[Harvard Must Read] Business Model Innovation 04: Beyond the Pipeline, Toward the Platform
경영전략 비즈니스모델 플랫폼 하버드머스트리드
Beyond the Pipeline, Toward the Platform
Pipelines, Platforms, and the New Rules of Strategy
Through Lesson 3 we looked at how to protect an existing model while growing a new one. Lesson 4 turns to the most powerful form that new model can take: the platform. A platform does not make a product and sell it. It connects producers and consumers and earns value from the interactions between them. Drawing on two Harvard Business Review articles, this lesson covers how pipelines differ from platforms, why every platform starts with a chicken-and-egg problem, and which choices separate platforms that succeed from those that fail.
In 2007, five companies (Nokia, Samsung, Motorola, Sony Ericsson, and LG) together captured about 90% of the global mobile phone industry's profits. That year Apple launched the iPhone. By 2015 the iPhone alone generated 92% of the industry's profits, and most of the former leaders were making no profit at all. Why did companies with strong brands and differentiated products collapse? The authors' answer lies in the structure of the business, not the quality of the product.
A pipeline company brings in raw materials, adds value at each step, and hands a finished product to the customer along a linear chain. A platform company connects producers and consumers and earns value from what happens between them. Apple's handset business is a pipeline. Once the App Store was added, Apple became a pipeline and a platform at once, and when the two kinds of business compete in the same market, the platform often wins.
Runs a linear chain from raw material to finished product efficiently. Value is created inside the firm and flows one way to the customer.
Builds a place where producers and consumers meet and interact. Participants create the value; the firm enables and governs the interaction.
Resources the firm owns and controls: factories, inventory, patents, brand.
The network of participants, the resources they bring, and the quality of their interactions. The firm does not need to own them, only to orchestrate them.
A platform is not a place where value is made directly. It is a place that makes possible the interactions through which value moves.
The article explains that when a pipeline firm moves to platform logic, the center of management shifts in three places.
From Resource Control to Orchestration
A pipeline's strength came from owning and controlling scarce resources. On a platform, the valuable asset moves to the network of participants and what they contribute. The winner is not the one who holds the most assets but the one who attracts and connects the people who hold them.
From Internal Optimization to External Interaction
A pipeline sharpens its internal processes to raise efficiency. A platform's value comes from outside participants meeting one another. The leader is no longer someone who gives orders, but someone who persuades, sets the rules, and designs governance.
From Customer Value to Ecosystem Value
A pipeline focuses on growing the value of each individual customer. A platform aims to grow the value of the whole ecosystem, sometimes taking a loss on one group in order to attract another.
A platform creates value only when both sides are present. Game developers will not build for a console nobody owns, and players will not buy a console without games. A credit card is useless without merchants, and merchants have no reason to accept it without cardholders. When the size of one side raises the value of the other, we call it a cross-side network effect. Value can also grow as more participants join the same side, but a platform usually begins with the cross-side effect.
The standard answer in Eisenmann, Parker, and Van Alstyne's "Strategies for Two-Sided Markets" (HBR, 2006) is to subsidize one side. Lower the barrier with a low price or a free offer for the side that is price-sensitive or that determines the platform's quality, and collect revenue from the side that gains most from the other side's growth. The first is the subsidy side; the second is the money side.
Consoles are sold at or below cost, and the profit comes from royalties paid by developers. The article puts those royalties at up to 20% of a game's retail price.
The reader side is given away free, building an installed base of hundreds of millions, while the side that creates documents pays for the software.
Users are the money side and developers the subsidy side. The article notes that Microsoft gave its Windows development kits away while Apple charged developers $10,000, and that Windows ended up with six times as many applications as the Mac.
The point is not to lower prices for everyone. Decide which side is more sensitive to price and which gains most from the other side's growth, and place the burden on one of them.
A platform is strong once it takes hold, but there are several ways to stumble on the way there and afterward. These are four patterns that recur across the two articles.
Covisint, a B2B exchange for the auto industry, lost momentum when many parts makers refused to join for fear of price pressure. Without both sides, the cross-side effect never starts.
Too much control and the community withers; too much openness and noise takes over. The article cites Myspace as a platform run top-down like a newspaper, and Chatroulette as one whose openness let noise grow.
Congestion, bad actors, and noise drive good participants away and start a downward spiral. That is why Airbnb, Uber, and app stores filter and rate their participants.
An adjacent platform bundles your function into its own service and offers it more cheaply. Microsoft bundled streaming software with its own products, and in 2003, 42% of North American internet users named Windows Media Player as their primary player, against 19% for RealPlayer.
The Core Interaction
A platform is built around one core interaction: a single ride, a single video view, a single "like." It can be tiny and need not involve money. Make that one interaction highly valuable first, then expand to adjacent ones.
Subsidy Side and Money Side
In a two-sided market, price is designed to fall on one side. Draw in the side that is price-sensitive or sets the quality, and earn revenue from the side that profits from the other's growth.
Two Faces of Network Effects
More participants can raise value for everyone, but congestion and bad behavior can erode it. Both are things to design for.
Balancing Openness and Control
A platform must decide how open to make its tools and resources (architecture) and its rules and rewards (governance). More openness brings more participants, but they leave if rewards are unfair, and noise grows if no one is filtered out.
On a platform, the quality of the interaction comes before scale. If participants join but do not meet in ways that create value, the network effect never begins.
Before price is a way to make revenue, it is a choice about which side to call in first.
A platform's enemy is not only the competitor in the same industry. Once an adjacent platform bundles your function into its own offer, industry boundaries stop mattering.
What stood out most in Lesson 4 is that a platform's fate is decided before the quality of the app or the algorithm, by the question of whom to bring in, in what order, and at what price. Technology can be copied, but the order in which one side gathers and calls the other is hard to recover once it is missed. Platform strategy, I came to feel, is less product planning than designing the order in which a market opens.
The lesson reads more clearly alongside Lesson 3. Apple's handset business is a pipeline, and the App Store layered on top of it is a platform. Two kinds of logic live in one company. The pipeline's metrics are margin and inventory turnover; the platform's are the number of participants and the quality of interactions. The problem the ambidextrous organization had to solve, not judging a new model by the old model's metrics, comes back here unchanged. The article's choice of platform metrics points the same way: interaction failures (for example, calling a ride and seeing no cars), engagement, match quality, and negative network effects.
What I consider most important in Lesson 4 is sequencing and balance: the order in which to call each side, the balance between openness and control, the balance between growth and quality. Platform management is decided by how you make these few choices. Habits from the pipeline era, making the inside more efficient and controlling more, can become obstacles on a platform.
A platform is not a question of what to make, but of who gets to meet whom. And it is a question of who starts that meeting first, and on what terms.
✦ Lesson 4 Summary
A platform is a business model that, instead of making and selling a product, connects producers and consumers and earns value from their interactions. In moving from pipeline to platform, the center of management shifts from resource control to orchestration, from internal optimization to external interaction, and from customer value to ecosystem value.
The starting point of a platform is solving the structure in which each side needs the other. Open the market by subsidizing one side, earn from the other, manage negative network effects through a balance of openness and control, and prepare for envelopment by adjacent platforms. Technology is only a tool; the strategy is which side you call in, and in what order.
Marshall W. Van Alstyne, Geoffrey G. Parker, and Sangeet Paul Choudary, "Pipelines, Platforms, and the New Rules of Strategy," Harvard Business Review, April 2016.
Thomas Eisenmann, Geoffrey Parker, and Marshall W. Van Alstyne, "Strategies for Two-Sided Markets," Harvard Business Review, October 2006.
The figures and examples in the text come from the two articles as of the time they were written. The situation may have changed since.