Note: While reading a book whenever I come across something interesting, I highlight it on my Kindle. Later I turn those highlights into a blogpost. It is not a complete summary of the book. These are my notes which I intend to go back to later. Let’s start!

  • A two-sided market (with both producers and consumers) gives rise to four kinds of network effects: same-side effects (positive and negative) and cross-side effects (positive and negative). A growing platform business must manage all four

  • The key to minimizing most negative network effects is quality curation, which increases the chances of a happy match between producer and consumer

  • Participants + Value Unit + Filter → Core Interaction

  • Google’s search engine acts in a fundamentally similar way. Google’s crawlers search the web, creating web page indices (value units). A consumer types in a query. Google combines the query with other specified inputs, such as social signals—the volume of “likes,” retweets, comments, and other responses received by a particular posting on the Internet. This combination of inputs constitutes the filter, which determines which value units are delivered to the consumer

  • When designing a platform, your first and most important job is to decide what your core interaction will be, and then to define the participants, the value units, and the filters to make such core interactions possible

  • Platforms must perform three key functions in order to encourage a high volume of valuable core interactions, which we summarize as pull, facilitate, and match. The platform must pull the producers and consumers to the platform, which enables interactions among them. It must facilitate their interactions by providing them with tools and rules that make it easy for them to connect and that encourage valuable exchanges (while discouraging others). And it must match producers and consumers effectively by using information about each to connect them in ways they will find mutually rewarding

  • One powerful tool that encourages users to keep returning to the platform is the feedback loop. A feedback loop in a platform may take various forms, all of which serve to create a constant stream of self-reinforcing activity. In the typical feedback loop, a flow of value units is that generates a response from the user. If the units are relevant and interesting, the user will be drawn to the platform repeatedly, generating a further flow of value units and facilitating more interactions

  • One kind of feedback loop is the single-user feedback loop. This involves an algorithm built into the platform infrastructure that analyzes user activity, draws conclusions about the user’s interests, preferences, and needs, and recommends new value units and connections that the user is likely to find valuable

  • In a multi-user feedback loop, activity from a producer is delivered to relevant consumers, whose activity in turn is fed back to the producer. When effective, this creates a virtuous cycle, encouraging activity on both sides and ultimately strengthening network effects. Facebook’s news feed is a classic multiuser feedback loop

  • Facilitating interactions may also involve reducing barriers to usage. Not long ago, a Facebook user who wanted to share photos with friends had to use a camera, transfer the images to a computer, use Photoshop or another software package to edit them, and finally upload them to Facebook. Instagram enabled users to snap, modify, and share pictures in three clicks on a single device. Lowering barriers to usage in this way encourages interactions and helps expand participation on the platform

  • In some cases, increasing barriers has a positive effect on usage. Sittercity is platform that helps parents find babysitters. To inspire trust among its users (the parents), Sittercity has imposed a stringent set of rules that restrict those who can sign up as producers (the babysitters)

  • A successful platform creates efficiencies by matching the right users with one another and ensuring that the most relevant goods and services are exchanged. It accomplishes this by using data about producers, consumers, the value units created, and the goods and services to be exchanged. The more data the platform has to work with—and the better designed the algorithms used to collect, organize, sort, parse, and interpret the data—the more accurate the filters, the more relevant and useful the information exchanged, and the more rewarding the ultimate match between producer and consumer

  • In its search for new drivers, Uber discovered that many of its best prospects were recent immigrants to the U.S. who were eager to supplement their incomes by driving for Uber but who lacked the credit histories and financial qualifications needed to finance car purchases. Andrew Chapin of Uber’s driver operations group came up with the idea of having Uber act as a middleman to guarantee car loans for its drivers, deducting repayments from driver revenue and sending them directly to the lenders. Finance companies like the program because loans backed by Uber’s massive corporate cash flow are almost risk-free, and local auto dealers are happy with the additional inventory turnover

  • By changing the value unit exchanged between existing users (as when LinkedIn shifted the basis of information exchange from user profiles to discussion posts)
    • By introducing a new category of users as either producers or consumers (as when LinkedIn invited recruiters and advertisers to join the platform as producers)
    • By allowing users to exchange new kinds of value units (as when Uber and Lyft made it possible for riders to share rides as well as arranging solo pickups)
    • By curating members of an existing user group to create a new category of users (as when LinkedIn designated certain participants as “thought leaders” and invited them to become producers of informational posts)
  • The design of a platform should begin with its core interaction—one kind of interaction that is at the heart of the platform’s value-creation mission

  • Three key elements define the core interaction: the participants, the value unit, and the filter. Of these, the value unit is the most crucial, and often the most difficult to control

  • In order to make the core interaction easy and even inevitable, a platform must perform three crucial functions: pull, facilitate, and match. All three are essential, and each has its special challenges

  • As a platform grows, it often finds ways to expand beyond the core interaction. New kinds of interactions may be layered on top of the core interaction, often attracting new participants in the process

  • It’s important to design a platform thoughtfully to make mutually satisfying interactions easy for large numbers of users. But it’s also important to leave room for serendipity and the unexpected, since users themselves will find new ways to create value on the platform

  • Efficient pipelines eat inefficient ones pipelines. Platforms eat pipelines

  • Platforms enjoy two significant economic advantages over pipelines

  • One of these advantages is superior marginal economics of production and distribution. As we’ve noted, when hotel chains like Hilton and Sheraton want to expand, they build new rooms and employ thousands of staff. By contrast, Airbnb expands with near-zero marginal costs, since its cost for adding an additional room to its network listings is minimal

  • A platform’s ability to scale rapidly is further enhanced by network effects. When positive network effects kick in, higher production leads to higher consumption, and vice versa

  • Pipeline businesses like Nike have traditionally scaled in one of two ways. Some expand by owning and integrating a greater length of the value-creation-and-delivery pipeline—for example, by buying upstream suppliers or downstream distributors. This is referred to as vertical integration. Others expand by widening the pipeline to push more value through it. This is horizontal integration. When consumer goods companies grow by creating new products and brands, it’s an example of horizontal integration

  • Lately Nike has also been developing apps—in this case, apps related to sports and fitness. On the surface, these might seem like traditional product-line extensions aiming at horizontal integration. But in reality, Nike is testing an approach that, if successful, will lead to a new form of growth—one pioneered by platform businesses like Apple

  • Platforms are able to outcompete pipelines because of their superior marginal economics and because of the value produced by positive network effects. As a result, platforms are growing faster than pipelines and taking leading positions in industries once dominated by pipelines

  • The rise of platforms is also disrupting business in other ways. It is reconfiguring value creation to tap new sources of supply; reconfiguring value consumption by enabling new forms of consumer behavior; and reconfiguring quality control through community-driven curation

  • The rise of platforms is also causing structural changes in many industries—specifically, through the phenomena of re-intermediation, separation of ownership and control, and market aggregation

  • Incumbent companies can fight back against platform-driven disruption by studying their own industries through a platform lens and beginning to build their own value-creating ecosystems, as Nike and GE are doing

  • One way to address this conundrum is to avoid the chicken-or-egg problem altogether by building a platform business on the foundation of an existing pipeline or product business. This approach is known as:
    1. The follow-the-rabbit strategy. Use a non-platform demonstration project to model success, thereby attracting both users and producers to a new platform erected on your project’s proven infrastructure
    2. The piggyback strategy. Connect with an existing user base from a different platform and stage the creation of value units in order to recruit those users to participate in your platform
    3. The seeding strategy. Create value units that will be relevant to at least one set of potential users. When these users are attracted to the platform, other sets of users who want to engage in interactions with them will follow
    4. The marquee strategy. Provide incentives to attract members of a key user set onto your platform. In many cases, there’s a single group of users who are so important that their participation can make or break the success of the platform. It may, therefore, make sense for the platform manager to incentivize their participation, either through a cash payment or through other special benefits
    5. The single-side strategy. Create a business around products or services that benefit a single set of users; later, convert the business into a platform business by attracting a second set of users who want to engage in interactions with the first set. Launching a service booking platform like OpenTable, the restaurant reservation system, poses a classic chicken-or-egg problem. Without a large base of participating restaurants, why would patrons visit the OpenTable site? But without a large base of patrons, why would restaurants choose to participate? OpenTable solved the problem by first distributing booking management software that restaurants could use to manage their seating inventory. Once OpenTable had enough restaurants on board, they built out the consumer side, which allowed them to start booking tables and collecting a lead generation fee from the restaurants
    6. The producer evangelism strategy: Design your platform to attract producers, who can induce their customers to become users of the platform. Platforms that provide businesses with tools for customer relationship management (CRM) can often solve the chicken-or-egg problem simply by attracting one set of users—producers—who then take on the task of bringing along the other set—consumers—from their own customer base. The platform helps the producers cater to their existing set of consumers, and over time, the producers benefit from data-driven cross-pollination as other consumers on the network become interested in their products and services
    7. The big-bang adoption strategy. Use one or more traditional push marketing strategies to attract a high volume of interest and attention to your platform. This triggers a simultaneous on-boarding effect, creating an almost fully-developed network virtually instantaneously
    8. The micromarket strategy. Start by targeting a tiny market that comprises members who are already engaging in interactions. This enables the platform to provide the effective matchmaking characteristic of a large market even in the earliest stages of growth
  • Four key elements are necessary to begin the process of viral growth for a platform business—the sender, the value unit, the external network, and the recipient. Let’s consider the viral growth of Instagram:
    1. The sender. A user on Instagram shares a picture that he has just created. This launches the cycle that will eventually bring in a new user
    2. The value unit. On Instagram, the value unit is the picture that the user shares with friends
    3. The external network. For Instagram, Facebook serves as a very effective external network, allowing value units (photos) to spread and be exposed to potential users
    4. The recipient. Finally, a user from Facebook gets intrigued by the picture and visits Instagram. This user may create her own photo and start the cycle all over again. Now the recipient is acting as the sender
  • Unlike its competitor Hipstamatic, Instagram didn’t simply allow users to save, organize, and filter pictures. It encouraged them to share their photos on external networks like Facebook, converting a single-user activity into a social, multi-user activity. Every time users engaged the app, they shared their creations. Every point of app usage became an instance of app marketing. In essence, Instagram converted all its users into marketers

  • One difference between platform businesses and traditional pipeline businesses is that, in the world of platforms, pull strategies designed to encourage virality are more important than the push strategies (such as advertising and public relations) used in conventional marketing

  • Successful platforms use one of eight proven strategies for solving the chicken-or-egg problem:
    1. follow-the-rabbit strategy
    2. piggyback strategy
    3. seeding strategy
    4. marquee strategy
    5. single-side strategy
    6. producer evangelism strategy
    7. big bang adoption strategy
    8. micromarket strategy
  • The speed of a platform’s expansion can be accelerated through viral growth. This depends on four key elements:
    1. sender
    2. value unit
    3. external network
    4. recipient
  • A well-managed platform can create excess value in four ways:
    1. access to value creation
    2. access to the market
    3. access to tools
    4. access to curation
  • Monetization is about capturing a portion of the excess value created

  • Techniques for monetizing a platform include charging a transaction fee, charging users for enhanced access, charging third-party producers for access to a community, and charging a subscription fee for enhanced curation

  • One of the most crucial monetization choices is deciding whom to charge, since the difference in roles played by various platform users means that charging them can have widely differing network effects

  • Given the complexity of the monetization challenge, platform managers should take potential monetization strategies into account in every decision they make regarding platform design

  • There are three kinds of openness decisions that managers face: those regarding manager/sponsor participation, developer participation, and user participation

  • Management and sponsorship of a platform may be controlled by a single firm, by different firms, or by groups of firms

  • The four possible combinations lead to differing patterns of openness and control, with various advantages and disadvantages

  • The open/closed dichotomy isn’t black or white. There are shades of gray, and benefits and drawbacks to every point on the spectrum Sometimes, similar platforms choose to compete on the basis of differing openness policies

  • Maturing platforms often evolve in the direction of greater openness. This demands continually reevaluating and adjusting curation processes to ensure consistently high quality of platform content and service value

  • Governance is the set of rules concerning who gets to participate in an ecosystem, how to divide the value, and how to resolve conflicts

  • Three fundamental rules of good governance:
    • Always create value for the consumers you serve
    • Don’t use your power to change the rules in your favor
    • Don’t take more than a fair share of the wealth
  • Market failure—a situation in which “good” interactions (fair and mutually satisfactory) fail to occur, or “bad” interactions do

  • There are four main causes of market failures: information asymmetry, externalities, monopoly power, and risk

    • Information asymmetry arises whenever one party to an interaction knows facts that other parties don’t and uses that knowledge for personal advantage. Consider the problem of counterfeit goods, when the seller knows that the goods are fake but does not inform the buyer
    • Externalities occur when spillover costs or benefits accrue to anyone not involved in a given interaction. Imagine that one of your friends provides your private contact information to a gaming company in exchange for a few digital points. This would be a bad interaction because it violates your privacy rights, and so it is an example of a negative externality.The concept of a positive externality is a bit more ambiguous. Consider what happens when Netflix analyzes the movie-watching behavior of someone whose tastes match yours and uses this data to give you a more accurate movie recommendation. This would be a positive externality, because it provides you with a benefit based on an interaction in which you’re not directly involved. Individuals who benefit from positive externalities aren’t likely to complain about them—but they are considered problems from a business design standpoint, since they reflect value that is not being fully captured by the platform. A concept closely related to the positive externality is the public good, whose value is not fully captured by the party that created it. Individuals generally create too few public goods unless some governing mechanism is designed to recognize and reward them
    • Monopoly power arises when one supplier in an ecosystem becomes too powerful because of its control of the supply of a widely sought good, and uses this power to demand higher prices or special favors. At the height of its popularity (2009–10), game-maker Zynga became excessively powerful on Facebook, leading to conflicts over issues such as the sharing of user information, the split of gaming revenues, and the cost to Zynga of ads on the social network
    • Risk is the possibility that something unexpected and essentially unpredictable may go wrong, turning a good interaction into a bad one. Risk is a perennial problem in all markets, not just on platforms. A well-designed market generally develops tools and systems that serve to mitigate the effects of risk, thereby encouraging participants to engage in more interactions
  • Alvin Roth described a model of governance that uses four broad levers to address market failures. According to Roth, a well-designed market increases the safety of the market via transparency, quality, or insurance, thereby enabling good interactions to occur. It provides thickness, which enables participants from different sides of a multisided market to find one another more easily. It minimizes congestion, which hampers successful searches when too many people participate or low quality drives out high. And it minimizes repugnant activity—which explains why platform designers forbid porn on iTunes, human organ sales on Alibaba, and child labor on Upwork. According to Roth, good governance occurs when market managers use these levers to address market failures

  • Self- Governance rules followed by Intel Architecture Labs in launching the USB standard:
    1. Give customers a voice in key decisions. Use a separate business unit, with a “Chinese wall,” to handle conflicting agendas
    2. For trusting relationships, open standards must remain open
    3. Treat IP [intellectual property] fairly, yours and theirs
    4. Communicate a clear road map and stick to it. Commitments to act or not act must be credible
    5. Reserve the right to enter strategically important markets with notice. Don’t surprise people and don’t play favorites with news
    6. In case of big investments, share risk and bet your own money
    7. Do not promise to not change the platform. Do promise early notice. Have skin in the game, so change bites the platform, not just the partner
    8. It’s okay to offer differential benefits to partners with differentiated assets. Just make sure everyone understands how to qualify
    9. Promote the long-term financial health of partners, especially smaller ones
    10. As the business matures, decisions increasingly favor outward progression from core platform, to complements, to new businesses that cannibalize the platform
  • Governance is necessary because absolutely free markets are prone to failures.

  • The basic tools for platform governance include laws, norms, architecture, and markets. Each must be designed and implemented with care in order to encourage platform participants to engage in positive behaviors, incentivize good interactions, and discourage bad interactions

  • Self-governance is also crucial to effective platform management. Well-run platforms govern their own activities following the principles of transparency and participation

  • To define success or failure for a platform, and to identify how to improve it, there are three main metrics: liquidity, matching quality, and trust

  • Liquidity in a platform marketplace is a state in which there are a minimum number of producers and consumers and the percentage of successful interactions is high. When liquidity is achieved, interaction failure is minimized, and the intent of users to interact is consistently satisfied within a reasonable period of time. Achieving liquidity is the first and most important milestone in the life cycle of a platform. Therefore, the most valuable metric in the early months of a platform is one that can help you determine when liquidity is reached

  • One reasonable way to measure liquidity is by tracking the percentage of listings that lead to interactions within a given time period. Of course, both the definition of “interactions” and the appropriate time period will vary depending on the market category

  • On the negative side, it’s important to look for and track the occurrence of illiquid situations. These are circumstances in which a desired transaction is impossible—for example, when an Uber user opens the app and discovers that no car is available. Illiquid situations discourage users from participating in the platform and so must be kept to a minimum

  • Matching quality is critical to delivering value and stimulating the long-term growth and success of the platform. It is achieved through excellence in product or service curation

  • Precision in matching leads to lower search costs for users—that is, they need to invest less time, energy, effort, and other resources in finding the matches they want. Thus, if the platform does a great job of linking users to one another quickly and accurately, those users are likely to become active participants and long-term members of the platform; if matching quality is poor, slow, and disappointing, users will soon dwindle in number, interactions will slow to a trickle, and the platform may be doomed to an early demise

  • Trust refers to the degree to which users of a platform feel comfortable with the level of risk associated with engaging in interactions on the platform. It is achieved through excellent curation of participants in the platform. Building trust, of course, is central to marketplaces, especially those in which interactions carry some level of risk—and in the world of online platforms, where initial connections among users as well as many interactions are conducted entirely in cyberspace, the perception of risk may be even more significant. A well-run platform is one in which participants on both sides have been successfully curated so that users are comfortable with the level of risk involved in engaging in interactions on the platform

  • Since the value of a platform is derived primarily from network effects, platform metrics should ultimately seek to measure the rate of interaction success and the factors that contribute to it. Interaction success attracts active users and thereby enhances the development of positive network effects
    • During the startup phase, platform companies should concentrate on metrics that track the strength of characteristics that enable core interactions on the platform, including liquidity, matching, and trust. These characteristics can be measured in a variety of specific ways, depending on the nature of the platform
    • During the growth phase, platform companies should focus on metrics that are likely to impact growth and enhanced value creation, such as the relative size of various portions of the user base, the lifetime value of producers and consumers, and the sales conversion rate
    • During the maturity phase, platform companies should focus on metrics that drive innovation by identifying new functionalities that can create value for users, as well as metrics that can identify strategic threats from competitors to which the platform needs to respond
  • Rather than re-dividing a pie of more-or-less static size, platform businesses often grow the pie (as, for example, Amazon has done by innovating new models, such as self-publishing and publishing on demand, within the traditional book industry) or create an alternative pie that taps new markets and sources of supply (as Airbnb and Uber have done alongside the traditional hotel and taxi industries). Actively managing network effects changes the shape of markets rather than taking them as fixed

  • Platforms turn businesses inside out, moving managerial influence from inside to outside the firm’s boundaries. Thus, a firm no longer needs to seize every new opportunity on its own; instead, it can pursue only the best opportunities while helping ecosystem partners seize the others, with all partners sharing the value they jointly create

  • Within the platform ecosystem, the lead firm negotiates dynamic tradeoffs involving competition at three levels: platform against platform, platform against partner, and partner against partner
    • At the first level, one platform competes with another, as in the video game console battles among Sony (PlayStation), Microsoft (Xbox), and Nintendo (Wii). Strategic advantage is based not on the attractiveness of particular products or services but rather on the power of entire ecosystems. The Sony PlayStation Portable was a stronger gaming device than the iPhone, which lacked specialized left and right controls. When Sony released the PSP-2000 in fall of 2007, after Apple’s summer release of the iPhone, Sony’s stock rose about 10 percent. But before long, the iPhone ecosystem vastly outstripped that of the PSP. As we’ve already noted, Apple has subsequently enjoyed far great financial success than Sony, thanks in large part to the size and value of its ecosystem
    • At the second level, a platform competes with its partners—for example, when Microsoft appropriates such partner innovations as browsers, multithreading, streaming media, and instant messaging and incorporates them into its operating system, or when Amazon operates as a platform for independent merchants while also selling some of the same goods on the same platform in competition with those merchants. This is a delicate and dangerous move. It can strengthen the platform, but at the expense of weakening partners—a short-term gain that can lead to painful long-term consequences
    • At the third level, two unrelated platform partners compete for positions within the platform ecosystem, as when two game app developers strive to attract the same consumers on the same console
  • Platforms seek exclusive access to essential assets. They do this, in part, by developing rules, practices, and protocols that discourage multihoming. Multihoming occurs when users engage in similar types of interactions on more than one platform

  • Platform businesses seek to discourage multihoming, since it facilitates switching—when a user abandons one platform in favor of another. Limiting multihoming is a cardinal competitive tactic for platforms

  • Here’s an example of how the effort to limit multihoming plays out in the new world of strategy. Adobe Flash Player is a browser app that delivers Internet content to users, including audio/video playback and real-time game play. Flash could have been used by app developers on Apple’s iPhone operating system—but Apple prevented this by making its iOS incompatible with Flash and insisting that developers use similar tools created by Apple itself. Developers and users responded with dismay, and some observers called the policy an anti-competitive gambit that might be subject to governmental sanction under antitrust regulations. Adobe had designed Flash developer tools to allow content and program porting from Apple iOS to Google Android and to web pages more generally. Apps developed in Flash could multihome, reducing the iPhone’s distinctiveness. Adobe also released extensions that allowed in-app purchases. By allowing developers to take interactions off the iTunes platform, Flash would cause Apple to lose its 30 percent cut of every interaction as well as control over the associated usage data—information that provides valuable clues concerning trends in the marketplace. If Apple had supported Flash, it would have granted users access to enormous amounts of Flash content already on the web while giving developers more ways to monetize their investments by multihoming across platforms. But it would have been a big loss for Apple. So they used licensing rules and technology to keep interactions from going off platform

  • When Alibaba was struggling to attract traffic, CEO Jack Ma and his team had made a counterintuitive decision: they created technological barriers that prevented Baidu from searching their website. Baidu is China’s biggest Internet search engine—the Google of China. Blocking Baidu’s bots from searching Alibaba for products being sought by Baidu users cut off a vast supply of potential customers. Doing this at a time when Alibaba was desperate for shoppers must have seemed a bit crazy. But Alibaba’s leaders were playing a long-term strategic game. They had their eye not just on the shopping interactions that would take place on their platform but on the potential to monetize the platform by selling advertising. They were determined to retain control of the community of would-be shoppers that they were gradually building on Alibaba—so that Alibaba alone would be able to sell ads aimed at those shoppers. Barring Baidu’s bots from Alibaba’s listings was a way of preventing Baidu from hosting the consumer-oriented ads that companies would ultimately want to direct to the growing numbers of Chinese online shoppers—and making sure that those ads would appear on Alibaba’s platform instead. The strategy worked. As Alibaba’s user base expanded, it gradually displaced Baidu as the most valuable online advertising platform in China

  • A platform business need not own all the inimitable resources in its ecosystem, but it should seek to own the resources whose value is greatest. This is why Alibaba (rather than Baidu) owns search on its platform, why Facebook (rather than Google) owns search on its platform, and why Microsoft (rather than some outside software developer) owns Word, PowerPoint, and Excel on its platform. These are all crucial resources that create value for the majority of the platforms’ users—that’s why it is critical for the platform owners to control them. Less valuable, or more niche, resources can be ceded to ecosystem partners without significantly weakening the competitive position of the platform itself

  • Some notable platform strategy battles have been won by companies that took advantage of data supremacy to outcompete their rivals. By most measures, Monster should have won the battle for supremacy among job placement platforms. As one of the earliest entrants into the market, it had a first mover advantage, and it quickly generated strong network effects in the two-sided market of employers and employees seeking one another. However, there were built-in limitations to the data Monster gathered. Because Monster targeted only active job seekers, it captured no information concerning users’ broader social networks. And once a particular job search interaction was completed, both employer and employee would leave the platform, halting the flow of data altogether. By contrast, LinkedIn targeted the social networks of all professionals, not just active job seekers. This led to a higher degree of ongoing engagement and captured data from those who were happily employed but willing to consider new job opportunities—a vastly expanded user base. LinkedIn also captured data from interactions among professionals interacting with one another as well as with recruiters, providing two separate feedback loops on the same platform. Later, LinkedIn began emphasizing content creation and sharing by its users to foster additional reasons for them to spend time on the platform. LinkedIn’s big edge in the scope, depth, and volume of marketplace data has given it a huge advantage in competing against Monster.

  • Platform design can be optimized in various ways to generate better user data

  • Platform businesses don’t need to own all the critical assets as long as they have access to them in their ecosystems, platform companies can pursue fewer M & A deals than many traditional firms feel compelled to do. In the process, they enjoy at least two significant benefits. First, claiming a portion of the value created by a platform partner is far less risky than buying that partner Second, keeping a partnership at arm’s length reduces the platform’s technological complexity. As the very term vertical integration implies, any new business a platform purchases must be integrated with the platform, which creates both technical and strategic challenges. A platform built from a dozen independently developed technologies will break sooner, cost more, and deliver a worse customer experience than one built on a lean architecture that conducts all its business activity through clean interfaces

  • Platform managers need to continually scan the horizon, observing the activities of other platforms—particularly platforms that serve similar or overlapping user bases. We call these adjacent platforms. When a new feature appears on an adjacent platform, it may represent a competitive threat, since there’s a possibility that users of your platform may find the new feature attractive enough to begin multihoming or even to abandon your platform altogether. To respond, the platform manager can choose either to provide a similar feature directly or to offer it indirectly via an ecosystem partner. When used most successfully, this strategy leads to the phenomenon we call platform envelopment. This occurs when one platform effectively absorbs the functions—and the user base—of an adjacent platform

  • Of course, the opportunities and threats run both ways. If Platform A is trying to envelop adjacent Platform B by developing a feature that competes with Platform B’s most attractive offering, Platform B may try to envelop Platform A by mounting the same kind of attack in reverse. In this kind of envelopment battle, the larger platform, with its more numerous initial user base and more powerful network effects, is generally triumphant

  • Platforms compete by trying to improve the quality of the tools they provide to pull in users, facilitate interactions, and match producers with consumers

  • In some cases, superior platform design enables a platform to dramatically outcompete a preexisting rival. Airbnb initially had far fewer users than the much older Craigslist, which also provides listings of rooms and apartments for short-term rent. However, Airbnb did a much better job at the key platform functions of facilitation and matching. On Craigslist, a person seeking a room for rent had to drill down through an unmanaged list of options clustered by city and organized by time of posting. By contrast, Airbnb allowed a person to search through options organized not only according to these characteristics but also by quality, number of rooms, price, and mapped geolocation. Furthermore, users could also strike deals directly through Airbnb, whereas Craigslist users had to go off platform to make a rental agreement. This made Airbnb far easier to use and enabled the platform to rapidly outgrow the erstwhile category leader

  • Winner-take-all market is a market in which specific forces conspire to encourage users to gravitate toward one platform and to abandon others. The four forces that most often characterize winner-take-all markets are supply economies of scale, strong network effects, high multihoming or switching costs, and lack of niche specialization

  • Opponents of regulation point to phenomena like regulatory capture to argue that government intervention in business is usually ineffectual. But history suggests that some level of societal regulation of business is healthy and beneficial both to the economy and to society as a whole

  • There are a number of regulatory issues that are unique to platform businesses or that require fresh thinking in the light of the economic changes that platforms are causing. These include access to platforms, compatibility, fair pricing, data privacy and security, national control of information assets, tax policy, and labor regulation

  • The flood of new data made available by today’s information age technologies suggests the possibility of new regulatory approaches based on after-the-fact transparency and accountability rather than restrictions on market access. But such new approaches will need to be designed thoughtfully and carefully to fully protect the public

  • Economic frameworks for industries with network effects suggest that dominance alone is not necessarily cause for government intervention. Rather, failure to manage externalities, abuse of dominance, manipulating populations, and delaying innovation can indicate when intervention in platform markets is necessary and appropriate

  • Some of the types of businesses that are most likely to join the platform revolution in the years to come:
    • Information-intensive industries. In most industries today, information is an important source of value—but the more crucial information is as a value source, the closer the industry is to being transformed by platforms. This explains why media and telecom are two of the industries that have already been disrupted so thoroughly by platforms. New entrants have created ecosystems that can create and disrupt content and software more quickly and easily than large firms with thousands of employees once did
    • Industries with non-scalable gatekeepers. Retailing and publishing are two examples of industries that traditionally have employed expensive, non-scalable human gatekeepers—buyers and inventory managers in the case of retail, editors in the case of publishing. Both are already undergoing disruption thanks to the rise of digital platforms, with millions of producers (artisans, craftspeople, writers) creating and marketing their own goods through platforms like Etsy, eBay, and Amazon
    • Highly fragmented industries. Market aggregation through a platform increases efficiencies and reduces search costs for businesses and individuals looking for goods and services created by far-flung local producers. Platforms ranging from Yelp and OpenTable to Etsy, Uber, and Airbnb have made it easy for customers to visit a single source to gain access to thousands of small suppliers 
    • Industries characterized by extreme information asymmetries. Economic theory suggests that fair, efficient markets require that all participants have equal access to information about goods, services, prices, and other crucial variables. But in many traditional markets, one set of participants has far better access than others. Used car dealers, for example, knew much more about the condition and history of the cars they sold, as well as about supply and demand variables, than their customers—hence the distrust in which they were held. Data aggregating and sharing platforms such as Carfax are now leveling the field, making detailed information about used car values available to anyone willing to pay a small fee. Other markets where information asymmetries have made fair dealing difficult, from health insurance to home mortgages, are ripe for similar changes
  • Industries that might seem to be susceptible to platform approaches, yet are likely to be resistant to such disruption, have certain characteristics:
    • Industries with high regulatory control. Banking, health care, and education are all highly regulated. Regulations favor incumbents and work against the interests of startups trying to unlock new sources of value. Emerging platforms are starting to attack this problem in an effort to create new sources of value, but regulatory control is holding them back
    • Industries with high failure costs. The costs of a defaulted loan or matching a patient with the wrong doctor are much higher than the cost of showing inappropriate content on a media platform. Consumers are reluctant to participate on platforms when the perceived costs of failure are high
    • Resource-intensive industries. Resource-intensive industries have typically not been dramatically affected by the Internet. Winning participants in these markets still depend on their access to resources and their ability to manage efficient, large-scale processes such as mining, oil and gas exploration, and agriculture, in which information has a limited role to play
  • Education is perhaps the prime example of a major industry that is ripe for platform disruption. Information-intensive? Check. In fact, the fundamental product being sold by schools, colleges, and universities is information of various kinds. Non-scalable gatekeepers? Check. Ask any parent whose child has recently had to navigate the slow, complex, inherently arbitrary process by which a lucky handful of students are admitted to the most prestigious and selective colleges, and you’ll hear plenty about the shortcomings of some of the world’s most powerful gatekeepers. Highly fragmented? Check. Within the United States, there are over 13,000 public school districts, as well as thousands of private school systems, colleges, universities, and proprietary schools, each fiercely independent and proud of its unique programs and standards. Information asymmetry? Check. Only a small percentage of parents feel competent to judge the qualifications and reputations of schools and colleges—hence the proliferating, competing, and confusing array of rating systems and the ever-increasing pressure on students to earn entry to the handful of institutions that are universally admired—the Harvards and Yales of the world. The drive to build education platforms is well under way, as businesses like Skillshare, Udemy, Coursera, edX, Khan Academy, and others suggest. Eager to avoid being rendered irrelevant or obsolete by upstart platform companies, a number of the world’s greatest universities are moving to position themselves as leaders in this educational revolution. Institutions including Harvard, Princeton, Stanford, the University of Pennsylvania, and many others are offering online versions of some of their most popular classes in the form of “massive open online courses” (MOOCs)—many in partnership with companies like Coursera. In the years to come, the spread and increasing popularity of teaching and learning ecosystems will have an enormous impact on public school systems, private schools, and traditional universities. Barriers to entry that have long made a first-class education an exclusive, expensive, and highly prestigious luxury good are already beginning to fall. Platform technologies are making it possible for hundreds of thousands of students to simultaneously attend lectures by the world’s most skilled instructors, at minimal cost, and available anywhere in the world that the Internet is accessible. It seems to be only a matter of time before the equivalent of a degree from MIT in chemical engineering will be available at minimal cost in a village in sub-Saharan Africa

  • Industries that are most prone to platform transformation in the near future include those that are information-intensive, those with unscalable gatekeepers, those that are highly fragmented, and those characterized by extreme information asymmetries

  • Industries that are less likely to be transformed by platforms in the short run include those with high regulatory control and high failure costs as well as those that are resource-intensive

  • It’s possible to foresee some of the specific changes that are likely to impact selected industries in the decades ahead, including education, health care, energy, and finance

  • The platform model will continue to shape transformations in the markets for labor and professional services as well as the operations of government

  • The burgeoning Internet of things will add a new layer of connectivity and power to the platforms of the future, linking people and devices to one another in new value-creating ways

  • The platform revolution will ultimately transform our world in unpredictable ways, calling for society as a whole to develop creative, humane responses to the challenges this change will produce