Jack Welch was wrong. Here’s why leaders should embrace chaos

Jack Welch, former CEO of General Electric, once proclaimed, “Variation is evil.” Through management practices like Six Sigma, Welch and many others sought to eradicate risk entirely from the operations of their companies and achieve extreme efficiency, predictability, and reduction of error.

Although Welch was an incredible leader and full of wisdom, in this regard he was spectacularly wrong. The idea that variation, or volatility, must be removed from companies for effective operation is unsound because it constrains ingenuity; it seeks to turn people into machines. It relies on a command-and-control system and hierarchy that does not enable flourishing and breakthrough empowering innovations. It creates an illusion of safety but squelches insight. Capital efficiency is useful, and decades of management practices are geared toward it. Companies should be efficient, but efficiency can be taken too far. 

The growing mountains of capital we see on corporate balance sheets are evidence of a decades-long march toward capital efficiency. Everyone knows companies need to innovate—CEOs are rightly compelled to provide evidence of innovation activity in their annual reports and speeches—yet very little of the activity seems to produce meaningful results. Far too often, innovation becomes theater, and corporate executives demonstrate uncertainty about how to invest balance sheet capital to renew and refresh their businesses. 

Contrast Welch’s approach with the management philosophy of Reed Hastings, founder and CEO of Netflix, who declared in a podcast interview, “Most companies overoptimize for efficiency. . . . The nonintuitive thing is that it is better to be managing chaotically if it’s productive and fertile. Think of the standard model as clear, efficient, sanitary, sterile. Our model is messy, chaotic, and fertile. In the long term, fertile will beat sterile.”

Netflix has applied this philosophy with remarkable success, executing wholesale business model transformations since its founding: first from mailed DVDs to digital delivery, then from digital delivery to content production. Very few large corporations are able to execute a successful transformation; Netflix’s acceptance of a degree of messiness (read: capital inefficiency to try new things) has enabled the company to thrive. The messiness is reinforced through incentive systems and culture, a culture that Netflix built and has maintained since its scrappy beginning.

Organizations wanting to adopt a similar degree of “messiness” through process and metrics alone will find failure if culture is not part of the equation. Strong and rigid organizational designs create fewer errors, less messiness, and fewer insights. 

Time will tell whether Netflix is able to maintain the messiness in the face of market pressure to adopt efficiency, particularly when new opportunities and challenges confront the business model in the form of online gaming, dynamically generated content, and globalized competition. 

The trick in transformative innovation, like in certain kinds of investing, is to remember that it is the magnitude of correctness that matters, not the frequency. This means you can be wrong a lot in the pursuit of transformative innovation because on the occasion when you are right, you are likely to achieve dramatic results. Conversely, in the operation of a scaled business, the frequency of correctness matters more than the magnitude—leaders try hard to avoid anything that looks like a mistake. That tension is what makes innovation so hard. 

The secret to resilience lies in creating systems that enable large-scale, big-magnitude correctness but also frequent misses. In the face of an unknown future, the best strategy for maximizing returns and minimizing risk is to run as many experiments as possible, at the lowest possible cost per experiment. 

Radical innovation requires learning, making mistakes, and messiness. It requires a deliberate degree of capital inefficiency, an acceptance of lower-than-possible near-term RONA. Corporations are optimized for the exact opposite: execution instead of learning, predictability instead of messiness. As companies grow and scale, their operating system—the culture, incentives, governance, processes, and talent—is optimized for scaled, efficient execution, not learning.

The mistake most corporations make in pursuing empowering innovation is that they try to repurpose their existing operating system—designed for efficient, scaled execution—to do something it was never designed to do: operate with a degree of inefficiency to create learning. 

In practice, this looks like a corporation setting up an internal innovation team, giving them some budget, and asking them to find and develop new opportunities for the corporation to pursue.

As Clayton Christensen said, “The worst place to develop a new business model is from within your existing business model.” These corporate innovation teams are far too often subject to the same governance systems, metrics, and processes the scaled organization relies on to operate the existing business. As a result, the innovation that does succeed ends up looking a lot like the existing business, not the transformation everyone hoped for.

What happens when a large corporation tries to redirect the efficient machine toward the inherently inefficient activity of learning and seeking insight?

Early in my career, I helped lead a corporate innovation and investment team. Our job, at a high level, was to consider how to grow the core business into adjacent and transformational opportunities through investment in external startups or by building new ventures internally. Our team reported through the CFO. We were funded as an operating activity by means of a small “tax” on operating businesses that contributed funds and ideas into our funnel. Because of this, our team constantly had to fight to acquire and maintain budget, often competing against alternative uses of capital, like marketing spend, where the return on investment was well understood and the time to payback was quick.

To “speak the language” of the business, we often found ourselves translating plans into IRR models so that we could argue why our projects were a better use of capital than alternatives. Our projects were not typically better as a driver of efficient returns; the opportunities we were considering were intended to create learning and future optionality for the organization, not near-term cash flow. But to win arguments, we had to make models that fit the constraints under which decisions were made so we could show what might happen if everything went exactly according to plan. 

Investment decisions were slow and bureaucratic; because the company was based in Switzerland, our team, based in the United States, sometimes had to travel abroad to secure investment approval from top executives. The pace of our venture building and dealmaking became slower over time as decisions became harder for the organization to make and capital became more difficult to allocate to our uncertain endeavors.

It’s ironic that while the CEO of the company would discuss our projects in speeches to internal audiences and public investors, we were struggling to get the support we needed to make them happen. Our work was exciting and promised a range of potential futures and growth opportunities for the company.

At a certain point, our team felt like animals in a petting zoo, brought out for show but with limited ability to really do what we knew we could do best. The organizational constraints—the incentives, governance, processes, and culture—designed to protect the status quo and ensure the company moved carefully hindered our ability to experiment, learn, and succeed. Ultimately, we recommended shutting down the entire innovation operation because we were not able to have a meaningful, positive impact on the company. 

These examples are typical of what we see across corporate innovation teams and approaches. The systems used to execute the core business efficiently, at scale, and with careful consideration are applied to challenges of learning, not execution. Despite the best intentions of all involved, the innovation efforts fail. The intention to root out errors from the system ironically ensures eventual failure.

The innovation efforts, in many cases, leave the company worse off than if nothing at all had been done. They actually destroy value, not because they do not ultimately produce meaningful results but because they provide an illusion of action when the company could instead be allocating capital and time to activities that would produce more impactful outcomes.

Elliott Parker is founder and CEO of High Alpha Innovation, a venture builder that partners with corporations, universities, and entrepreneurs to cocreate startups that solve compelling problems. He built his career in strategy consulting at Innosight, the firm founded by Clayton Christensen, in corporate venturing, and as an entrepreneur bringing new ideas to market.

Excerpted with permission from The Illusion of Innovation by Elliott Parker. Courtesy of Ideapress Publishing.