When a platform is self-service, even improbable ideas get tried, because there’s no expert gatekeeper ready to say “That will never work!” Guess what? Many of those improbable ideas do work.

—JEFF BEZOS, 2011 LETTER TO SHAREHOLDERS

Imagine for a moment that you own a business that provides a service to your community, a market, or an industry. To make your service the best in the world, you also design, build, test, and use innovative tools and capabilities that not only make your business better but can be sold to other businesses. Your business gets bigger and bigger because your service gets better and better, and your innovative tools become more and more in demand. It’s a virtuous circle.

In other words, you’ve become a platform company—your platform and the innovative tools you build are your service.

Principle 6: Building a platform business model, which allows others to leverage your capabilities to build and grow their own businesses, creates a stronger sustainable, competitive advantage for your business. The Internet of Things creates exciting possibilities for companies to develop a platform business model, leveraging their connected devices for other companies use.

In this chapter, we’ll explore how Amazon identifies, tests, and builds its own potential platform businesses using IoT, how to figure out if a platform business model might be right for you, and how to think about using the Internet of Things to build your platform business model.

Building platform businesses is the Amazon way. It’s the reason Amazon.com currently sells more goods online than its next twelve biggest competitors combined, including Walmart, Apple, Macy’s, and Target.²⁶ It’s a major reason why Amazon sales are growing faster than Internet sales as a whole, accounting for more than 50 percent of all US e-commerce revenue growth. On the enterprise side, it’s a huge reason that Amazon Web Services (AWS) leads the cloud-computing market in size, innovation, and mind share.

The Amazon Marketplace platform powers many of Amazon’s retail capabilities and core assets and enables millions of third parties to sell and deliver items to customers. Amazon’s retail business needed scalable technology infrastructure and tools. It turns out other businesses needed the flexible technology infrastructure as well, and the Amazon Web Services (AWS) cloud business was formed, serving both Amazon the retailer and thousands of other clients.

Now Amazon sees an opportunity for an Internet of Things platform. Enter Alexa, poised to become Amazon’s Trojan horse for IoT.

Through Alexa, the Internet of Things is developing as yet another platform business for Amazon. I expect it to expand for them in much the same way that AWS and Marketplace have. Already Alexa is a voice-controlled interactive platform leveraged by thousands of other companies. This is good news for you because Amazon and other IoT platforms can fuel your IoT plans. These companies also create a model for how to build an IoT platform business.

Not every company has what it takes to build a platform—but if you’re interested in getting your feet wet, there are a few questions you should ask yourself before diving in.

What does it mean to be a platform business? How does it relate to IoT?

Where is the opportunity? Let’s explore…

SIMPLIFICATION EPITOMIZED: AMAZON’S PLATFORM BUSINESSES

Traditional businesses tend to offer a core capability and value proposition to a customer and then organize capabilities to meet that customer’s needs. Take Warby Parker, which makes private-label glasses and sells them online and in their stores, direct to consumer. They sell specific goods—stylish prescription glasses at a great price—that are easy to try on and return, simplifying the glasses-purchasing process.

Platform businesses, on the other hand, create the fundamental capabilities for customers and traditional businesses to do business with one another. They may provide some products or services themselves, but for the most part, they rely heavily on other parties to supply the goods offered to customers. iTunes, which allows artists to upload and sell their songs to customers, is an example of that. Artists direct their fans to the platform, creating a network effect that builds the value of iTunes itself—another hallmark of a platform business.

Platforms knit together complex processes and tasks that are normally more complex for both suppliers and consumers. Amazon’s third-party merchant platform is successful because it makes the process of setting up an online shop, building access to customers, and fulfilling orders—which would normally be complicated and expensive—seamless for small businesses. At the same time, it makes comparison shopping and purchasing seamless and trustworthy for consumers.

In its ideal state, a platform’s capabilities can also be extended and improved upon by other technology developers and other companies. Earlier in this book, we talked about some of the third-party businesses developing software and other capabilities for Amazon’s Echo. As third parties innovate on a platform, those companies and businesses transition from acting as traditional suppliers or vendors to being partners in the business. This is a major advantage of running a successful platform business—you benefit from other partners innovating on the services and capabilities you provide.

Some platform businesses not only let partners use and innovate on services or capabilities, but they allow them to embed these capabilities in other products and services. Take PayPal, which allows users to send and receive money on its platform but also allows sellers to embed PayPal payment processing directly into their own websites. Or take the iTunes App Store, which allows third-party developers to create new capabilities for Apple phones. Alexa Voice Service can be used in other products as the software powering the voice-activated device.

When you enable and incentivize other businesses to build on your capabilities, you are crowdsourcing ideas, innovation, and investment to the betterment of your capabilities.

PLATFORMS -- THE AMAZON WAY

A “platform” is a system that can be programmed and therefore customized by outside developers—users—and in that way, adapted to countless needs and niches that the platform’s original developers could not have possibly contemplated, much less had time to accommodate.²⁷

—MARC ANDREESSEN

When I was at Amazon in the early 2000s, we developed the concept that Amazon was fundamentally two types of businesses. First, it was an online retailer selling everything from books to shampoo. Second, though, it was a platform business that built capabilities for outside companies. Those capabilities were used by Amazon the retailer as well as other companies.

I ran two of Amazon’s platform businesses: First, I launched and scaled the third-party Marketplace business, which today is responsible for more than 50 percent of all units shipped and sold through Amazon. Second, I ran the Enterprise Services business, which ran other large retailers’ e-commerce infrastructures for them, including website infrastructure and management software and branded fulfillment and customer service. Target.com, Toys “R” Us, Marks & Spencer, Sears Canada, and the NBA have all used Amazon’s Enterprise Services platform.

As I worked to build those businesses for Amazon, I was following four clear rules for successful platform business.

1. Platforms Should Simplify Complexity. The first rule of thumb is that you should build a platform business to make hard things simple, often making them self-service. Rather than trying to support confusing or complicated tools with lots of account support required, teams focused on building services that were ridiculously obvious and easy to use. In addition to creating better user experiences and products, this approach provides major benefits to a platform’s ability to scale—adding more users does not directly require more headcount to manage them. Doing this right means making hard things simple and accessible to the right audiences. “Someone shouldn’t have to talk to you to use your stuff,” was a common refrain among Amazon leadership.
2. Platforms Use APIs to Make Embedding Easy for Customers. A platform should be made programmable by creating an application programming interface, or API. Think of an API as a well-defined set of requirements for how one software program talks to another. Creating an API makes technology integration—the use of your platform by other companies—simple and feasible at scale.
3. Platforms Provide Ongoing Value. To create a sustaining platform business, there needs to be an ongoing value proposition that the platform provides and protects. For the Amazon Marketplace business, the heart of this was protecting customer trust by providing payment and account services and extending the A-to-Z Guarantee to third-party products and protecting third-party trust by providing customer demand.
4. Eat Your Own Dog Food. As was the case with AWS, Amazon generally builds its platform businesses just as much for internal customers—Amazon Prime or e-commerce teams, for example—as for external customers. Amazon Retail runs on AWS infrastructure, and many of the innovations of AWS—scaling, reliability requirements, security requirements, and international capabilities—have come from having to support such a massive retail business. Products take lessons from Amazon’s early and exacting internal users to drive R&D and scale much more quickly.

Examples of platforms that Amazon has built and operates today include the following:

• Amazon Web Services. As we mentioned above, AWS is the leading cloud-computing-innovation company. In case you’ve been hiding in a computer closet for the last ten years, its diverse set of on-demand infrastructure, data management, and solutions make up a $10 billion business annually, with operating margins greater than 20 percent. It also happens to be growing at more than 60 percent year over year.
• Fulfillment by Amazon. Fulfillment by Amazon, known as FBA, allows sellers and companies to use Amazon warehouses to store inventory and fulfill orders (sold either on Amazon.com or through other websites). Thousands of small and medium-sized companies use Amazon’s more than three hundred worldwide fulfillment centers to store and ship goods, giving them a global distribution network they could only dream about. Over one billion items were shipped by Amazon on behalf of third-party sellers in 2015.²⁸
• Amazon Machine Learning. Machine learning can be extremely technically complicated, based on complex algorithms and technology. But Amazon’s machine learning service gives developers at all levels a set of visualizations and wizards to help walk them through the process of creating machine learning models. This platform was first developed as part of Amazon’s internal machine learning capabilities, then turned into an external service.
• Amazon Marketplace. As we mentioned above, Amazon Marketplace accounts for more than 50 percent of all units sold at Amazon. The key to Marketplace’s success was threefold—we didn’t compromise on the buying experience for customers shopping with third parties; we created a simple, intuitive seller experience despite the complex integration and set of choreography and data required between the seller and Amazon; and we built a seamless combination of Amazon Prime, Amazon FBA, and Amazon Marketplace that really drove growth.
• CreateSpace. Amazon’s self-service, on-demand publishing platform lets authors like me easily write, design, and publish books. It facilitates print and e-book distribution without authors having to go through the traditional-publishing-house gatekeepers. It was an easy decision for me to self-publish through CreateSpace rather than go through a traditional publisher. One huge advantage is Amazon’s on-demand printing capability. I often speak in front of audiences and teams, which requires a large book order. With CreateSpace, I can order the books I want just a few days in advance rather than having to order large batches with several weeks of lead time.

The examples cover some of Amazon’s most substantial platform businesses, but Amazon is also developing and incubating a host of smaller platform businesses. Amazon Video Direct, which lets users upload and profit from professional-grade video content, is an embryotic platform business that could someday grow into a huge business like the others. Mechanical Turk is a platform to outsource small tasks to a worldwide labor pool.

As Bezos wrote in the 2011 shareholder letter, “The most radical and transformative of inventions are often those that empower others to unleash their creativity—to pursue their dreams. That’s a big part of what’s going on with Amazon Web Services, Fulfillment by Amazon, and Kindle Direct Publishing. With AWS, FBA, and KDP, we are creating powerful self-service platforms that allow thousands of people to boldly experiment and accomplish things that would otherwise be impossible or impractical. These innovative, large-scale platforms are not zero-sum—they create win-win situations and create significant value for developers, entrepreneurs, customers, authors, and readers.”²

Amazon’s business platforms are enablers. They enable writers and booksellers and small business owners. They enable developers and IT professionals to focus on adding business value rather than worrying about technology infrastructure. Amazon platforms enable businesses to outsource labor and empower entrepreneurs to pursue personal and professional growth.

And all of the energy it creates eventually comes back. The virtuous cycles of Amazon business platforms circulate and expand energy just like the Amazon flywheel itself.

THE ECHO: AMAZON’S IOT TROJAN HORSE

Amazon’s Echo looks simple enough—it’s a cylinder-shaped consumer electronic device, sleek, dark, and unassuming. Don’t be taken in. That trim black exterior is just the pretty face on a bundle of features, each designed to enable the Internet of Things.

With its sophisticated form—a seven-microphone array, subwoofer and tweeter speakers, advanced voice-recognition software, remote control, and onboard computer with processors, memory, and power supply—even the earliest version of the Echo seemed to early adopters an innovative gadget for their home. They could ask Alexa all kinds of convenient things. Things like,

• “Alexa, play my country playlist.”
• “Alexa, what is the weather forecast?”
• “Alexa, add Gatorade to my shopping list.”

Over the past year, the Alexa teams have continued to add to her list of skills, with over one thousand skills built by other companies and developers. Alexa’s integration with third-party devices and services is growing exponentially:

• “Alexa, arrange an Uber to the Seattle airport.”
• “Alexa, tell Garage.io to close the garage door.”
• “Alexa, how much gas is in my Ford car?”

The teams behind Alexa and Echo are constantly working with third-party vendors like Uber, Garage.io, and Ford to add integrations with their products. Each of these third-party applications can be downloaded to and managed through your Echo.

It takes coordination to integrate each of these new capabilities. Within Amazon’s organizational structure, Echo is composed of several hardware teams (to create both internal and external device hardware), several software teams, and several partner teams to create these external relationships with partners. Each of these teams has its own independent strategy, product roadmap, business plan, and adoption scenario. And each of these teams is actively coordinating with the strategies, roadmaps, business plans, and adoption scenarios of Echo’s other teams.

Watching the growth of Echo’s “talents” over the last year, it has become clear that, in Amazon’s mind, the Echo is far more than just an interactive home speaker. To Amazon, it’s a new type of computing interface that helps customers interface with their connected devices.

Think of Echo as Amazon’s first Internet of Things PC. It’s a computing device that performs calculations and beams queries to your connected devices, but instead of a keyboard and mouse acting as the primary interface, it uses voice for input and sound for output. And if Echo is Amazon’s connection to the Internet of Things, the bridge between home and Internet, then Alexa herself becomes much more than just a disembodied voice spouting facts and playing music.

Alexa is Amazon’s first hands-free operating system.

THE SECRET SAUCE

When Echo launched in 2015 as a plug-in, always-on listening device, it was the only device using Alexa. Since then, Amazon has introduced the Tap, a portable battery-operated version of the Echo, and the Dot, a smaller Alexa interface. There are rumors also of a voice-activated Kindle tablet on the way.

In the end, though, all of these devices are really just the container for Amazon’s secret sauce—Alexa.

At its simplest, Alexa provides three critical capabilities:

1. Sophisticated Speech-Recognition Capabilities. Similar to Apple’s Siri, Alexa Voice Services allows connected devices like Echo to recognize and assign meaning to users’ vocal commands. Alexa Voice Services uses cloud computing and machine learning to improve its recognition capabilities based on individual interactions with users. And, much like Google Search uses machine learning to improve itself with each use, Alexa Voice Services pools its learning across all users to improve user interactions and the accuracy of search results.
2. Event Triggering. Alexa provides event recognition, a rules engine, and the technology interface to run third-party applications. At Amazon, each custom application is referred to as a skill. For example, an Uber skill allows a user to request an Uber ride. There is a combination of tools built by Amazon to enable this event recognition capability, but it essentially consists of defining a keyword to trigger your skill, a set of voice-command keywords, such as “start” or “order,” and then a list of rules on what to do or what to return. There are hundreds of third-party skills available today, and the platform is just getting started.
3. Software Platform for Voice Interaction and Integration. Finally, Alexa can be used as a software platform on other products, similar in nature to an operating system. Amazon allows other companies to download and license the Alexa software and use it on their own devices and products for free.

Along with all of this, Amazon has created an investment fund, the Alexa Fund, providing “up to $100 million” to help developers, manufacturers, and startups build new voice-driven technologies and applications using Alexa’s developer tools and APIs.²⁹ All types of companies, products, and services can embed Alexa into their products.

“We’ve made adding Alexa incredibly easy for developers,” said Greg Hart, vice president of Amazon Echo and Alexa Voice Services. “Any device with a speaker, an Internet connection, and a microphone can integrate Alexa with just a few lines of code. From startups to established companies, we can’t wait to see how developers integrate Alexa.”³⁰

What can companies do with this? Toymail uses Alexa Voice Services (AVS) as the technical backbone and operating system for toys that allow kids to send and receive voice messages. Scout Alarm uses AVS for voice control and integration in their connected home-security system.

What is Amazon’s motivation or rationale for giving Alexa away? They’re making a long-term bet on the development of Alexa: they’re betting that they can get other companies to develop Alexa’s skills and to use Alexa as the voice-recognition system in their products—that Alexa will become a major platform for devices that use voice interaction and event triggering, which, as it turns out, is a large segment of IoT scenarios.

Alexa is a Trojan horse. It’s a Trojan horse that creates many opportunities.

First, like the Amazon Dash, Alexa (via the Echo device) provides customers with a direct line to Amazon product purchases. The more customers that use Alexa, the more product purchases Amazon will see.

Second, Alexa will be most natively integrated into AWS, a fact that will drive the consumption of a remarkable amount of storage, computing, network bandwidth, and other higher-end cloud capabilities. At the moment, as it incubates and supports its new ecosystem, Amazon is paying for the computing costs Alexa incurs on AWS infrastructure. Down the line, though, as the ecosystem grows more and more robust, they will likely start working with application providers to charge for consumption of AWS services. (We’ll talk about this more in principle 10.)

Third, both Alexa’s voice-recognition and search-processing capabilities are built using machine learning. Machine learning requires more data to learn. That means that the more consumers use Alexa, the better the results will be—much like Google Search, which increases its accuracy and value with each new piece of history and context. So, in these early days of voice-recognition capability, it is critical to become the market leader and drive improvement through use.

THE IOT PLATFORM OPPORTUNITY

What can other companies take from this IoT platform strategy? Other than “consider using Alexa as your voice-recognition software solution” (this is not a trivial suggestion!), the potential is this: building connected devices could be a Trojan horse for becoming the embedded infrastructure for your customers. You have a chance to build a platform that is valuable for other companies to participate and leverage your connected devices in ways you could monetize.

When you become the operating infrastructure, your products become much harder to replace. Your partners will teach you a tremendous amount about the usage and problems of your product and services. Your revenue model will likely be accretive and recurring. And when your product and capabilities can be “programmed” into environments, likely through an API, the cost of replacement skyrockets. Ripping and replacing programmed infrastructure is hard and expensive.

Of course, Amazon isn’t alone in recognizing the power of platform. Apple, Facebook, and Google have all leveraged some variation with similar traits. Together, Apple, Facebook, Google, and Amazon have come to be known as the Gang of Four. As Eric Schmidt commented, “It seems to me that there are four companies that are exploiting platform strategies really well.”³¹

Among these, consumer-based platform strategies—like the Apple App Store and Facebook—are easiest to recognize. But business-to-business (B2B) platforms, though fewer, present more opportunity.

In B2B industries, there will be many opportunities to create unique platforms within the realm of IoT based simply on where your company’s products are already located and how they’re already being used.

A company that manufactures and installs hand dryers, for example, could leverage the physical position of its equipment to create a platform for monitoring and serving bathrooms and patrons. Example scenarios could include notifying cleaning teams of needed supply replacements, reordering supplies themselves, and providing security services. The maintenance ecosystem could leverage the “hand-dryer platform” to monitor water on the floor to avoid slip-and-fall situations. Video monitoring, voice and event identification, and other data sensors, all potentially positioned from the physical unit, could be valuable to many other companies.

You now have a platform.

Bigbelly, a trash-equipment company, is already building out an IoT platform strategy around another unglamorous but huge opportunity—garbage cans. By thinking in concentric circles, Bigbelly has created a better municipal and public trash receptacle—solar-powered sensors monitor capacity and supply power to a compactor, which creates up to eight times more capacity. This has also allowed Bigbelly to add new services and revenue opportunities—a “clean management system” provides real-time data to drive operational improvements and serves as a platform for an ecosystem of additional services. Security and Wi-Fi hotspots are two early use cases.

This pattern of leveraging infrastructure to allow others to access, operate, and extend their services is the same platform strategy Amazon has repeatedly used to innovate and gain scale.

But this is so much more than a technical and architecture question. The most common mistake companies and leaders make is thinking that platforms are, at their core, a technology challenge.

As Phil Simon wrote in The Age of the Platform, “An overall mind-set based on openness and third party collaboration is absolutely essential in building a true platform. That mentality is much more important than any individual API or technology.”³²

How to monetize? How to partner and build the ecosystem? How to operate and update the infrastructure? How to manage risk and liability? How to provide high availability? How to do updates? How to manage security? A successful platform play will need to address all of these and more.

IoT is providing this set of opportunities to many more companies going forward, as each device category opens up a new set of scenarios. Amazon is positioning the Echo to be the house IoT platform, car brands around the world are maneuvering to become the auto IoT platform, and GE is turning the aircraft carrier to become the industrial IoT platform. But there are many more IoT platform opportunities, essentially each unique operations or physical domain—buildings, bathrooms, emergency rooms, basketballs, delivery trucks, even shoes.

Any device that is in a situation that is a good “perch,” or crow’s nest, for monitoring, collecting data and events, and allowing other entities to leverage (and pay) for the access to that perch is potentially in a position to become a platform.

That being said, though, building a platform will only be the right strategy for a minority of companies. Becoming a platform is yet another big change to add on top of the transitions needed to move from a product company to an IoT-enabled service company.

ARE YOU IN A GOOD POSITION TO BECOME A PLATFORM COMPANY?

As you consider whether your company might benefit from investing in the infrastructure and services it takes to create an IoT-based platform, evaluate the following.

1. Location and Positioning of Product. The location and positioning of your product or company assets can determine the strategic nature of the data and events that could be captured. Do you have a “perch”? A well-positioned product provides a valuable crow’s nest, whereas a product with narrow visibility has less potential to act as a platform. Positioning is also a big determinant to the connectivity options available to the device, which impacts both costs and the amount, speed, and reliability of data connection and transport.
2. Form Factor of Product. Does the form factor of your product provide a basis for integrating other sensors and the associated infrastructure? Could your device allow for other sensors to be located on it? Incumbents, because of their install base and industry relationships, can have an advantage here, but their success should not be taken for granted. Look at connected-home startups like Nest and Scout as companies that are disrupting incumbents.
3. Technology Chops. What’s your pedigree or your interest in being a technology company? You will become a software and technology company if you decide to proceed with a platform strategy. This is a major strategy decision that should include many considerations. Should the platform be generally “open” or “closed” with a limited number of partners?
4. Operating Environment Complexity. You now have to operate a high-availability platform within a complex ecosystem. When an issue happens or a software update needs to be made to a device, that adds complexity to testing and release. Data-sharing rights, security, updates, and reporting will all need to be added to your list of operating considerations.
5. Play the Long Game or the Short Game? Becoming a successful platform company typically requires playing the long game. It takes time and investment to build competencies like testing platforms, APIs, and developer engagement. It takes time to drive customer adoption and use of the platform. If you’re a company that needs to find short-term, revenue-generating wins, becoming a platform may not be the strategy for you.

Old-school technology people are likely familiar with the acronym SMOP—“simple matter of programming.” Typically used as an ironic response to “All you need to do is program the technology to accomplish blah,” SMOP points out that it is always more complex than just “programming the capability.” Perhaps SMOP needs to stand for “simple matter of (being a) platform.” Being a platform company may sound like “just develop the software, and make it available,” but becoming a platform company, or releasing IoT products that you want others to build on and integrate with, is much more complicated than just a SMOP.

In my first book, I wrote a section called “The Critical Capabilities of a Platform Company,” which gives a framework for thinking through the implications of becoming a platform company. Business-model implications, operational implications, ecosystem implications, and a baseline set of critical capabilities that a platform likely needs to provide are outlined. For example, a critical capability a platform needs to provide is to help manage data and content in a trusted manner.

For traditional product and service companies evaluating making an IoT platform, these considerations hold true. The best IoT use cases flow across multiple products and typically across multiple enterprises. The prior discussed example, Audi, DHL, and Amazon have partnered in Germany to enable a package to be delivered to the trunk of your car. Structuring and operating ongoing partnerships like this is complex and vital for compelling IoT capabilities, and the skills to build these partnerships put a premium on “partnership” capabilities in organizations that may traditionally not have partnerships like this. Partnership capability will be an organizational muscle that is developed for IoT business to flourish.

As has been and will be throughout this book, “think big, but start small” is the ethos that innovators and companies becoming innovators will adopt. So get going on your IoT platform strategy!