How to Design a Platform Business
I like the way Marshall van Alstyne conceptualizes platform business design.[i] He defines platforms as an open architecture with rules of governance designed to facilitate interactions. Platform design consists of three factors: interactions, architecture and governance. See also his work with Geoff Parker and Sangeet Paul Choudary as you can read in the book: Platform Revolution.[ii]
I built my work on this conceptualization of platforms.[iii] Let’s look at the three factors:
Interaction
It is the starting point of platform design since interactions are the source of value. When a GoPro surfer uploads an image or video to the GoPro site and shares it with other like-minded action sports enthusiasts, the simple transaction creates value. The simpler the transaction, the easier it is to scale the platform, but I believe that to solve many of today’s challenges, more complex transactions are necessary.
Van Alstyne distinguishes between volume and value of a set of interactions. A single internet search has trivial value but trillions of occurrences. By contrast, a single stay at Airbnb has far more transaction value but far less frequency.[iv] Without frequency, there is no value creation.
I conceptualize an interaction in terms of volume or frequency and in terms of quality.[v] Three characteristics contribute to quality: meaning, reciprocity, and value/benefit. Meaning is about the degree to which the interaction expresses the intent of the business, its mission or purpose, or its brand. Take John Deere as an example. The collection of data from a farmer is a transaction that occurs with relative frequency on farm fields today, but it is not an interaction that has a lot of meaning. However, when the data is being collected by a company like John Deere whose brand promise is to create shared business value and whose mission is to improve farm profitability, the interactions around data collection take on much greater meaning.
The quality increases as the meaning grows deeper. Deere has a mission of “serving those close to the land.” As they say, the company is for those who cultivate and harvest the land, for those who transform and enrich the land, and for those who build upon the land. In this regard, the interactions that the company enables have a good deal of meaning for farmers but also for other participants in the ecosystem. This is a company that has been involved in farming for nigh on two centuries, has a brand trusted by farmers, and is committed to goals beyond its own corporate growth and profit.
Reciprocity is about the mutuality of the exchange. Is it truly an interaction? Do all the participants actually participate? Do they take and give? When the farmer contributes data to Deere, does it share data of similar significance in return? (Yes, it does.) Or is the interaction more like a disguised transaction? For example, when I do a lot of transactions with a credit card company, an airline, or a bank, I accumulate points in what seems like an interaction. But when I take a look at what the points are worth, the amount seems insignificant and almost insulting. That is all the company thinks of me and my information?
The third attribute of interaction quality is value, in terms of the benefit to participants. An interaction might have a good deal of meaning and reciprocity but not much value. Deere, for example, might provide the farmer with a lot of information that looks interesting, but is actually generic or doesn’t have much relevance or applicability to the farming operation. The question here is, how does an interaction create value of improving farm yield?
There is also the prospect of new value. FBN, together with Deere, enables higher price transparency for corn and soybean seeds, which lowers the overall seed costs. Interactions, therefore, can create negative externalities for some participants in the interaction field. If Deere offered an electric tractor with fewer mechanical parts, this would have a negative impact on the revenue that dealers could earn by providing maintenance services. The farmer will always be calculating the value of the interaction. If Deere’s analytics enable him to reduce costs, cut the time spent in operation, or increase yield, it has value.
We must also look at value in terms of how many participants it affects. If it benefits the community, the industry or other participants in the farming ecosystem, and the society—as well as the company and the farmer—it is of higher value. The more cohesive and tightly connected or linked the interaction is—that is, the higher the number and greater the quality of interactions—the more velocity it will achieve and the more everyone will benefit from the three effects: virality, network effect, and learning.
The number of interactions increases through virality. As word spreads about the Deere offering, participants join at a faster rate and the number of interactions climbs. A large number of interactions is also essential to gaining the network effect; that is, the product or service becomes more valuable as more participants join. Finally, the learning that results from the work of an interaction field is plowed back into the company and its activities. When the whole system is thriving—the number and quality of interactions is growing — a virtuous cycle is created that keeps the field healthy and sustainable.
Architecture
Van Alstyne describes various types of architecture.[vi] There are those with a narrow versus broad focus and an open versus closed architecture. These are important tradeoffs especially when starting a platform from scratch. Van Alstyne describes the initial failure of Alibaba that started too broad but found success with a narrower vertical structure. Lyft launched in San Francisco and got traction there before it expanded in other cities.
I think of architecture in terms of three different layers since the interactions are very different from layer to layer and hence value creation is different as well. These are the nucleus, ecosystem, and market-makers.
The nucleus of participants is typically the company, like John Deere or GoPro, and the customers—anyone who contributes to the core interactions on a regular basis. The traditional company has already established a business relationship with the participants in the nucleus, which is the foundation of the interactions. The ecosystem of contributors is composed of partners in the company’s business activity. It is not uncommon that some data is shared between the nucleus participants and the ecosystem participants. Business ecosystems are built on relationships that have been established over years. An example is the supplier relationship between Bosch, the automotive electronics company, and Daimler, the car manufacturer. They have a well-established supplier-buyer relationship based on the development, manufacture, and sale of electronic components for Mercedes-Benz vehicles.
The third group of participants, or layer in a platform architecture, is the market-makers. These are entities that exert influence and enable velocity – namely, the frequency and quality of interactions. There are many types of entities that can be market-makers, and the types differ from one interaction field to another. The US Department of Transportation, for example, regulates the automotive industry and is one type of market-maker in an automaker’s interaction field. Consumers who could potentially be attracted to the platform because they want to solve their transportation needs, but have not yet purchased vehicles, are another type of market-maker. Daimler has merged the Car2Go carsharing interaction field with that of BMW’s DriveNow. Potential drivers who don’t currently use the offerings of these two companies are important market-makers. The better the merged service is positioned to pull new drivers toward it, the more velocity the field gains.
Market-makers can also be entities such as research institutes, like the Fraunhofer Institute, or university researchers who develop automotive technology. Velocity depends greatly on the market-makers. Whether they are new consumers attracted to the field, competitors, government agencies or regulators, or participants in other platforms, market-makers can significantly determine the success or failure of the company in creating value. High interaction velocity is achieved when the three elements of the field work together to create network effects, learning effects, and virality.
Governance
Governance provides the rules of who may participate, how they create and divide value, and how to resolve conflict.[vii] The success of a platform business rises and falls with proper fair governance and regulation. One of the most recent discussions on regulation that facilitates data sharing that delivers value for all participants in a platform business can be found here: [viii]
Designing a platform business requires a complex set of decisions or choices regarding three areas: first is the governance structure which is about defining the participants, the roles and responsibilities, defining decision right and powers, etc. Second is the definition of the decision-making structure. If John Deere builds the platform, should John Deere be making the decisions or should the ecosystem participants be included in the decision-making? Third is, how will the rules be enforced? Amazon has penalized many third-party sellers for their acts and deeds, for example. Sometimes these rules are reinforced automatically through bots or AI, and sometimes they are monitored through employees or third-party contractors.
Flatiron Health is an example that shows that governance be managed even in complex industries such as healthcare, where patients share incredibly sensitive information for the greater benefits of everyone. Flatiron Health is now a Roche Pharma company that brings together patients, care providers in cancer hospitals, and even competing pharmaceutical companies and regulators like the Food and Drug Administration. Over two million cancer patients, and 2,500 clinicians in 800 unique sites of care share “research-grade” data with the larger healthcare and pharmaceutical community regulatory bodies such as the FDA or the National Cancer Institute. Competitors such as 14 of the top 15 life sciences companies participate. The success of Flatiron Health has significantly cut the average years of FDA approval of new therapeutic solutions and has let to much more effective treat cancer patients. The success wouldn’t happen without a well-thought-out governance structure.
Three Types of Platform Businesses
The three factors of a platform lead to three types of platform businesses: platforms, digital ecosystems, and interaction fields. I believe platforms and digital ecosystems are transitory models. Ultimately, these models need to evolve toward interaction fields models. What, then, is an interaction field business model or company? Let me describe three major features:
First, interaction fields are interactional and not just transactional. Platforms are known to be a powerful force of competition. They build an infrastructure to orchestrate transactions between providers and consumers, between riders and drivers (Uber or Lyft are examples), between hosts and travelers (Airbnb), between buyers and sellers of books, for example.
They learn a great deal from the frequency and number of transactions. They make markets more efficient; they remove frictions and create value for customers and consumers. They are so successful in competing against incumbents, they become incredibly disruptive across a swath of categories and industries.
In New York City, where I live, driving a taxi used to be a way to make money for many immigrants. It was a way to establish a new life and earn a living as an entrepreneur. Long hours were required to catch enough fares, but there was also the value of the license or “taxi medallion.” This license would go up in value and could be sold by a driver at the end of his career, much like an entrepreneur could sell his or her restaurant, retail store, or small business after many years of labor.
If you started early in the 1960s, you could get a license for $25,000. By 2005, the license cost $325,000. Five years later, it was $600,000 and in 2013, the value was over a million dollars. Then, suddenly the price dropped by 45% over two years. By 2019, a medallion license was worth nearly nothing. In one auction, sixteen medallions were offered; three sold for less than $140,000 and 13 medallions had no bidders.[ix]
What happened? Uber and Lyft and other ridesharing platforms had come to the city. There are now about 80,000 for-hire vehicles on the road. Two-thirds are from ridesharing platforms. 13,500 are traditional medallion taxis.[x] There is now a lot more congestion in midtown New York City, and most drivers don’t make enough anymore since there aren’t simply enough fares out there. Economists call it “negative externalities.” Taxi drivers have lost their investments. If you now need to make a living in the city by driving a for-hire car, you join the gig economy. That means, minimum wage (if any at all), and no benefits, you need to take care of those for yourself. As Douglas Rushkoff writes, they turn lifelong jobs into the temp jobs for the gig economy.[xi]
Uber and the other ridesharing companies have moved on and have become digital ecosystems. There is now Uber Eats, which delivers food from restaurants, and dozens of other businesses, for example.
Platforms are about disruption. They see opportunity in frictions and inefficiencies or lack of innovation. It is often said that taxis were not innovative. The last innovation in the taxi industry was the taxi meter, which was introduced soon after World War II.
Platforms do two things when they are done with an industry. They evolve toward ecosystems like Uber, and in the process look for disruption in adjacent markets, or they move on to another market. I could tell the same story about Amazon starting with book retailing, and when that industry was kaput, it moved to other categories, one after the other. Some people say that’s the nature of business. Tough luck.
But I don’t think that is correct. Platforms and digital ecosystems can be good when they are interaction field models. Interaction field models build on collaboration and participation. They are not just transactional but interactional. An example is Alibaba. Alibaba’s mission is, “Our mission is to make it easy to do business anywhere. With our platform model, we are bringing buyers and sellers from all over the world together, and are best placed to partner with them to meet the needs of the nearly 700 million users on our platform.”[xii] Alibaba is not in the business of disrupting small retailers; they are in the business of making them efficient, removing frictions and enabling them to sell more. In China, there are 6 million small retailers, mom-and-pop shops who sell locally and operate at a relatively basic level. Alibaba offers them a retail management software called Long Shou Tong for free, which helps the stores to become more efficient, optimize the assortment, and sell more across many more parts of China.
Everyone gains; Alibaba gains valuable data from millions of small stores and earns a fee for any online sales. Hundreds of thousands of merchants also can now sell their millions of products locally. Alibaba builds on collaboration, not disruption – it is interactional and benefits everyone.
Second, interaction fields solve new problems that are often intractable challenges or pain points that often haven’t been solved before. Platform and digital ecosystems also do that, but they typically focus on narrow, often existing, problems. GM tried it with Maven, a subscription service to compete with Zipcar. It started in 2016 solving for more flexibility of transportation or mobility. It closed the business in 2020. GM is not an exception; I could now spend the rest of my day describing other car companies’ attempts to do the same.
Tesla is, in my mind, a company that isn’t about a platform or a digital ecosystem. It wants to be an interaction field company. It solves for a lot more than just electric cars, as you know. Besides, it is not even just a car company. It solves for some degrees of autonomous driving and for lower CO2 emission. It is well known that a car is 95 percent idle and so Tesla built its own network where you can post your car while you travel globally. Others can pick up the car and use it during that time. It solves for lower cost of ownership of a car among a horde of other issues. I don’t need to tell much more; the story has been told so many times. Some things are still vision, and some are reality, but the direction is apparent. If Tesla has it, it will have a much more significant share of our lives, solving for multiple problems and challenges we have daily, rather than just selling us a lease of a new Model 3 every three years.
Third, interaction fields are platforms and digital ecosystems that are open and welcome other participants. They propagate being inclusive rather than being exclusive. In today’s discussion of digital ecosystems, the term “ecosystem competition” has become popular.[xiii] This is the notion that ecosystems compete now against other ecosystems and that you as a company need to decide which ecosystem you join if you can’t build your own.
This idea is exactly what was once the dated idea of competition between firms, which came down to the disruption of industries and categories for over hundreds of years, and the same logic applies to platforms and digital ecosystems.
We should be wiser. Is it really about competition between ecosystems? Is it really all about who wins and who loses? Should we really, in this day and age, just think about capturing or extracting value, driving more shareholder value through ecosystem competition?
I don’t think so, I think we need to build interaction fields with healthy governance to ensure fair value distribution – as Marshall Van Alystne says, a situation where you create more value than you take.
[i][i] Platforms and Ecosystems: Enabling the Digital Economy, page 9.
[ii] Parker, Geoffrey, Marshall Van Alstyne, and Sangeet Paul Choudary, Platform Revolution: How Networked Markets Are Transforming the Economy – and How to Make Them Work for You, W.W. Norton & Company, 2016.
[iii] Erich Joachimsthaler, The Interaction Field: The Revolutionary New Way to Create Shared Value for Companies, Customers and Society, Public Affairs, 2020, forthcoming.
[iv] Ibid, p. 10
[v] Erich Joachimsthaler, ibid p. 41 ff
[vi] Ibid, p. 10
[vii] Ibid, p. 10, Parker, Van Alstyne, and Choudary.
[viii] Parker, Geoffrey and Petropoulos, Georgios and Van Alstyne, Marshall W., Digital Platforms and Antitrust (May 22, 2020). Available at SSRN: https://ssrn.com/abstract=
[ix] Wikipedia entry: Taxi Medallion, accessed June 6, 2020. https://en.wikipedia.org/wiki/Taxi_medallion#:~:text=The%20price%20rose%20steadily.,around%202013%20at%20over%20%241%2C000%2C000.
[x] https://www.wired.com/story/new-york-city-flexes-extending-cap-uber-lyft/
[xi] Douglas Rushkoff, Team Human, W.W. Norton, New York, 2019.
[xii] Statement by Terry von Bitra, General Manager, Europe, Alibaba Group, Germany, page 9 of this excellent report
[xiii] https://www.mckinsey.com/business-functions/mckinsey-digital/our-insights/competing-in-a-world-of-digital-ecosystems