I’ve noticed the term “innovation” seems to mostly be used as an abstract way to describe new products or ideas that impress us. As an engineer, this abstraction can be frustrating because it does not provide much insight into how innovation might be achieved. The dictionary’s definition is not much help either: “introducing new ideas; original and creative in thinking.” Certainly we all introduce new ideas on a daily basis. But ideas are like opinions, and opinions are like… well… you know. So what makes some ideas innovative, and other ideas stink? This is my attempt to discern that; to construct a more concrete definition of “innovation”, as well as how it might be achieved.
“Innovation is the discovery of new truths that change current assumptions.”
The Scientific Method
I believe the most important invention, discovery, process, or idea ever created has been the scientific method. The scientific method is a set of instructions by which we can discover what is true. Although it has been critical to the advancement of mankind, it is beautifully simple: construct a hypothesis about what is true, list assumptions that led to the hypothesis, then test the hypothesis to discover if it is actually true. This is an iterative process that may take days, months, years, centuries, or even millennia but it has led to important discoveries, such as: Descartes realizing “I think therefore I am”, Newton discovering gravity, Herschel discovering infrared, and the development of every piece of technology we use today.
I used to believe the scientific method was reserved for science class – when it was time to blow shit up or freeze bananas. I have since learned that need not, nor ought to be the only environment for which the scientific method should be applied. For example, I believe the scientific method may also serve as the perfect basis by which innovation can be achieved. I believe this because of the definition of “innovation” I have come to develop – “the discovery of new truths that change current assumptions”. If this is an applicable definition, then the most critical activity of innovation is the search for [and discovery of] truth; and no other process is better suited to discover truth than the scientific method.
This definition is itself a hypothesis. Following are three examples of individuals whose innovations – I believe – support it.
Steve Blank and Eric Ries
Today’s entrepreneurs often adhere to Steve Blank and Eric Ries’ Lean Startup Principles to ensure they build companies that create products and services people actually want. Ries and Blank realized most startups went bankrupt because they produced offerings no one was willing to pay for. Lean Startup combats this by posing that before executing a business model, the assumptions of the model ought to be tested and validated by “getting out of the building” and speaking with potential customers and stakeholders. The consensus has been that their approach better ensures profitability and reduces the risk of failure.
“Lean Startup is a business adaptation of the scientific method.”
While Blank and Ries have developed sophisticated names for aspects of Lean Startup, such as: “business-hypothesis-driven experimentation”, “iterative product releases”, and “validated learning”, at its core, Lean Startup is a business adaptation of the scientific method. It is a method for developing hypotheses, noting assumptions that led to those hypotheses, and testing the hypotheses in order to iteratively discover what is actually true about the business model. Hypotheses are validated (proven true) or invalidated (proven untrue) and new hypotheses are developed as a result.
We have seen businesses of all sizes be innovative by changing our assumptions: Apple changed our assumptions about what a telephone can be used for, Google changed our assumptions about how to best find relevant web pages, Uber changed our assumptions about paying someone to drive us, AirBnb changed our assumptions about where we should stay on vacation, and so on.
The way by which we could reliably create new and innovative products like these used to be something of a mystery – perhaps even to the innovators themselves. Steve Blank and Eric Ries’ were able to solidify a process by which any person or company could develop new and innovative products; and the process they developed strongly correlates to the scientific method.
As a result of the financial crisis, asset management may not impress us as a particularly innovative industry – at least, not in a positive way. However, Ray Dalio – founding manager of the world’s largest hedge fund, Bridgewater Associates – is certainly an exception. Ray’s great innovation was his discovery of a portfolio that could survive [and thrive] in all types of “economic weather”.
I recently read MONEY: Master the Game by Tony Robbins in which Tony prompted Ray to disclose his “All Weather” approach to asset allocation. To summarize, Ray revealed he had discovered three important truths about financial management: (1) A portfolio that invests 70% of capital in stocks and 30% in bonds actually allocates over 95% of risk to stocks and less than 5% to bonds, (2) there are four fundamental economic seasons that depend on the following environments: higher than expected inflation, lower than expected inflation, higher than expected economic growth, and lower than expected economic growth, and (3) there exists at least one portfolio that is designed to perform well in each of the four economic seasons.
“Ray discovered new truths that changed assumptions about how assets ought to be allocated.”
In this context, it is evident to see the scientific method inherent in Ray’s success. Conventional wisdom suggested a 70/30 asset allocation was the best way to optimize a portfolio’s risk/reward characteristics. Ray challenged that assumption by calculating that stocks are three times riskier than bonds, and therefore such an allocation is far from diversified or balanced. Ray then hypothesized how the economic machine works in order to identify four fundamental economic seasons. Finally, he hypothesized and calculated a portfolio that would be well-suited to withstand every such environment; and he has validated his hypotheses over decades'-worth of successful tests.
Broadly speaking, I believe Ray discovered new truths that changed assumptions about how assets ought to be allocated.
I am currently reading Work Rules! by Laszlo Bock – Google’s SVP of People Operations. It is an inside look at how Google operates. Much like MONEY: Master the Game, it has blown me away.
Google used to be infamous for asking interview questions like: “How many golf balls can fit into a school bus?” or “why are manhole covers round?” Then Laszlo looked at the data and discovered interview questions like those were not actually indicative of an employee’s future performance. Instead they were bias toward a subset of the population that were good at solving brain teasers, and served more to pat the egos of interviewers than to identify and measure the type of intelligence Google valued. As a result of his analysis, Laszlo began to champion a hiring approach aimed at measuring attributes that did predict an employee’s future performance, which included: general cognitive ability, emergent leadership, cultural fit, and role-related expertise.
Similar to their evolution of hiring practices, Laszlo and his team used data to look at the effect and importance of management at Google. Early on in Google’s history, they hypothesized management may be unnecessary for engineering teams. They tested this by removing manager roles from those teams, only to realize it left executives exposed to resolving simple issues – such as time off requests – which greatly reduced efficiency and caused chaos. This led them to conclude management was necessary – even for engineers.
Additionally, Laszlo and his team proceeded to test whether the quality of management mattered to the overall performance of a given team. They found that it did. So Google reconstructed the way by which they measured a manager’s performance. The data they collected then allowed them to develop a training program for managers whereby managers who were strong in a given area could coach their peers in that area, and they could quantify the benefits in meaningful ways.
Though their processes may be labeled a number of things (namely data-driven whatever), my conclusion is that Laszlo and his team simply used the scientific method. Take the interview questions; they hypothesized brain teasers would be effective in identifying the best candidates because they assumed a correlation between ability to solve brain teasers and the type of intelligence Google valued. They also assumed a correlation between that type of intelligence and an employee’s future performance at Google. However, when they tested their hypotheses and assumptions, they invalidated them. But in the process, they discovered new truths that improved recruiting efforts as well as the culture at Google.
Much like Steve, Eric, and Ray, Laszlo did not develop his strategies out of isolated intellect. Laszlo took an iterative approach to improving his hypotheses by testing his assumptions about what was true. Ultimately, he discovered new truths that led him to change Google’s assumptions about everything from interview processes to management reviews.
I believe innovators are people who discover new truths that change current assumptions. This may come in the form of new product offerings, cultural restructurings, or even whole changes to a business model. Regardless of its form, I am convinced innovation must be based in truth. I refer to advancements not based in truth as pseudo-innovation. Pseudo-innovations are often temporary and result in reversion back to old assumptions and paradigms. We’ve seen how destructive this can be (e.g. LTCM, CDOs, Enron, etc.).
If innovation must be based in truth, then the processes we design to mediate innovation ought to be designed to discover truths about whatever it is we are interested in innovating. That is why I believe the scientific method serves as the ideal basis by which we can develop such processes. And I believe the examples outlined in this article support that assertion. But I may be wrong. These are simply my hypotheses.