From Push to Pull

Viewpoint: October 16, 2005

The McKinsey Quarterly has just published a new article that JSB and I wrote – “Push to Pull – The Next Frontier of Innovation.” For those who want more detail, I posted on my web site a longer working paper that I wrote with JSB on “From Push to Pull – Emerging Models for Mobilizing Resources.

This is a particularly important piece because it represents the opening salvo in our research for a new book. Yes, we’ve just barely published our previous one and we are already starting to work on a new book. As we were writing The Only Sustainable Edge, we became convinced that there is a much bigger story yet to be told.

“Push to Pull” is only one slice of this bigger story, but it is an important slice. It describes a fundamental shift in models for mobilizing resources. Organizational success depends upon effective mobilization of resources. Getting the right resources to the right place at the right time can make the difference between desired impact and catastrophe – something we learned in graphic detail from Hurricane Katrina and New Orleans.

Over the past century, institutions have been perfecting highly efficient approaches to mobilizing resources. These approaches may vary in their details, but they share a common foundation. They are all designed to “push” resources in advance to areas of highest anticipated need. 

– In education, we design standard curricula to expose students to codified information in a pre-determined sequence of experiences.
– In business, we build highly automated plants or service platforms supported by standardized processes seeking to deliver resources to the right place at pre-determined times.
– In technology, we write massive enterprise applications specifying activities that must be performed and resources that must be deployed to meet anticipated demand.

In the past decade, we have seen early signs of a new model for mobilizing resources. Rather than “push”, this new approach focuses on “pull” – creating platforms that help people to reach out, find and access appropriate resources when the need arises. 

Now, when JSB and I talk about pull platforms, many executives immediately assume that we are referring to the lean manufacturing techniques pioneered by companies like Toyota back in the 1950’s. As we make clear in our writing, though, lean manufacturing represents a hybrid between push and pull models – it still contains significant elements of push models. We are talking about something even more profoundly rooted in the principles of pull.

Pull approaches differ significantly from push approaches in terms of how they organize and manage resources. Push approaches are typified by “programs” – tightly scripted specifications of activities designed to be invoked by known parties in pre-determined contexts. Of course, we don’t mean that all push approaches are software programs – we are using this as a broader metaphor to describe one way of organizing activities and resources. Think of thick process manuals in most enterprises or standardized curricula in most primary and secondary educational institutions, not to mention the programming of network television, and you will see that institutions heavily rely on programs of many types to deliver resources in pre-determined contexts.

Pull approaches, in contrast, tend to be implemented on “platforms” designed to flexibly accommodate diverse providers and consumers of resources. These platforms are much more open-ended and designed to evolve based on the learning and changing needs of the participants. Once again, we do not mean to use platforms in the literal sense of a tangible foundation, but in a broader, metaphorical sense to describe frameworks for orchestrating a set of resources that can be configured quickly and easily to serve a broad range of needs. Think of Expedia’s travel service or the emergency ward of a hospital and you will see the contrast with the hard-wired push programs.

Why are pull platforms beginning to emerge? Many organizations are adopting pull platforms as a way to create more flexibility and cope with greater uncertainty. But early adopters are realizing that there is another more compelling value as they gain more experience with these pull platforms. Pull platforms are particularly powerful in fostering innovation, learning and capability building. 

These new pull platforms are emerging in very diverse arenas. Pull platforms are helping to transform the production and distribution of digital media in areas like blogging and music remixing. But it would be a mistake to view pull platforms as limited to digital “fringes”. Global process networks built upon pull platforms are reshaping the global operations of such different industries as apparel, motorcycles and consumer electronics. Learning institutions as diverse as the University of Phoenix and Brown University are deploying pull platforms. These are not just isolated examples – powerful forces are at work shaping the need for an alternative approach to mobilizing resources. These forces ensure that this new model will spread to all arenas of human activity.

Five broad forces are shaping the emergence and evolution of pull platforms:

– Increasing uncertainty
– Growing abundance
– Intensifying competition
– Growing power of customers
– Greater emphasis on learning and improvisation

In environments shaped by these forces, push models are breaking under the strain and pull models are beginning to fill in the gaps.

Here’s the irony. Push models were originally designed to promote efficiency. Advocates of these models acknowledged that these models might limit flexibility and constrain creativity, but they argued that was a small price to pay for the opportunity to cut costs. Yet, as uncertainty increases and competition intensifies, it turns out that push models are less and less able to deliver efficiency. The fundamental assumption of push models is that demand can be predicted reliably enough to define the procedures required to deliver resources to pre-specified locations before the demand actually materializes. Push models therefore require accurate forecasts to function effectively. Uncertainty undermines the ability to forecast and this in turn undermines the ability to push resources to the right place at the right time. Efficiency becomes harder and harder to achieve.

Pull platforms are highly scalable as well as flexible because they embed specialized capabilities into distinct layers that can evolve independently. The lower layers of pull platforms, including such activities as communication and logistics networks tend to focus on high tech capabilities. Upper layers, concentrating on mobilizing individuals and communities to innovate and create new value, tend to focus on high touch capabilities.

Of course, pull platforms and push programs are not mutually exclusive. In fact, pull platforms often contain push programs that can be accessed through their platforms. For example, Amazon or eBay provide robust pull capability for consumers to access on demand an extraordinary abundance of products like books and CD’s. These products were originally made using traditional push manufacturing programs. On the other hand, reflect on the opportunities to further build upon these pull distribution systems by reconfiguring production processes to deliver publishing on demand.

More broadly, however, the forces outlined earlier make it more and more attractive to deploy pull models rather than push models. At the same time, broader deployment of more flexible technologies, tools and infrastructures makes it more viable to design and manage pull models. As a result, pull models will increasingly displace or marginalize push models in broader arenas of human activity.

We are talking about a profound architectural shift. Yes, it is related to technology architecture shifts, but it really requires redefinition of business architectures. The implications are broad:

– Mindsets: key assumptions about what is required for business success will need to be challenged and redefined
– Business definition: companies will face difficult choices regarding business focus
– Strategy: companies will need to determine the best timing for transitioning to pull platforms and what layers of the pull platforms to provide themselves
– Operations: pull platforms reshape the core operating processes of firm – everything from supply chain management to customer relationship management needs to be reconfigured
– Organization: leadership styles and governance mechanisms for effective participation in pull platforms differ significantly from push models
– Rationale for the firm: the firm has traditionally been viewed as a vehicle for efficiently designing and deploying push programs – what role can and should it play as pull platforms become more pervasive?

Like many of the most profound business changes, this architectural shift is beginning at the edge:

– It is starting at the edge of enterprises, rather than deep inside of the enterprise, because it is here that the greatest uncertainty exists. It is also here that push models, with their assumption of centralized control, are less viable (unless a company has enormous market power like Wal-Mart).
– Pull platforms are also beginning to take hold in emerging economies like China and India because these platforms are particularly powerful in supporting bootstrapping activity.
– Finally, pull platforms are emerging at the demographic edge – younger generations more comfortable with the technologies and tools emerging on electronic networks are pioneering both the creation and use of pull platforms to create businesses that grow extremely fast with relatively modest investment.

It is easy to become overwhelmed by the magnitude of change ahead. In our article and working paperwe conclude by focusing on the bottom line for business executives, suggesting tangible and pragmatic actions they can take now. These actions will help to ensure that they are well-positioned for the transition ahead. 

By mastering the techniques required to make this new model work, companies will be able to create substantial economic value. Those who adhere rigidly to the old model will likely destroy significant economic value.