Platforms are becoming increasingly central to business value creation.  Yet, not all platforms are created equal – some platforms have far more potential to trigger powerful forms of increasing returns that will ultimately marginalize other forms of platforms. In a rapidly evolving environment, it’s important to understand not just the structure, but the dynamics, of different kinds of platforms. I recently had an opportunity to step back and reflect on this issue as I prepared for a speaking engagement.

I had the good fortune of being invited to speak at the Platform Strategy Summit held at the MIT Media Lab a while back. This Summit was organized by Geoffrey Parker, Marshall Van Alstyne and Sangeet Choudary – three guys who have spent a lot of time studying and working with platform-based businesses. There were great speakers and, perhaps equally important, the audience included a lot of people who have been building platform-based businesses.

What is a platform?

So, what did I learn? Well, first of all, that platforms, at least as the organizers think of them, cover a broad range of institutional arrangements. It was clear that they didn’t simply view platforms as technology platforms. In fact, the definition they offer of business platforms is: “a nexus of rules and infrastructure that facilitate interactions among network users” or, alternatively, “a published standard, together with a governance model, that facilitates third party participation.”

Just to get the semantics aside, I tend to prefer the term “performance ecosystems” to describe this space rather than platforms simply because, to me, the term “platforms” creates an image of a much more tightly designed and specified environment than the term “performance ecosystems.” For the purpose of this posting, though, I’ll defer to my hosts and use the term “platform” since it’s clear from their definition that they are talking about roughly the same domains.

So, are all platforms created equal? I argued in my talk that they are not. To make my case, I distinguished four different categories of platforms that are becoming increasingly prominent in the business world (and elsewhere).

Aggregation platforms

The basic focus of these platforms is to bring together a broad array of relevant resources and help users of the platform to connect with the most appropriate resources.  We can think of data aggregation platforms, marketplaces and broker platforms like eBay and Etsy, and contest platforms like InnoCentive or Kaggle.  These platforms tend to be very transaction or task focused – the key is to express a need, get a response, do the deal and move on.  They also tend to operate on a hub and spoke model – all the transactions are brokered by the platform owner and organizer.

Social platforms

These are similar to aggregation platforms in the sense of aggregating a lot of people – think of all the broad-based social platforms we’ve come to know and love: Facebook, Twitter, etc.  Social platforms also include more tightly defined communities of interest that come together around specific shared interests like certain genres of music, types of sports or academic disciplines like history or economics. They differ from aggregation platforms on some key dimensions.  First, they end up building and reinforcing long-term relationships across participants on the platform – it’s not just about doing a transaction or a task but getting to know people around areas of common interest.  Second, they tend to foster mesh networks of relationships rather than hub and spoke interactions – people connect with each other over time in much richer and diverse ways that usually do not involve the platform organizer or owner.

Mobilization platforms

Now, we’re talking about something more than individual tasks or networks of relationships around common interests.  Mobilization platforms ultimately focus on mobilizing participants to engage in some kind of collaborative effort that will take considerable time to accomplish.  Here, we can think about open source software platforms like Linux or Apache.  There are also “process network” platforms that bring together participants in extended business processes like supply networks or distribution operations that help to select and orchestrate participants who need to collaborate in flexible ways over time.  We might also think of mobilization platforms that support social movements like the Arab Spring movement.  Bottom line, these platforms are not just about conversations and interests, they are ultimately focused on moving people to act together to accomplish something beyond the capabilities of any individual participant.

Because of the need for collaborative action over time, these platforms tend to foster longer-term relationships rather than focusing on isolated and short-term transactions or tasks. But the key focus here is to connect with, and mobilize, a given set of people and resources to achieve a shared goal.  The participants, in other words, are often viewed as “static resources” – they have a given set of individual capabilities and the challenge is to mobilize these fixed capabilities to achieve the longer-term goal.

Learning platforms

Now, it starts to get interesting.  These platforms have as their explicit goal to create environments where participants can learn faster and individually achieve higher and higher levels of performance as more and more participants join the platform. These platforms are what I’ve called “creation spaces” in some earlier work.  They have a distinctive configuration.

Their primary unit of organization is a small team or work group that takes on particular performance challenges and where participants work closely together to come up with creative new ways to address the performance challenge.  The emphasis on small teams or work groups is essential because the focus is on a powerful form of learning that involves accessing tacit knowledge. This in turn requires the formation of deep, trust based relationships.  These relationships evolve quickly in small teams or workgroups but are very challenging to scale.

That’s where the second key element of creation spaces comes in – they provide participants with ways to connect with each other beyond the individual team or workgroup to ask questions, share experiences, and get advice.

To my knowledge, there are very few examples of these creation spaces in business, but we can find very large scale creation spaces in arenas as diverse as online war games (for example, World of Warcraft) and online platforms to help musicians develop and refine their remixing skills (ccMixter). They also have emerged in a broad array of extreme sports arenas, including big wave surfing and extreme skiing.

As with social platforms and mobilization platforms, learning platforms critically depend on the ability to build long-term relationships rather than simply focusing on short-term transactions or tasks. Unlike the other platforms, though, learning platforms explicitly do not view participants as “static resources.” On the contrary, they start with the presumption that all participants have the opportunity to draw out more and more of their potential if given the right environment.

Doubling network effects

Let’s be clear.  All four of these platform categories have the potential to create significant economic value for the platform owner and for the participants.  All four have the potential to tap into network effects – the value created by the platform accelerates as more and more participants join the platform. Of course, the flip side of this is that, until and unless the platform draws in a critical mass of participants, it will have only minimal value to participants.  That’s why the vast majority of platform initiatives end up failing – the very thing that creates value once the critical mass is achieved, diminishes value before that point.

But, the fourth category of platform offers a second level of network effect, one that is uniquely associated with that platform.  To understand this, it’s useful to go back to the somewhat outdated example used to describe network effects: the fax machine. If you only have one fax machine, it has negative value – you paid money for it and it does nothing.  The more fax machines that are connected, the more valuable each fax machine becomes.  But, in that example, the individual fax machine is a static resource – its features and functions are a given and don’t change as the number of fax machines multiplies. The increasing value comes from its connections with other machines.

What if we change the assumption, though?  What if each fax machine acquired more features and functions as it connected with more fax machines? What if its features multiplied at a faster rate as more fax machines joined the network? Now, we’d have a second level of network effect – we’d still have the network effects that come by simply increasing the number of fax machines, but now there’s an additional network effect that accrues as each fax machine adds more and more features as a result of interacting with other fax machines.

That’s the potential of learning platforms.  In a world of mounting performance pressure and accelerating change as described in our Shift Index, that learning potential isn’t just an opportunity, it becomes an imperative.  Those platforms that don’t help participants to learn faster and faster as they work together will tend to be marginalized over time, especially once learning platforms become more prevalent. And this also applies to the participants. If we choose to participate on platforms that are not explicitly driven to accelerate learning, we’ll learn at a slower rate than participants who choose to spend their time on learning platforms.

I also want to be clear that learning can and does occur on all four of the platforms, but on the first three platforms, it occurs as an unintended byproduct of participation on the platform. In the course of a conversation with someone on a social network, I’ll often learn something new – but this is incidental and tends to be learning in the form of knowledge transfer.  Knowledge transfer is inherently a diminishing returns game – the more the knowledge spreads across participants, the more the rate of learning will slow down. The only thing that can turn this from a diminishing returns game to an increasing returns game is to shift the focus from knowledge transfer to knowledge creation.

Platforms that are explicitly designed to drive new knowledge creation through collaborative efforts offer the potential for increasing returns to learning – we’ll learn faster as more and more diverse participants join the platform, especially if we can combine the learning that occurs within high performing teams and workgroups with the learning that comes from connecting across a broader environment.

Bottom line

Which platform would you rather be on? In my book, The Power of Pull, I make the case that we’re moving to a world where value creation will depend on harnessing the power of scalable pull platforms, where we can draw out the people and resources that we need, when we need them and where we need them.  The highest form of pull platform is the learning platform – one that helps all participants to learn faster as more and more participants join in.

The good news from a platform owner perspective is that any of the first three categories of platforms can evolve to become learning platforms over time. But, it will likely not happen on its own. It requires thoughtful observation of where the opportunities for learning might be and initiatives to create more functionality on the platform to support and accelerate learning processes.

Four questions become central to harnessing the power of accelerating learning on platforms:

• What is the primary design objective of the platform today – aggregation, social, mobilization or learning?
• Where are there opportunities to accelerate learning (especially through new knowledge creation) among participants today?
• What features or design principles could we deploy to help tap into those learning opportunities today?
• What are the metrics that we could track to determine the rate of learning and how can we monitor those metrics on an ongoing basis to refine our initiatives and accelerate learning even more among participants?

John Hagel is co-chairman for Deloitte LLP’s Center for the Edge with nearly 30 years of experience as a management consultant, author, speaker and entrepreneur. He is the author of numerous books, including “The Power of Pull,” “Net Gain,” “Net Worth,” “Out of the Box” and “The Only Sustainable Edge.” Previously, he was Global Leader of McKinsey’s Strategy Practice and Electronic Commerce Practice (which he founded and led from 1993-2000). John holds a B.A. from Wesleyan University, a B.Phil from Oxford University and a J.D. and MBA from Harvard University.