The biggest symptom of a failed innovation capability is a failure to make the bet to win. There are several reasons for this. The new venture may pose different kinds of risk than those in the core business, and executives may be uncomfortable making a big bet given the risk. Executives may fear that the new business will undermine the existing business and defer investment for that reason. Or executives may not feel prepared to defend the necessary investments to the investment community. These are all reasonable concerns.

But the failure to bet on new ventures often masks something more fundamental. In almost every corporation, there is a systematic avoidance of decisions that challenge the status quo.

Chris Argyris, a professor at Harvard Business School, has described this tendency in terms of models of organizational learning [Argyris, 1992].

Single loop learning is akin to the operation of a thermostat. If the temperature of the room is not at the set point of the thermostat, it closes a switch to turn on the heat or air conditioning. Single loop learning like this presents no issues for organizations. If something is amiss, according to established metrics of operations, people within the organization respond. A failure may implicate the performance of an individual or a department, but it does not challenge the organizational norms themselves.

Figure 9.1 depicts a schematic for single loop learning. The PDCA (Plan-Do-Check-Act) cycle is at the heart of the model. An important addition in Argyris’ model is the box on the left, which represents the governing variables that guide decision-making. These may be thought of as the norms of the culture.

Most well-managed organizations excel at single loop learning. Its apotheosis is Six Sigma. In single loop learning, people observe and address any divergence from expected results.

Single loop learning can be a rigorous discipline, seeking root causes for deviations and addressing them rather than simply resolving the specific instance of a problem. Single loop learning has helped many companies significantly improve their operations. It is not, however, well-suited to most radical innovation.

Double loop learning, on the other hand, is learning about the system itself. There are circumstances under which any system ceases to function effectively. This failure often manifests itself in repeated attempts to solve a problem without success. An intervention may be designed to correct the problem, but the results of the intervention do not meet expectations. The cycle can go on ad infinitum, with repeated initiatives (which often, over time, are recycled).

Success with double loop learning requires challenging the governing variables of the system. To use the analogy above, it requires questioning the notion of the thermostat itself. Figure 9.2 is a schematic for double loop learning.

To be clear, a lot of innovation is possible without challenging the governing variables of the company. In fact, a useful categorization of innovation distinguishes between those types of innovation that operate within the bounds of the dominant business model (Type I innovation) and those that require changes to that model (Type II innovation). Figure 9.3 depicts these Basic Innovation Types.

Type I innovation includes most process innovation. Process innovation fits well within the established governing variables of the business even if it involves radically new technologies. Also included in Type I innovation is the development of faster, better, or cheaper versions of your current offering. These offerings may integrate radically new technologies, but because the new product simply does a better job of delivering the same basic benefit to the customer, it does not require any change in operating assumptions about what works. Similarly, new product features may themselves be quite innovative and may create significant competitive advantage, but they do not challenge “the way things are done around here.” Most innovation of this type succeeds most of the time in well-managed companies.

Type II innovation requires changes to some of the underlying assumptions about what works and does not work for the company. It usually requires violating one or more of the company’s well-established governing variables. Type II innovation includes new business innovation, disruptive innovation, and some forms of “me, too” innovation. This type of innovation is generally very difficult inside existing corporations.

There are three governing variables that are active in most corporations and that can inhibit innovation. The company seeks always to act in accordance with these dicta:

1. To maintain focus
2. To do nothing that could damage the core business
3. To shepherd resources and direct them to the highest value return.

These operating rules make immense sense, all other things being equal. If your core markets are growing and if your competitive position is sustained or growing—in other words, if you have both the potential and the capacity for growth in your current business—then Type I innovation is sufficient for growth—and much lower risk than radical innovation. For Type I innovation, the existing governing variables will very likely support your innovation efforts.

If core markets are declining or your competitive position is eroding, on the other hand, it is very likely that some form of Type II innovation will be required. And such innovation will almost always challenge a core belief. Core beliefs are hard even to surface and challenge; Argyris makes the point that they are often so deeply embedded that they become undiscussable.

How does what Argyris calls an Organizational Defensive Routine (ODR) play out in practice? A goal is stated (and even embraced) that conflicts in some way with an (unstated) governing variable. People deny the conflict and suppress any discussion of it. This suppression assures that the conflict will not have to be addressed, and it acts to protect the core business from external challenges – even those that it might be vital to address.

Take an example related to innovation. Perhaps the espoused goal is to drive substantial growth outside the core business through new business innovation. The dysfunctional pattern starts with an inherent contradiction:

• Explicit Goal: We need to grow outside the core. We will invest in any new business ventures that make sense
• Implicit Guiding Principle: Protect (the current definition of) the core business at all costs.

These statements conflict with one another, but the conflict is not acknowledged. In fact, it is usually undiscussable. Why? It is part of many cultures to have a “can do” attitude; it is simply not acceptable to talk about apparently distant, hypothetical problems; one’s focus should be on whatever it takes to move the project forward.

This seems to make sense, but it can prevent leaders from addressing an underlying issue. When a conflict between the new venture and existing norms arises—when the new business requires significant investment, for example—decision makers tend to default to the (implicit) guiding principles of the core business. The system acts to preserve resources for the core. In these battles, the governing variables of the dominant business almost always win, which means that the innovative venture almost always loses.

It is difficult work to both operate with excellence in the current business and challenge its most established norms. What is required is what Michael Tushman terms “ambidextrous leadership” [O’Reilly and Tushman, 2004]. Ambidextrous leaders develop the capacity both to exploit the current business and to explore new businesses, in spite of the fact that the skills and mindset required for these roles are so different. Figure 9.4 compares the two.

A key to ambidextrous leadership is the ability to discuss conflicts that are created when a company tries to innovate. This means getting past the organizational defensive routines to understand the differing cultural norms required for success with the two types of innovation [Argyris, 1992].

Tushman has identified two types of ambidextrous leadership. The first is structural ambidexterity. In structural ambidexterity, the new venture is managed completely separately from the core business, often in a different location. The new venture has its own functions and develops its own culture and set of governing variables. The executive leading the NewCo likely has attributes of an “explore” mindset. The new business comes together with the established business only at the level of the CEO. This type of ambidextrous leadership works if the CEO has the skills to hold the tension himself or herself and to make difficult decisions about resource allocation in the face of resistance from other key executives.

The second type of ambidextrous leadership is team-centric ambidexterity. It depends on developing, over time, a cadre of leaders who can both exploit the existing business and explore new businesses. These skills must be buttressed by a set of leadership practices and incentives that encourage the right behaviors. According to Tushman, team-centric ambidextrous leadership is more difficult to create and sustain. It will be increasingly important in the future, however, as the half-life of business models becomes shorter and shorter.