When we say that entrepreneurs and managers’ behaviors should incorporate aspects of the scientific method, we do not refer to the findings of science, but to a general method of thinking about and investigating problems. This scientific attitude in entrepreneurial - and more generally - managerial behaviors comprises three elements. The first is the hypothetical spirit, that is the feeling for tentativeness and caution, the respect for probable error. The second is experimentalism, that is the willingness to expose ideas to empirical testing, to procedures, to action. The third is criticism, that is the openness to collective discussion, confrontation and critique, to achieve a collective common understanding and set a guide to common action. The intuition that entrepreneurs should behave like scientists and that the entrepreneurial process should be grounded on the application of the scientific method has innervated entrepreneurship and management studies since their dawn. A classic reference is Frederick Winslow Taylor's ìScientific Managementì, but other streams of management research, including management science and operations research, industrial and system dynamics, and some of the work by Peter Drucker, have explicitly paralleled the work of scientists and the work of entrepreneurs and managers. On top of these antecedents, there are at least two other related sets of contributions that pertain to this analogy. The first derives from the application of statistics to managerial decisions and problem solving. In the 1920s, Walter Shewhart conducted some pioneering studies on statistical process control applied to production quality at the Bell telephone laboratories. These studies initiated a more general trend of applying data and statistical analysis in managerial decision-making. At Bell telephones, Shewhart met in the early 1930s William Deming, a young engineer then, and one of the most quoted management gurus later. And together they started to elaborate on the fact that statistical analyses should not be limited to support quality control, but rather analyze and understand the root cause of any defect, solving such problems effectively in order to allow the systematic improvement of organizational performance. These pioneering studies engendered in parallel with the development of multivariate statistical analysis and econometrics, a series of management systems, as total quality management and - later - Lean Thinking and Six Sigma, grounded on the application of the scientific methods to business decisions. These management systems have in common a decision-making discipline based on structured cycles of problem solving. Example of such cycles are the PDCA, Plan-Do-Check-Act cycle, at the basis of Lean Thinking, and the DMAIC cycle, Define-Measure-Analyze-Improve-Control, at the basis of Six Sigma. The most famous is the Deming cycle, also known as a PDCA, Plan-Do-Check-Act, or - more precisely - PDSA, Plan-Do-Study-Act. This cycle represents the logical steps that should be taken in order to solve problems, and therefore making decisions according to the scientific method. In the first phase, Plan, managers identify the problem, develop their own understanding of the situation, their theory, and elicit the hypotheses, their predictions. In the second phase, Do, managers test the hypotheses using experiments, direct observation, and data. In the third phase, Study, managers analyze and interpret the data, and compare this to their predictions. In the fourth phase, Act, managers reflect, evaluate and draw the practical implications of what they learned, making the appropriate decisions. More recently, innovation and entrepreneurship research has elaborated on this idea, alternatively emphasizing the diverse components of the above-described scientific attitude. For example, Teppo Felin and Todd Zenger conceive entrepreneurs as theory developers, engaged in deliberated problem-framing and solving, and suggest that any strategy is also a theory, and can't be a mere trial-and-error search process. When they try to innovate, entrepreneurs and managers can create value as they formulate, identify, and solve problems. The above-described experimental spirit is, instead, epitomized by the burgeoning entrepreneurship literature on customer development and the Lean Startup method. This method of creating and delivering innovation to the market was originally proposed by Eric Ries in his bestseller, ‘The Lean Startup’. It explicitly refers to lean thinking and the PDCA cycle, and suggests that entrepreneurs and managers develop ideas and innovate iterating an experimental process which comprises three major steps: build, measure, and learn. This iterative cycle suggests that any innovation decision should: first, formulate and test hypotheses through business experiments to validate the customer problem and the solution technology designed to solve it. Second: gather data on actionable metrics and run analyses to infer if the problem is real and the solution effective. Third: analyze, reflect, and interpret the results to learn, possibly abandoning the idea, sticking to it or pivoting to something different. Consider, for example, the case of a new product, like an application. The Lean Startup method starts with the question: Should this product be built? Does it address a real problem, creating value for the customer? In order to validate this problem, it is necessary to test - with experiments or data - a set of hypotheses. Once this is done, then solutions to the problem must be validated, testing different options of minimum viable products through A/B tests. Eventually, product market fit must be achieved, again through validation in order to make the idea viable and scalable on the market. All these activities take place using the above-described Build-Measure-Learn cycle, where the key factor is to shorten the total time through the learning loop. One famous case of application of the Lean Startup method is Dropbox. Through the Build-Measure-Learn cycle, the company started iterating their product much faster in order to test what customers really wanted. They did it early on and often. Using the Lean Startup method and principles, in just 15 months, Dropbox went from 100,000 registered users to over 4 million. The Lean Startup method has been applied also in large established enterprises like General Electric, which developed its own version FastWorks, for example, to develop and test new home appliances like refrigerators. The PDSA cycle and the Lean Startup method are examples of entrepreneurial and managerial approaches that have incorporated aspects of the scientific attitude we described at the beginning of this module. We believe that adopting a scientific approach to the innovation or entrepreneurial process comprises a general method of thinking about and investigating problems which has four components: theory, hypotheses, tests, evaluation. Entrepreneurs and managers engaged in product or process innovation, will be better off, if they articulate a vision or a theory, that typically leads it to the definition of a business model grounded on correctly framing the problem the startup or the R&D function wishes to solve. As we will see in the next modules, tools like the Business Model Canvas can be used to this aim, in order to make sure that the idea being pursued is grounded on reality and logically sound. Theories are built on deep, rigorous thinking and on effective reasoning. In practice, good theories - once told - sound like persuasive, compelling stories. A theory comprises a set of hypotheses, that is a set of ‘if… then’ statements that constitute the basis of any tests, and consequently of any discovery and learning process. We will see that these hypotheses should be formulated explicitly and should be testable and falsifiable. Hypotheses need to be tested empirically, collecting data, and possibly running experiments to understand whether the hypotheses are true or false. These tests need to be rigorous, data-driven, and should be characterized by the appropriate choice of counterfactuals, data gathering and analysis methods, decision rules, as well as by valid and reliable metrics. Once the outcomes of the analysis of data and experiments are available, they should be discussed and scrutinized collectively, in order to make sense of them, learn, and make follow-up decisions. Individual and collective judgment, as well as critical appraisal of evidence, are essential when using data to inform a specific decision in a particular setting. Summarizing, managers and entrepreneurs following a scientific approach to innovation decisions, have a behavioral discipline that allows them to reduce uncertainty, mitigate the risk of incurring in false positives and negatives and make better stay, abandon, or change decisions.
