and offer more pre-built algorithms.
In Bezos’ recent shareholder letter recapping 2016, he spent a significant portion writing about the AI strategy going forward. In no uncertain terms, Bezos highlights how its tools in voice NLP, computer vision, Amazon Go, and AI infrastructure will carry it forward:
“We’re in the middle of an obvious [trend] right now: machine learning and artificial intelligence. Over the past decades computers have broadly automated tasks that programmers could describe with clear rules and algorithms. Modern machine learning techniques now allow us to do the same for tasks where describing the precise rules is much harder. At Amazon, we’ve been engaged in the practical application of machine learning for many years now. Some of this work is highly visible: our autonomous Prime Air delivery drones; the Amazon Go convenience store that uses machine vision to eliminate checkout lines; and Alexa our cloud-based AI assistant. (We still struggle to keep Echo in stock, despite our best efforts. A high-quality problem, but a problem. We’re working on it.) But much of what we do with machine learning happens beneath the surface. Machine learning drives our algorithms for demand forecasting, product search ranking, product and deals recommendations, merchandising placements, fraud detection, translations, and much more. Though less visible, much of the impact of machine learning will be of this type – quietly but meaningfully improving core operations. Inside AWS, we’re excited to lower the costs and barriers to machine learning and AI so organizations of all sizes can take advantage of these advanced techniques.”
Bezos also mentions “quieter” machine learning and AI products working in the background. Amazon owes development of Alexa hardware to Lab126 and many of its early recommendation algorithms to A9, both of which are secretive divisions Amazon started in Silicon Valley. These bold bets on skunkworks teams have ended up paying off big, and the fruits of their labor are now front-and-center in Amazon’s computing strategy.
Of course, as everything becomes more digitized, applying AI to more novel areas like healthcare could create new breakthroughs. Amazon’s investment in genomics startup GRAIL was a vote of confidence in the area, and Amazon recently worked with Merck to make an Alexa skill for diabetes patients to check glucose levels. (Already, healthcare is one of the hottest areas for AI startups.)
Being that AWS is the go-to place for anything big data-related means Amazon is well-positioned to pivot once AI has its watershed moment. AI is at peak hype in the startup world currently, but from Bezos’ words it seems that AI could likely be Amazon’s next pillar, up there with Prime and AWS.
Media and publishing
Bezos has said that Prime is one of Amazon’s 3 current pillars. To bolster its Prime subscriptions, which is Amazon’s largest source of revenue after e-commerce and its 3rd party marketplace, the company has introduced another special benefit to membership: Prime Video streaming.
Bezos has publicly explained the flywheel effect of Prime Video. With a suite of premium streaming, users are more likely to renew their (decently large ticket) Prime memberships, and are also inclined to buy more, which, in turn, make for more Prime memberships and e-commerce sales:
“Amazon Studios is making original content for Prime Video…from a business point of view for us we get to monetize this content in an unusual way. Winning a Golden Globe helps us sell more shoes and it does that in a very direct way. If you look at Prime members, they buy more on Amazon than non-Prime members. One of the reasons they do that is because they’ve paid their annual fee, they’re looking around to see how to get more value out of the program. They look across more categories… We’ve monitored that Prime Video customers renew [Prime] at higher rates, and they convert from free trials at higher rates.”
Interestingly, Bezos also said that he doesn’t see Netflix as a competitor. He argues they don’t necessarily compete head-to-head with Netflix on the demand side: “when it comes to these over-the-top subscriptions, I think people are going to subscribe to Netflix, and Prime Video, and Hulu, and HBO, and so on.”
But similar to Netflix, Amazon has moved beyond simply distributing content. Amazon Studios produced movies that racked up three Academy Award wins at the 2017 Oscars: “Manchester By the Sea” won Best Actor and Best Original Screenplay, and “The Salesman” won best Foreign Film. (That’s 2 more Oscars than Netflix, which won for a short-subject documentary.) Amazon’s interest in content could further transform the entertainment industry: Amazon is expected to spend $4.7B on content this year, which is twice HBO’s budget. But that’s still trailing Netflix’s aggressive $6B budget for 2017 content, which is double what it spent in 2016. Amazon also recently won a $50M contract for NFL streaming, which could only further the company’s reach.
Fred Wilson of Union Square Ventures recently expressed worry that Amazon has an unfair advantage in its ability to bundle content with other Prime services:
“I worry Amazon has an almost unfair advantage in the content business because of Prime. It’s the craziest thing ever. Who would have ever thought shipping physical products to people at a loss would become this incredible competitive moat in the television business? Who dreamt that up? I happen to believe that it’s an accident, that Bezos woke up one day and said ‘Wow, actually we can take Prime we can use that revenue to subsidize our way into the content business.’ But honestly, if you’re a Prime subscriber it doesn’t cost you any more to get their TV. And I gotta come out-of-pocket with Netflix. If you’re only going to have one, which one are you going to have?…People who don’t have hundreds of dollars a month to spend on entertainment might choose just one [entertainment provider], and Amazon might be the one they choose because they’re already Prime. They gotta buy their groceries anyway. It’s just this incredible bundle.”
The Amazon store’s position as a hub for eyeballs also makes it a valuable resource for selling ads. Amazon has implemented a programmatic ad product that’s expected to generate over $1B in online ad revenue in 2017. If it catches on, it could break up the Facebook-Google duopoly in internet advertising. Martin Sorrell, the head of advertising giant WPP, said Amazon’s potential in this area is what keeps him up at night, since it could connect directly with the brands and manufacturers that are now WPP’s clients.
Amazon also upended the bookselling industry from its start with features like “Look Inside,” and later brought books into the digital era with its Kindle e-readers, another product of Amazon’s R&D arm Lab126. Its acquisition of Audible continues to supply audiobooks, which is currently the fastest-growing format in publishing. Back in its infancy, Amazon would throw its weight around the publishing world by threatening to lower the store rankings of publishers who didn’t meet their digitization demands.
As e-sports — competitive video-gaming — continues its rise to prominence, Twitch, the video game broadcast site Amazon bought for just shy of a billion dollars, now boasts being one of the largest bandwidth users in the US, and has 9.7M daily active users. The company has attracted YouTube gaming stars, and is planning to start selling video games, an expansion that would compete with the game marketplace world of Steam and Valve. Twitch membership is free with Prime, making joining Amazon attractive to gamers who are already well accustomed to buying their gaming titles online.
With its unique incentive to sell more shoes by way of giving away more content and services, Amazon is able to offer a premium media suite at a cost its competitors cannot.
AWS and enterprise cloud
AWS is now Amazon’s largest source of revenue after its core business, posting $12.2B in sales in 2016 and more than $3B in profit.
AWS’ genesis was a result of Amazon overhauling its own internal capacity for cloud services. This allowed startups to migrate from expensive server hardware and software which represented a fixed cost to a variable cost based on usage, and it played no small part in the new wave of startups that flourished in the aftermath including present-day unicorns like Palantir and Slack.
In fact, AWS’ relationship to the venture world started off somewhat fraught. Amazon CTO Werner Vogels said that VCs initially loathed AWS because it “robbed them of the opportunity to get significant chunks of young businesses” by providing previously unaffordable computer infrastructure. But he also notes that it has increasingly been accepted because it allows VCs to spread risk across a greater number of smaller startups. Nowadays, VCs often give away AWS gift certificates to their startups.
As Brad Stone wrote, “It is not hyperbole to say that AWS, particularly the original services like S3 [storage] and EC2 [VMs], helped lift the entire technology industry out of a prolonged dot-com malaise.” Stone also wrote that, more importantly, Amazon was also able to shake its own image as a simple e-commerce player and transform itself into a big-league tech company:
“Perhaps the greatest makeover was of Amazon’s own image. AWS enlarged the scope of what it meant to be the everything store and stocked Amazon’s shelves with incongruous products like spot instances and storage terabytes. It made Amazon a confusing target for Walmart and other rival retailers and gave the company fresh appeal to the legions of engineers looking to solve the world’s most interesting problems. Finally, after years of setbacks and internal rancor, Amazon was unquestionably a technology company, what Bezos had always imagined it to be.“
With cybersecurity being another hot startup category presently, it’s also no surprise that AWS wants to beef up its offerings in this area. As indicated by their patents and acquisitions, Amazon is very focused on improving and securing the stack. Last fall, AWS suffered DDoS attacks, causing many websites across the internet to be affected because of AWS’ status as the go-to for corporate computing. It wouldn’t be surprising to see more cybersecurity M&A like the recent acquisition of Harvest.ai to continue the company’s efforts to strengthen its offerings.
AWS likely got a head start because of classic Innovator’s Dilemma nearsightedness, where many of its competitors didn’t see infrastructure-as-a-service (IaaS) as a threat. Now that’s reached a critical mass, and the strategy going forward will likely be centered on making code deployment even more seamless. We’ll likely see more “as-a-service” type offerings that make deployment flexible and dead simple. One example in this arena is AWS’ Lambda’s serverless computing, which easily runs a snippet of code upon request, so developers don’t need to consider provisioning or managing servers. In essence, this is the micro-service version of EC2, but EC2 is still a full server that has to be managed (think: Airbnb vs renting an apartment). With Lambda, developers can quickly go live with quick, one-time bits of code in a way that was previously cost and time-prohibitive. To build its moat, AWS will need to offer more services like Lambda, along with its AI and voice APIs that empower small-time coders.
Hardware & Devices
Along with its aforementioned line of Echo hardware products running its voice software, Amazon has a host of products like Fire TV stick, dash buttons, and a line of tablets.The hardware R&D group Lab126 was the visionary behind these successes, and the Echo’s popularity has allowed Amazon to be the go-to name for voice computing. But Lab126 also was responsible for the massive failure of the Fire phone, which was caused by a string of strategic errors, with some coming straight from the top: Bezos was known to have played a large part in the demise, obsessing over the 3D feature that fell flat with consumers. (His taste in design also proved questionable with the Kindle, for which he demanded a clunky keyboard that in later iterations was removed.)
Amazon’s had its share of mega-hits, but hardware might be its Achilles heel. Increasingly, it seems the company has recognized this by playing to its strengths in the cloud and focusing on the software platforms behind their IoT devices. This was first apparent with its plans for the Alexa Voice Service, and opening it up for third-party device makers like Invoxia. But now, Amazon is also opening up the hardware tools it made for the Echo and recently announced it’s opening its 7-microphone array for voice processing to commercial developers. The move is yet another example of Amazon disseminating the essential tools and allowing creativity to evolve naturally.
How it will build on its latest success with Echo remains up in the air, and perhaps Amazon has realized that from here it can only open up its ecosystem to the creativity of third parties. Another option is that we could see Amazon’s interest in mobile phones resurface in the future, but given its place as a black spot in the company history, we could be waiting a while.
Other new businesses
Amazon has been experimenting with financial technology that could widen its reach. In India, which is expected to become the world’s fastest-growing e-commerce market, as discussed, the company is offering thousands of loans to e-sellers so suppliers can expand their operations and manage seasonal spikes.
Amazon is also expanding its financial reach by launching Amazon Cash, which allows users to add to their Amazon.com balance by showing barcodes at brick-and-mortar checkout locations. The move is said to help appeal to the “underbanked” who are accustomed to dealing with cash and might be new to online commerce.
In addition to fintech, Amazon might have its eyes on augmented and virtual reality (AR/VR). Amazon’s Lumberyard, a game development engine on AWS, could play a greater role in the development of VR content. The company is also rumored to be integrating more augmented reality tech into its brick-and-mortar effort so patrons could picture how items would look in their homes.
In short, fintech and AR/VR could help Amazon provide more frictionless commerce.
Given its mutually reinforcing lines of business in commerce, cloud computing, and AI-as-a-service, it’s not easy to find chinks in the Amazon armor. But in the AI arena at least, Amazon is up against more research-oriented peers in Silicon Valley who are aggressively recruiting the top AI talent and have data troves at least as large as Amazon’s to train their algorithms on.
And Amazon is fighting the technical AI battle while it is simultaneously working across more than 5 industries and waging bare-knuckle fights for e-commerce market share in India and the Middle East.
Jeff Bezos, who has made relatively few missteps in twenty years of tenure, seems capable of carrying the company forward. Healthy and relatively young, much will still be demanded of him. As Brad Stone wrote, “In a way the entire company is scaffolding built around his brain — an amplification machine meant to disseminate his ingenuity and drive across the greatest possible radius.” It’s almost impossible to imagine Amazon carrying forward without Bezos at the helm, and there’s hardly an inkling that he is going anywhere. Yet, it’s also not clear that Bezos has cultivated the next generation of leadership to move toward a central vision without him driving the company forward.
Instead, the biggest threat to growth could be Amazon itself. Its outsize role in the commerce ecosystem, reports the Economist, could attract the ire of regulators: “If Amazon does become a utility for commerce, the calls will grow for it to be regulated as one.”
Another question is how long investors will stand for a dream deferred. As the Times wrote back in 2013:
“In its 16 years as a public company, Amazon has received unique permission from Wall Street to concentrate on expanding its infrastructure, increasing revenue at the expense of profit. Stockholders have pushed Amazon shares up to a record level, even though the company makes only pocket change. Profits were always promised tomorrow.”
Despite having earned a reputation as a “profit miser,” Amazon recently began posting profits in large part due to the success of Amazon Web Services. As Amazon dips deeper into its coffers to streamline its logistics network, boost its web offerings, and develop artificial intelligence, it will face increasing pressure to also reward patient shareholders. The tensions could one day reach a breaking point.