Scenario Planning vs. Time-series Forecasting
When discussing ways to prepare a business or organization for future events, there are many types of projection methods that can be employed. Scenario planning and time-series forecasting are two approaches that have their respective benefits and drawbacks. In assessing the differences and similarities between these two approaches, it is necessary to examine them both to understand why all projections are not made with the same intention.
Looking at the traditional approach to forecasting, it is important to break down the concepts that frame how projections are made. A time series is a sequence of observations taken sequentially in time (Brownlee, 2019). In this format, the data that is collected is arranged in a manner that creates an order within the dimension of time from which projections about specific variables can be drawn (Brownlee, 2019). It has been asserted that the use of the term “scenario” in the context of time-series forecasting is a misapplication of the terminology in that the projections are actually describing variables and not scenarios (Wade, 2012). This distinction is an important aspect that will separate traditional forecasting from scenario-planning.
On the other hand, scenario planning anticipates that there is no way to predict a single possible outcome and prefers to depend upon projecting multiple future scenarios as an array (Wade, 2012). Scenario planning involves looking at the strengths, weaknesses, opportunities, and threats (SWOT) facing an organization to inform the array of most likely future scenarios (Wade, 2012). The inclusion of uncertainty in the process makes the scenario-planning approach much more agile in response to real-world and real-time change when compared to traditional forecasting (Wade, 2012).
While both of these approaches have their benefits, in the context of modern organizations the appeal to uncertainty that is inherent in scenario-planning makes it more useful for projecting than traditional time-series forecasting (Wade, 2012). Time-series forecasting will likely be used to help give machine learning algorithms the front end appearance of making scenario-planning suggestions in the form of UX/UI like Siri or Alexa (Brownlee, 2019).
This is where the projection methods that are no longer useful for business can be useful for machine learning algorithms that can then reproduce the most novel and useful automated projection methods in the form of a front end UX/UI. As time-series forecasting relies heavily on mathematical concepts and frameworks, the application of these ideas into machine learning will be a natural transition that will synergize the experience humans have making projections with the ability of computers to make accurate data extrapolations quickly.
References:
Amer, M., Daim, T. U., & Jetter, A. (2013). A review of scenario planning. Futures, 46, 23-40.
Brownlee, J. (2019, August 21). What Is Time Series Forecasting? Retrieved from https://machinelearningmastery.com/time-series-forecasting/.
Wade, W. (2012). Scenario planning: A field guide to the future. John Wiley & Sons.