This week's guest is an innovation guru from one of the world's leading business schools. Stefan Thomke is a William Barclay Harding Professor of Business Administration Chair at Harvard Business School and a widely published author around innovation processes. We will discuss Stefan's new book, “Experimentation Works: The Surprising Power of Business Experiments,” and how experimentation functions as the engine of innovation.

Experimentation

Stefan says he got involved in experimentation while working as an engineering intern. He got tasked with optimizing a chip manufacturing process and was lost on how to do it. Someone mentioned looking into experimentation, and he decided to research and study it. He was able to solve the problem, prompting his realization that experimentation is the engine of innovation.

For many innovators I know, innovation is viewed as stumbling in the dark, hoping you have that “eureka” moment, verses having a methodology. Is that the case for innovators you interact with? Stefan says part of the problem is how we use the word experimentation itself. When people say they experiment, often they mean they are just trying something. They did something, and it didn't work; therefore, it must be an experiment. I am talking about disciplined experimentation using principles employed in a scientific method. It is important to note because you can't learn much from an experiment without implementing a process into it.

Failure and Incrementalism

In your book, “Experimentation Works: The Surprising Power of Business Experiments,” you call out the difference between an experiment and a mistake. In your opinion, what is the difference? Stefan says a mistake is something that you learn nothing from. The difference is that a failure has a learning objective, where a mistake does not. Failure is okay when you are learning from it.

Does experimentation get correlated to incrementalism? Stefan says that most innovation in the world is incremental. Most of the significant performance changes we see are the result of the cumulative impact of small changes. Microsoft changed the way they displayed their headlines and increased revenue by 100 million dollars a year. I call it high-velocity incrementalism, which means that you need to run fast but also go for scale. You need to be able to link cause and effect as a business. You want to have a high level of confidence that action A will produce outcome B.

Experimentation Culture

What are some attributes that are needed for an experimentation culture in business? Stefan says that companies often assume if they put the right tools into place, the experimentation will just happen, which is not the case. There are a few elements needed to be successful. Firstly, a company needs to have curious people who value surprises. Secondly, they need to insist that data trumps opinions. Often companies will only accept results that confirm their biases and challenge results that go against our assumptions. Thirdly, we need to empower people to perform experiments. If people have to run it up the chain for every experiment, you're not going to get the scale that you need run on. Next, you need to ethnically sensitive because we will all react differently to experiments. Lastly, you need to embrace different leadership mo