Case study on Amazon’s approach to innovation and competition in the knowledge economy

Introduction

Amazon is generally regarded as one of the most innovative companies in the world (Reed, 2017). In considering how Amazon approaches innovation within the knowledge economy we’ll frame the analysis of new technologies by looking at McKinsey’s research on disruptive technologies that have potential for economic impact, how Amazon has approached innovation in each of these new technologies, and consider how innovation has impacted Amazon’s revenue growth.

Amazon’s approach to innovation

Since beginning in 1995 as an online bookstore Amazon has expanded into ecommerce marketplace, digital advertising, cloud computing, groceries and apparel, and artificial intelligence industries. Amazon’s investment strategy for innovation is to act like a growth investor, spreading it’s investments across a diverse range of sectors and industries. This is a markedly different strategy to other tech giants who choose to focus the majority of their innovation efforts within their core competencies e.g. Facebook with social networks and Apple with consumer electronic devices (Bowman, 2017).

Jeff Bezo, CEO of Amazon, explains, “Because of our emphasis on the long-term, we may make decisions and weigh tradeoffs differently than some companies… We will continue to make investment decisions in light of long-term market leadership considerations rather than short-term profitability considerations” (Bezos, 1997). It is through this approach to innovation that Amazon seeks to develop monopolies in all of the sectors that it enters.

McKinsey Global Institute’s report (Fig. 1) on disruptive technologies identifies “12 technologies that could drive truly massive economic transformations and disruptions in the coming years” (Manyika et al, 2013). Amazon is publicly investing in at least eight of the twelve technologies, through investing in companies that are working on new technologies, utilising the new technology to build organisational capacity and improve productivity, and through commercialising the new technologies in ways that enable Amazon’s business customers to implement within their companies.

McKinsey gallery of disruptive technologies
Figure 1. Source: McKinsey Global Institute

Amazon is known for its secrecy around innovation, necessarily so in order to protect its trade secrets, but by looking at how Amazon approaches the six most impactful disruptive technologies we can gain an understanding of how Amazon approaches innovation.

Mobile Internet

“Increasingly inexpensive and capable mobile computing devices and Internet connectivity”

Amazon’s Kindle eReader, which launched in 2007, wasn’t the first eReader on the market but with it’s innovative WiFi hardware and Kindle Direct Publishing, the self-publishing platform, it enabled customers to have thousands of books available within seconds and authors to publish their writing without relying on the publishing industry (Fox Rubin, 2017). Whilst the design of the device was very similar to other eReaders, it was Amazon’s move to create its own ebook format and the Kindle Direct Publishing Network to allow ebooks to be published in its own format that fits with Amazon’s approach to innovation.

Whilst the majority of manufacturers were focused on developing eReader devices that could support .epub as the main format for ebooks, Amazon was instead establishing a core competency around its own format and publishing network. Developing the device was a complementary competency for Amazon, although one important enough for Amazon to ensure it controlled the device as part of the value chain for ebooks.

The digitization of books was a technological breakthrough which following Anderson and Tuishman’s evolutionary model of technological change, resulted in lots of technical variation in the formats available. Over the first decade the variations in formats reduced until the current situation of having two formats, epub and Amazon KIndle format, available. The ebook market hasn’t yet arrived at a single dominant design (Anderson & Tushman, 1990. Suárez & Utterback, 1995) however as Amazon currently dominates the eReader market with 60% of worldwide device sales in 2017 (Fox Rubin, 2017), only time will tell if the Amazon format for ebooks becomes the dominant design.

Automation of knowledge work

“Intelligent software systems that can perform knowledge-work tasks”

In 2018 Amazon “reorganised around Artificial Intelligence” (Morgan, 2018). This reorganisation focused other teams and departments at Amazon to utilise AI in their products and services, including warehouse management, recommendations on Amazon Music, Prime Video and on the ecommerce marketplace, Alexa and the Amazon Go store (Levy, 2018). This demonstrates Amazon’s approach to automating knowledge work. AI isn’t considered a single product that remains within a single team, it is a technology and capability that Amazon clearly regards as a core competency that should be utilised in as many ways as possible in order to give Amazon a competitive advantage in all of the sectors it operates in.

This is an example of what Tushman and Anderson explain when they say, “Technological innovation affects not only a given population, but also those populations within technologically interdependent communities” (Tushman & Anderson, 1986). Amazon leveraged the technological innovation of AI to gain benefits across all areas of its business, however it remains unclear whether this new technology was a competence-destroying because it required completely different skills and knowledge to operate or competence-enhancing because it built “on existing know how yet did not render skills obsolete” (Tushman and Anderson 1986).

Having realised the benefits Amazon went on to commercialise it’s AI by creating AutoGluon, a service that enables developers to build applications involving machine learning on top of AWS (Hepburn, 2020). “Commercial AI has enjoyed what we at Amazon call the flywheel effect: customer interactions with AI systems generate data; with more data, machine learning algorithms perform better, which leads to better customer experiences; better customer experiences drive more usage and engagement, which in turn generate more data.” (Sarikaya, 2019).

Internet of Things

“Networks of low-cost sensors and actuators for data collection, monitoring, decision making, and process optimisation”

The Amazon Dash, an internet-enabled button for making repeat purchases, was Amazon’s move into Internet of Things devices. On sale for less than four years the device fell foul of consumer protection laws in Germany, Amazon’s second biggest market at the time, where a court ruled that Amazon Dash didn’t provide customers with enough information to make informed purchases (Jagannathan, 2019). Although a regulatory and revenue-generating failure, the device may have been more of a success in establishing Amazon’s first-mover advantage into the market of consumer IoT devices and in collecting data on buying behaviour (Newman, 2016) to inform the next generation of devices.

Echo and Dot, the home speakers with the Alexa voice technology, soon replaced the Dash as a means of making purchases easier for consumers and as a means of collecting data on buying behaviour, data which could also be used to train the machine learning algorithms that powered Alexa. Voice-powered machine learning algorithms are intangible assets that require investment but have different economic characteristics to tangible assets (Haskell & Westlake, 2017):

  • Sunk costs – represent an investment that is unlikely to deliver a return in the way a tangible asset would if resold as intangible assets are difficult to sell as they are often bespoke to the company developing them, as in the case of Alexa algorithms.
  • Spillovers – are benefits competitors may gain from appropriating intangible assets such as the design of a device which is easy to reverse engineer. Amazon’s defense is to focus more on things that are difficult to copy such as bespoke algorithms.
  • Scalability – a characteristic of an intangible asset that can be leveraged in ways that tangible assets cannot without increased investment, such as the Alexa algorithm which works on all Alexa powered devices, but also the ‘brand’ of Alexa as a likeable, humanised, ‘part of the family’ voice assistant in comparison to Google choice to call its voice assistant Google.
  • Synergies – occur when intangible assets become more valuable together than in isolation. Alexa has more value because it connects to Amazon’s ecommerce systems and allows customers to make purchases, and because Amazon allows developers to build other services on top of the Amazon ecosystem that enables customers to control the heating and lighting in their homes.

Amazon’s approach to investing more in its intangible assets, such as algorithms, than in the physical devices seems to suggest that they recognise the competitive advantage intangible assets can give them a over other companies, but also that they recognise the risks Haskell and Westlake point out can be associated with this kind of investment (Haskell & Westlake, 2017). The economic value of intangible assets in the case of Alexa comes from strategic choices about how they are leveraged to drive purchasing behaviour in customers.

Cloud

“Use of computer hardware and software resources to deliver services of the Internet

Amazon Web Services is a leading (Gartner, 2018) infrastructure-as-a-service provider. Gartner calls out AWS’ “prioritisation of being first to market” along with being the “most mature enterprise-ready provider, with the strongest track record of customer success” as key aspects of being a leading cloud provider (Bala, et al, 2018). AWS started from the needs of Amazon’s ecommerce business, which required reliable, scalable technology to power its growth in the early 2000’s. By 2003, providing infrastructure services and reliable, scalable data-centers was considered a core competency by Bezos and Amazon senior executives. When Amazon launched AWS in 2006 they were “first to market with a modern cloud infrastructure” (Miller, 2016). AWS holds 40% of the market share in cloud computing (Carey, 2019), a position it gained by building on core competencies it owned in other areas of its business and being years ahead of competitors (Miller, 2016).

Teece talks about the ‘perplexing’ problem of how many companies who are first to market with an innovation are not the ones to commercialise and profit from it. With AWS, Amazon demonstrated that it’s approach to innovation can deliver on significant commercial success. Teece’s framework for determining which company will win from introducing innovation involves understanding the appropriability; the environmental factors that affect the ability to capture profits from an innovation, the design phase; whether a dominant design has emerged, and the competencies necessary for the commercialisation of the innovation.

In the early 2000s, cloud infrastructure services had what Teece describes as a “tight appropriability regime” (Teece, 1986). The environments in which the technology for providing infrastructure services over the internet existed was easy to protect simply because competitors were not yet building their core competencies in cloud. Having a “tight appropriability regime” for cloud services gave AWS the time it needed to launch its products and services before the regime weakened and other entrants could imitate the technology.

At the time of launching AWS, cloud infrastructure services were pre-paradigmatic, the majority of infrastructure providers weren’t even considering cloud, so there was no dominant design. Teece says that, “when imitation is possible and occurs coupled with design modification before the emergence of a dominant design, followers have a good chance of having their modified product anointed as the industry standard, often to the great disadvantage of the innovator.” (Teece, 1986), but this did not happen to Amazon.

Amazon already owned the complementary assets required to commercialise AWS successfully (procurement, marketing, sales, etc.) which removed any bargaining power issues that may have arisen from contracting assets, and put AWS in a good position to quickly establish the dominant design for cloud infrastructure services and so leverage its position as a first-to-market pre-paradigmatic innovator and as a paradigmatic market leader.

Advanced robotics

“Increasingly capable robots with enhanced sensors, dexterity, and intelligence; used to automate many tasks”

In 2012, Amazon acquired Kiva Systems, a small robotics company for $775 million providing Amazon with mobile robots and the technical expertise to begin automating its warehouses and sorting facilities. (Del Rey, 2019). This automation of the work of pickers and packers enabled Amazon to increase efficiency in its warehouse operations by reducing the time taken to pick items for delivery to its customers (Simon, 2019), and so driving the success of its ecommerce business. In 2019 Amazon introduced machines to automate putting customer orders into boxes ready for delivery, a job that was previously performed by thousands of workers. It “would amount to more than 1,300 cuts across 55 U.S. fulfillment centers for standard-sized inventory. Amazon would expect to recover the costs in under two years, at $1 million per machine plus operational expenses.”, reported Reuters (Dastin, 2019). Amazon currently has more than 200,000 mobile robots working inside its warehouse network, alongside hundreds of thousands of human workers.

Amazon, as a low margin business, seeks to organise its supply chain more effectively than its competitors to maximise profits (Teece, 1986). Automation increasingly allows for this in Amazon’s fulfilment business as it replaces the routine work (Autor, Levy & Murnane, 2003) of pickers and packers. In making capital investments in technologies to replace workers with robots Amazon could be said to be taking a skills-biased approach; that is, that it favours more highly skilled workers such as programmers, engineers and mechanics at the cost of lower skilled workers and assumes thats increased productivity for the company comes from fewer highly skilled workers over more lower skilled workers. Ordinarily we would expect that companies would make decisions about how much to invest in automation technologies by considering economic factors such as the cost of labour in a particular geographic market however, from what we’ve seen of Amazon’s investment strategy in innovation it seems more likely that Amazon is playing the long game with automation and betting on machines being capable of performing non-routine cognitive and manual tasks in the future (Frey & Osborne 2013) and so replacing lower skilled workers completely.

The adoption of automation in warehouses and fulfilment centres has been congruent with Amazon’s approach to innovation involving massive investment in technology that provides increased internal capabilities enabling Amazon to become a market leader and then selling that capability to businesses to deliver long term revenue gains. The question of whether robots will replace workers, at least in Amazon warehouses, seems to have an inevitable answer.

Autonomous or near-autonomous vehicles

“Vehicles that can navigate and operate autonomously or semi autonomously in many situations”

Amazon has invested $700 million in Rivian, the electric vehicle manufacturer and $530 million in Aurora, an autonomous driving startup. “For Amazon, this small investment is a good way to enlarge their bet on the E.V. [electric vehicle] market without having to tool up a plant to find out if it will fly. Over time, the Rivian investment could give Amazon a starting point to own and operate an in-house package delivery business.” (Mitchell, 2019).

Amazon has been developing delivery drones “that can fly up to 15 miles and deliver packages under five pounds to customers in less than 30 minutes.” (Vincent, 2019). Developing delivery drones and getting FAA approval might be considered a big enough innovation for most companies, but Amazon goes a huge step further by developing its own Air Traffic Control System for drones. “The system also gives aviation authorities, like the FAA, the ability to track the drones in the airspace to ensure safety and create “no fly zones” in times of emergency. The traffic management system is easy to use for various operators in the same airspace because it will connect via the internet” (Amazon, 2019). In a similar strategic play to the Kindle, Amazon realises that controlling the platform that controls the devices creates considerable more competitive advantage than simply developing the best drones.

Developing drones and autonomous electric vehicles will reduce Amazon’s reliance on third-party delivery partners and own the supply chain (Prosser, 2019), and conceivably it could commercialise the service to compete with FedEx, UPS, etc., and thus drive increased revenue for the company, but in order to do so it needs to protect the design of the drones and vehicles from competitors. Archibugi & Pianta explain that, “technological change impinges on codified and tacit knowledge… innovations can either be embodied in capital goods and products or disembodied, i.e. the know-how included in patents” (Archibugi & Pianta, 1996). As it is almost impossible to protect designs for publicly available machines like drones through trade secrets in the way Amazon does for its software and algorithms, Amazon needs to file patents to protect its disembodied codified knowledge in order to continue to be innovative.

In 2019 Amazon filed over two thousand patents (Capriel, 2019) many for drones and autonomous vehicles, and since 2010 Amazon has grown its patent portfolio from less than 1,000 active patents in 2010 to nearly 10,000 in 2019 (Columbus, 2019), a ten-fold increase in less than a decade (Fig. 2).

Patents Owned by Amazon, United States Patent and Trademark Office
Figure 2. Source: United States Patent and Trademark Office

Patents can be used strategically by companies in a number of ways; to protect inventions with the intention of commercialising them, or simply to prevent competitors from entering the space. This makes the number of patents filed a poor indicator of innovation, and so it seems that the number of patents Amazon has filed has increased over time because they have become involved in more sectors and industries rather than because they have become more innovative.

Amazon’s sources of competitive advantage

These six examples demonstrate Amazon’s superior ability over its competitors and how they employ the same approach towards innovation; not constrained to sectors or industries that they have previously operated in, investing huge amounts to own the sector they move into, building core competencies in their value chain to protect their own competitiveness, and making new technologies available outside their own ecosystem to allow their customers to leverage the technology in ways that support and scale Amazon’s business model, in many cases the customer becoming reliant on Amazon in their value chain, for example Netflix using AWS (Uenlue, 2018).

Amazon’s economic growth from innovation

Amazon’s approach to innovating across multiple sectors and industries has given them significant competitive advantage and commercial success, growing from $6.92 billion in 2004 to $280.52 bn in 2019 (Clement, 2020), an almost 4000% increase.

Source: Statista, Amazon revenue
Source: Statista

The breakdown of Amazon’s commercial performance by it’s main areas of business in 2018 shows it’s longstanding ecommerce business as the main revenue producing business (Day & Gu, 2019):

  • Ecommerce: $234.61 bn sales
  • Cloud computing: $25.6 bn in revenue
  • Groceries: $25.4 bn in sales
  • Online apparel: $24.61 bn in sales
  • Consumables: $23.6 bn in sales
  • Digital advertising: $7.4 bn in revenue

Of the six disruptive innovations discussed above, only cloud computing, where Amazon is the market leader, generates significant income. This reflects Amazon’s approach to innovation involving long-term investment to establish commercial success.

Conclusion

Amazon has earned its reputation as one of the most innovative companies in the world. Amazon’s approach innovation can be broadly summed up in three parts:

  • Large investments and acquisitions in software and hardware startups spread across multiple sectors and industries. This puts Amazon in control of the value chain and reduces the risk of suppliers holding strong bargaining positions.
  • Use the technology that is produced to develop efficiency and productivity gains in products and services in a diverse range of sectors and ensure competitive advantage over the long term.
  • Commercialise those products and services, allowing other companies to leverage them, generating revenue and creating lock-in network effects (Katz & Shapiro, 1994) for those companies and Amazon’s customers.

This approach to innovation has enabled Amazon to develop significantly successful businesses in ecommerce, cloud computing, digital advertising and retail, and is likely to contribute to Amazon’s continued success into the future.

References

Reed, N. 2017. Why Amazon Is The World’s Most Innovative Company Of 2017. Fast Company.

Bowman, J. 2017. Amazon’s Strategy Is a Perfect Example for Growth Investors. The Motley Fool.

Bezos, J. 1997. Letter to Shareholders. US Securities and Exchange Commission.

Manyika, J., Chui, M., Bughin, J., Dobbs, R., Bisson, P., and Marrs, A. 2013. Disruptive technologies: Advances that will transform life, business, and the global economy. McKinsey Global Institute.

Anderson, P. & Tushman, M. 1990. Technological Discontinuities and Dominant Designs: A Cyclical Model of Technological Change. Administrative Science Quarterly.

Suárez, F. F. and Utterback, J. M.. 1995. Dominant Designs and the Survival of Firms. Strategic Management Journal, Vol. 16, No. 6 (Sep., 1995), pp. 415-430. Wiley

Fox Rubin, B. 2017. 2 big innovations that made Amazon’s Kindle a success. CNET.

Morgan, B. 2018. How Amazon Has Reorganized Around Artificial Intelligence And Machine Learning. Forbes Media LLC.

Levy, S. 1990. Inside Amazon’s Artificial Intelligence Flywheel. Wired.com.

Tushman, M. L., & Anderson, P. 1986. Technological Discontinuities and Organizational Environments Administrative Science Quarterly, Vol. 31, No. 3. (Sep., 1986), pp. 439-465.

Hepburn, M. 2020. Amazon’s AutoGluon helps developers deploy deep learning models with just a few lines of code. Amazon.

Sarikaya, M. 2019. How Alexa Learns. Scientific American.

Jagannathan, M. 2019. Amazon’s Dash buttons may be gone, but it’s getting even easier to spend your money. Market Watch.

Newman, J. 2016. The Secret Power Of Amazon’s Dash Buttons: Not Sales, But Data. Fast Company.

Haskel, J. and Stian Westlake, S. 2017. Capitalism without Capital: The Rise of the Intangible Economy. Princeton University Press.

Bala, R.,Gill, B.,Smith, D. and Wright, D. 2019. Magic Quadrant for Cloud Infrastructure as a Service, Worldwide. Garnet, Inc.

Miller, R. 2016. How AWS Came To Be. Techcrunch.

Carey, S. 2019. The history of AWS: A timeline of defining moments from 2002 to now. Computerworld. IDG Network.

Teece, D. (1986) Profiting from Technological Innovation: Implications for integration, collaboration, licensing and public policy. Research Policy, 15, pp. 285-305.

Del Ray, D. 2019. How robots are transforming Amazon warehouse jobs — for better and worse. Vox.

Simon, M. 2019. Inside the Amazon Warehouse Where Humans and Machines Become One. Wired.

Dastin, J. 2019. Amazon rolls out machines that pack orders and replace jobs. Reuters.

Autor, D. H., Levy, F. & Murnane, R. J. 2013. The Skill Content of Recent Technological Change: An Empirical Exploration. The Quarterly Journal of Economics, Volume 118, Issue 4, November 2003, Pages 1279–1333.

Frey, C. B. & Michael A. Osborne, M. A. 2013. The Future Of Employment: How Susceptible Are Jobs To Computerisation? Oxford Martin School, University of Oxford, Oxford.

Mitchell, O. 2019. Amazon keeps on truckin’ with autonomous vehicle investments. The Robot Report.

Vincent, J. 2019. Here’s Amazon’s new transforming Prime Air delivery drone. The Verge.

Amazon Day One Staff, 2019. Another new frontier for Prime Air. blog.aboutamazon.com.

Prosser, M. 2019. Where Is Amazon Headed With Its Self-Driving Car Initiatives? Singularity Hub.

Archibugi , D. & Pianta, M. 1996. Measuring technological change through patents and innovation surveys. Institute for Studies on Scientific Research, National Research Council, Via Cesare De Lollis 12, 00185 Rome, Italy.

Capriel, J. 2019. The most interesting Amazon patents of 2019: From drones that follow you home to virtual reality noses. Washington Business Journal.

Columbus, L. 2019. 10 Charts That Will Change Your Perspective Of Amazon’s Patent Growth. Forbes.

Uenlue, M. 2018. Amazon Business Model: Amazon Web Services. Innovation Tactics.

Clement, J. 2020. Annual net sales of Amazon 2004-2019. Statista.

Day, M. & Gu, J. 2019. The Enormous Numbers Behind Amazon’s Market Reach. Bloomberg.

Katz, M. L. & Shapiro, C. 1994. Systems Competition and Network Effects. Journal of Economic Perspectives. Vol. 8, No. 2, Spring 1994. (pp. 93-115).