The economics, structure, and behavior of platform ecosystems and organizations

Organizational structure has been a central theme in the 'future of work' discourse. The transition away from hierarchical to networked forms, and the increasing autonomy of individuals and groups with the larger organization has been going on for decades. Nonetheless, as discussed in earlier sections, hierarchical dynamics still pervade the world of business. Why?

One hypothesis is that no other organizing principle has come along to decisively displace hierarchy. Yes, many organizations have tried to relax command-and-control thinking and management theories, as with Stanley McChrystal's task force and Zhang Ruimin's Rendanheyi. However, we may now have an approach to organizational and inter-organizational coordination that will lead to the displacement of hierarchical institutions: ecosystems.

Ecosystem is a term borrowed from biology and reapplied to strategy research and practice. The term has grown in use, increasing sevenfold over the past few years. At the most basic, the term is used to characterize a group of interacting firms that rely on each other's abilities. The term is used in many different ways. For example, focusing on a single company and a community of partners, as in the Slack ecosystem. Alternatively, a given innovation -- like Android -- could be described as the heart of an ecosystem. Or a technology platform, like Amazon's commerce platform, can be seen as the center point of an ecosystem.

Ecosystems are 'a distinct solution to the problem of inter-firm coordination, distinct from the use of alliances, supply chains, or market-based interactions', as Michael Jacobides and his coauthors wrote in Towards a Theory of Ecosystems. Alternatively, we may look at ecosystems as a new way to structure economic relationships between partner, specifically with the goal of lowering the costs of participation and therefore allowing for non-linear scaling.

Ron Adner, in Ecosystem as structure: An actionable construct for strategy, defined an ecosystem in this way, with greatest attention on the structure of alignment of the members:

The ecosystem is defined by the alignment structure of the multilateral set of partners that need to interact in order for a focal value proposition to materialize.

The form of the 'partners' alignment is one of the imprints of those that kick off the ecosystem, and demonstrates one of the ways that the 'hub' plays an outsized role in defining the ecosystem's essential nature.

Note that ecosystems are different from institutions, because they are explicitly about coordinating inter-organizational activities. However, an ecosystem isn't a supply chain, totally dominated by the supply-chain owner: see WalMart, for example, that controls almost every facet of its suppliers' operations.

Here's how Jacobides and co-authors represent the form of an ecosystem and contrast with other forms of value chains:

figure 4-1 Market-based Value System (from Jacobides et al)

Consider the figure above, which could represent a customer buying new tires from Goodyear for his existing Volkswagen. They value chain relates on market standards and 'generic complementarities.

Consider the somewhat different circumstance when Volkswagen manufactured the car, electing the various parts from a vertical supply chain. In the simple case, that includes acquiring tires from someone like Goodyear, but in more complex circumstances The leader in the supply chain can compel participants to cooperate in the joint design and manufacture of parts, as shown in figure 4-2.

figure 4-2 Hierarchy-based Value System (from Jacobides et al)

Contrast with an ecosystem, where 'complementors' -- participants in the ecosystem -- add new value over and above that from the baseline platform product, for example, app developers building on the Android or iOS OSs.

figure 4-3 Ecosystem-based Value System from Jacobides et al)

An ecosystem isn't like a market, either, although it has qualities of regulation and standardization built into its structure. Jacobides makes the case that ecosystems are the result of a 'partly designed' process, where 'hub' partners (the more powerful companies in the ecosystem), shape the rules and the process of ecosystem development:

To give a specific example, even within the Google/Android ecosystem, with Google as the hub and clear rules for complementors, some key handset manufacturers such as Samsung and Motorola are starting to create sub-ecosystems. They allow key app developers to connect via APIs in ways that are specific to their device, so as to “lock them in”.

This is a great example of ecosystems arising from technology platforms. Android is a 'multi-sided' ecosystem based on the open source Android OS, and while Google is the leading hub in the ecosystem, it does not control the ecosystem to the extent that Apple does with iOS. In Jacobides et al's analysis, Android is an example of a 'supermodular/supermodular' ecosystem, where the value of the system is non-linearly increased by each ecosystem participant and the cost of participation is kept low to incent more participation, and innovation is at its greatest because the central product in open source.

In contrast, iOS is firmly under the control of Apple, who maintains a higher bar for participation, and siphons off a much higher proportion of the benefits of the ecosystem (profits, brand, etc.). This is what Jacobides refers to as a 'unique/supermodular ecosystem'. A unique ecosystem's overall value also increases as more participants join, but a greater proportion of the increased value redounds to the hub than to the other participants. In a sense, this ecosystem has more of the 'supply chain' feel to it.

This table capture these and other distinctions, with the two axes being the type of complementarity of production (how the participants align to make the products) versus the type of complementarity of consumption (how end users experience the products).

figure 4-4 Ecosystems Complementarity (after Jacobides et al)

So-called 'two-sided' markets such as Uber, eBay, and Airbnb are not ecosystems in this formulation. They offer supermodular experience to customers of the service, since the value increases as the platforms' offering increase in size. However, the overwhelming majority of Airbnb hosts do not invest much time or energy in getting the rooms they rent integrated into Airbnb: there are only generic complementarities.

I believe Amazon's e-commerce platform is much more demanding for participants, and as such it is a unique/unique ecosystem. Vendors have to invest significant effort to work through the core platform, and those vendors wanting to get the most from the experience may go much farther, for example storing products at Amazon fulfillment centers and developing Amazon 'stores' for their products.

Given that background and conceptual characterization of ecosystems, I'd like to dig into a recent real-world example, and link ecosystems dynamics to the nature of organizations operating vertically within them.

Haier's Life Cycle of Clothing ecosystem

Rather than focusing on the limited idea of improving the performance and cost of washing machines -- while a worthy goal -- Haier and a group of partners reconsidered and recast the purpose of washing machines.

The starting point was to move away from a device-centric view, the life cycle of the washing machine, and shift to the experience of the customer which is really about clothing.

figure 4-5 The Old Lifecycle of Washing Machines

The conventional approach to thinking about washing machines is based on a market-based value system (as in figure 4-1): the washing machine, detergents, and clothing are independent purchases, and only generic complementarities of the various products are brought to bear.

What if, instead of just a means to clean clothes, washing machines are recast as elements in a complex lifecycle of clothing, supported by an ecosystem of clothing manufacturers, detergent makers, and washing machines by Haier and other appliance vendors?

Figure 4-6 Lifecycle of Clothes Ecosystem

Instead of a supply chain, Haier worked with a number of detergent companies and clothing manufacturers to create an ecosystem around the lifecycle of troublesome clothing. They envisioned a system that would allow end customers to avoid the high costs and inconvenience of difficult clothing, such as silk and fur. The ecosystem includes the innovation of new detergents, the matching of specifically designed detergents for specific clothing (see the interaction of detergent and clothing partners), and the creation of new cleaning modes in the washing machines by Haier.

One key element, perhaps the core product of the ecosystem, are new washing machines designed to handle delicate and difficult to handle materials, like silk and fur, innovations like 'air cleaning', and RFID tagging of clothing by clothing manufacturers. This RFID tagging is not solely for supply chain tracking and inventory, but also to indicate how the clothing should be washed in a way that the 'smart' washing machine can read and act on.

One layer of the ecosystem that is not immediately obvious is that the customers' machines are networked, and can see data to Haier, and Haier can supply that data to the other partners. This allows Haier to improve its machines and their operation, but also provide feedback to the clothing and detergent manufacturers. Note as well that automatic restocking of specially formulated detergents can be managed by the networked washing machines directly to partners.

In principle, data about a piece of clothing -- say a silk blouse that has reached its hypothetical end of life after some number of washings -- could lead to the clothing manufacturer and the store where the blouse was purchased being informed, allowing them to take action. For example, the store could send the blouse's owner a promotion for a new silk blouse.

In this ecosystem, all the partners benefit from the interconnections, and the value of information being exchanged and shared is made greater through the various ways partners apply the data for the benefit of the customer, first and foremost, but also for the benefit of all participants.

  • The customer -- because of all the additional shared design and planning by the partners -- has a much better experience: clothes that are correctly cleaned based on their fine-grained properties -- what sort of materials, what degree of dirtiness -- by creating smarter machines that are programmed to use the right detergents, and wash cycles.
  • The clothing manufacturer -- can sell 'difficult' clothing to customers in greater volume since the customers believe the life cycle costs of the clothing will be lower, and the experience -- home cleaning -- will be easier.
  • The detergent manufacturer -- can lock in customers to the right detergents for their clothes, therefore having a stronger value proposition, and the automated restocking increases volume of sales. The ecosystem acts as a barrier to non-ecosystem vendors.
  • The store or ecommerce vendor selling clothes and detergents --  the stores that stock the clothes and detergents (not really shown in the diagram, but implied) benefit from the ecosystem's value proposition to the end customer, and the data provided.
  • The washing machine vendor -- Haier washing machines are the central product in the ecosystem, and the standards for data and protocols for sharing information shape the ecosystem. Since Haier plays such a key role in recruiting and integrating partners, this seems like an Apple iOS analogue, in Jacobides terms a 'unique/supermodular' ecosystem.

I won't overelaborate this discussion with digging into Haier's ecosystem (not shown) for designing and manufacturing the washing machines used in this ecosystem model. In figure 4-3, Jacobides presented the model of an ecosystem where the 'focal firm product' was manufactured and designed through a conventional vertical supply chain. However, Haier operates internally around another ecosystem, one involving both internal groups and external organizations to deign, build, and provision washing machines and a host of other products. As a result, 'below' the clothing lifecycle ecosystem is another, more foundational ecosystem that Haier has crafted to manufacture appliances. And ecosystem economics are at play at both levels.

Extending the Ecosystem

Rent The Runway is a fast-growing women's rental clothing service in the US, one that is attracting a great deal of interest. In essence, customers sign up to rent either a single element of clothing for a sized period -- think wedding dress or formal gown -- or alternatively, sign up to rent a certain number of clothing items, and when one is returned, another is sent. This second model, which works like Netflix DVD rentals is the largest source of revenue and growth for the company.

Consider how Rent the Runway's clothing rental service could figure into an ecosystem. Rather than buying the clothes, Rent the Runway's customers -- all women, at present -- pay a monthly fee to be able to 'rent' a certain number of clothes: dresses, blouses, jeans, jackets, accessories, handbags, etc. The company's Unlimited service ($150 per month) allows customers to rent four wardrobe pieces, and exchange them as many times per month as desired. There is always an option to buy any piece at a discount from full retail, which happens quite a lot.

In this model, Rent the Runway acts as the cleaning and repair service, which means that for the rented pieces the customer may completely sidestep cleaning the clothes, but is likely to nonetheless clean clothes they own, or rent for a prolonged period of time. Note that clothing repair is currently not treated in the Haier clothing lifecycle ecosystem: it is left to end customers to deal with individually.

Currently, Rent The Runway buys its clothing in wholesale instead of creating partnerships with clothing manufacturers. It also buys commercial washing machines and runs the largest dry cleaning service in New Jersey.

I could envision that these two ecosystem models could converge. For example, Haier could find an additional partner who would emulate the Rent the Runway model -- or create a microenterprise within Haier to create such a service -- and then we'd have an extended ecosystem with all the elements.

Implications of Ecosystems Economics for the Organization

In Thriving in an Increasingly Digital Ecosystem, Peter Weill and Stephanie L. Woerner posed the following question with senior executives across 13 large corporations:

Describe a breakthrough organizational change enabled by digitization where your company has significantly changed the way you operate with early indications of good results.

And they analyzed 77 initiatives, leading to a framework with four business models: supplier, omnichannel, ecosystem driver and modular producer.

This model shows the strong dimensional advantage of moving to the upper right quadrant, ecosystem driver, and the relative weakness of the other three quadrants.

One takeaway from the paper's introduction [emphasis mine]:

We had an important insight in the course of our research: that in this period of digital disruption, businesses focused narrowly on value chains were at a disadvantage; they needed to think more broadly about their business ecosystems. Indeed, we found that companies that had 50% or more of their revenues from digital ecosystems and understood their end customers better than their average competitor had 32% higher revenue growth and 27% higher profit margins than their industry averages.

So, the starting point of organizational implications: companies may find higher profits from participating in ecosystems, first of all. But there may be even more benefit to restructuring the organization itself into an ecosystem, foundationally. That is what Zhang Ruimin is seeking to do at Haier.

I believe that shifting to an ecosystem model may require that the hub organization have reached a certain tipping point, where adding the right technology -- the platform -- leads to a comprehensive transition from ‘organizational culture’ to ‘organizational society’. I maintain that this is analogous to the transformation that happens when a town becomes a city.

Geoffrey West and Alan Bettencourt researched the nature of cities and determined that they are non-linear in productivity and performance. As they grow bigger they increase non-linearly in productivity and decrease in waste non-linearly as well. For example, a city of two million people will create more than twice as many patents as a city of one million. Conversely, doubling the size of a city will reduce the per capita number of gas stations.

As discussed at length in previous sections the transition to organizational society means moving past the organization as an army to the organization as a city.  Instead of hierarchical command-and-control, as in a top-down organization, we have horizontal interactions between relatively autonomous players in the context of hub-initiated governance through protocols which rely on members assuming defined roles in the ecosystem, and the sharing of benefits in a direct and emergent way. This is the Haier Rendanheyi model.

The largest remaining question is this: can platform organizations inherit the economics of ecosystems, and become more productive as they grow? Or art least lose productivity at a slower rate than non-platform organizations? Can we finally escape the trap of the 'Fortune 500 disease' that has motivated Chairman Zhang for decades, and let companies grow more efficient as the grow?

Section 1, Social Evolution

Section 2, The Era of Networks

Section 3, Haier, Rendanheyi, and Zhang Ruimin’s Vision

Section 4, The Future of Networks: Platforms and Ecosystems

Section 5, On The Horizon