Don’t Waste Time Testing Business Theories. Use Them Instead

futurelab default header

Strategists and management consultants love theories. Executives like to consume them, as it allows them to make sense of things. People are now criticizing how Clayton Christensen’s theories about “Disruptions” are now outdated and were created based on incomplete data.

For a generation of CEOs and executives, Christensen’s The Innovator’s Dilemma was a must-read, and helped them think about industry disruptions. Lately, however, journalists and academics have questioned the accuracy of Christensen’s industry analyses and challenged his broad generalizations. But I’m not here to defend whether his theories were right or wrong, or comment on the accuracy of his data. There is little to be gained from this, other than a philosophical debate. His book and seminar were successful and helped popularized the term “disruption.” The criticism is not exactly fair either, as he is an academic and academics create theories.

And what is a theory, anyway? A theory presents a systematic way of understanding events, behaviors, and situations. The notion of generality, or broad application, is important. Thus, theories are by their nature abstract and broad. Theories vary in the extent to which they have been conceptually developed and empirically tested; however, “testability” is a feature of a theory. Most business theories (like social sciences) are better understood as models that work in a limited range of settings, rather than laws of science, which hold and apply universally.

Today, almost every theory in management, organization behavior, strategy, economics, and marketing can be proved wrong and mostly irrelevant. It is easy to critique the academics that created them to advance management education. They were useful at a particular time, and then they expired, just like any business model. The world is changing fast. We are now in an era of hyper-competition, where technologies such as computing, networks, sensors, artificial intelligence, and robotics are advancing exponentially. Converging, disrupting, and innovating are now the core of every business. The MBA should be rebranded as a Master of Innovation and Disruption (MID). And even MBA schools are struggling to teach these new skills.

It takes a long time to develop theories and then present them as business school case studies. By then, the world has changed. Business lessons cannot all be taught in an amphitheater anymore; they happen in the real world. Theories are being prototyped as we speak. And in the world of startups, perhaps the only theory that applies is “chaos” theory. In chaos systems, the biggest challenge is to determine what is random or chaotic. One of the hardest things about being in a startup in this state of chaos is actually figuring out what leads to randomness, or what leads to a signal to act upon. There is no theory, just instincts to follow—and a bit of foolishness.

Christensen’s disruption theory is not wrong, but it’s also not comprehensive enough. Competition doesn’t just come from the lower end of a market; it comes from everywhere as industry boundaries are blurred. For the taxi industry, Uber came out of nowhere. For the hotel industry, Airbnb came out of nowhere. Tesla came out of nowhere too. Snapchat and Spotify came out from nowhere. And the same goes for the 500+ fintech startups that all aim to disrupt banking.

A recent study published in the MIT Sloan Management Review suggests that the vast majority of cases that Christensen cited in support of his theory do not actually fit the model, and the theory can only be narrowly applied. I am not defending Christensen, but every management author’s job is to find stories that fit his or her theories and provide context for illustration. Some academics have nothing to do, and would rather spend two years asking people about these cases. Instead, they should be studying our fast changing world in order to come up with newer and better theories, which at some point will also be proved wrong. No theory is meant to be predictive, so don’t waste time testing them. Use them before they expire. Prototype them and iterate over them.

Image via flickr

Original Post: