Better Decision-making Through Iterative Experimentation

Complexity, uncertainty and unpredictable customer needs are characteristics of current and future business environments. To survive in this complex environment and make more enlightened decisions, organisations need to learn faster than their competitors. Concurrently, wise decision-making requires ability for holistic thinking and sensitivity for interdependencies throughout the organisation and fields of business. Organisations’ increasing need to adapt to changes and to make good, fast decisions is no longer supported by the conventional ways of development. Therefore, the need for non-predictive approaches that support organisational learning and decision-making exists.   

In more conventional, planning-based development, specs of the desired outcome (for instance a product, a service or a way of working) are settled after careful planning before any actual work is done or any of the assumptions are validated. In industries and fields of business where the characteristics and features of the market and products are well known and easily measurable, planning-based development works well. For instance, in the car industry, the phases of development and the details of the manufacturing process are known in advance.

However, when aiming for strategic innovations and new-value creation, we are dealing with complex systems, uncertainties and unpredictable outcomes that cannot be assessed through planning only. Thus, it is imperative to test the idea early, instead of waiting until after the implementation phase is over.

Final decisions about the detailed specs are delayed as long as reasonable

Experimentation-driven development refers to an iterative trial-and-error process in which initial ideas are tested and improved through several small experiments. Each experiment provides new knowledge about a problem, a product or a service through the feedback gained. The crucial part of an experimentation process is to reflect and to learn from this gained knowledge in order to improve the initial idea. Failing early saves resources and iterative experimentation validates the end result. Final decisions about the detailed specs are delayed as long as reasonable, i.e. until there’s enough empirical data from real end users in real environments to support them.

Experimentation on an individual level is hard. The associated uncertainty and risk-taking combined with the very common fear of failure effectively discourage most people from adapting the experimentation mindset into their daily work. Thus, it is up to the organization to encourage employees, limit the risks and tolerate failure, in addition to dispelling the illusion of clarity and certainty brought by the stage-gate development model. Indeed, small experiments are a fast way of acquiring accurate information piece by piece as opposed to the unreliable information produced by the planning stage of the stage-gate model, and as such are a better way of handling uncertainty. Even failed experiments provide valuable information and therefore should be embraced as much as the successful ones.

Experimentation is one way to approach organisational development and learning, and it also supports clever decision-making by providing new knowledge and information throughout the process. The experimentation mindset requires a new, positive attitude towards failure, which can and should be practiced; an initial push from the organisation’s side is often helpful. In the long run, the organisation should provide encouragement, support and resources for experimentation in order for it to thrive.

An experimentation-driven approach is fueled by uncertainty and failure; it turns them into opportunities for growth and leads to better decision-making.

​