Introducing Our New Scenario-Building Method Based on Principal Component Analysis (PCA)
To help organisations successfully navigate an uncertain future, Futures Platform has developed a cutting-edge scenario-building method that utilises the scarce knowledge we have about the future to its fullest potential.
FUTURE PROOF – BLOG BY FUTURES PLATFORM
With so much uncertainty in today's operational environments, organisations need more than traditional strategic planning to succeed.
Scenario planning is emerging as a critical tool for building resilient strategies that can withstand a range of possible future scenarios.
To help organisations navigate an uncertain future, we've developed a cutting-edge scenario-building method that utilises the scarce knowledge we have about the future to its fullest potential.
Essentially, a scenario is just one alternative future development out of countless possibilities. There are various methods for creating scenarios, depending on how many future uncertainties one wants to explore. A simple forecast explores the future direction of a single uncertainty, while the two-axes scenario method explores two, and morphological analysis examines several.
At Futures Platform, we used to rely primarily on the Futures Table scenario-building method, which allowed us to factor in multiple uncertainties in our scenario-building process. But we found this method too demanding for some of our more complex cases. As we saw the imminent need for a more advanced method, we began developing a new method that is based on the applied statistical method for analysing multidimensional data, called Principal Component Analysis (PCA).
INTRODUCING THE PCA METHOD
Principal Component Analysis (PCA) is a novel explorative-inductive scenario-building method developed by Futures Platform. It is designed to build a set of scenarios that maximises three criteria: diversity, plausibility and coherence.
The PCA scenario-building method allows us to explore multiple uncertainties simultaneously and generate manageable scenario settings out of the millions of options. It also eliminates all randomness associated with future development paths of uncertainties. Compared to deductive methods, PCA provides more comprehensive coverage of the spectrum of possibilities with a minimal number of scenarios.
The analysis method derives from an applied statistical method called Principal Component Analysis, which was created by Pearson in 1901. Our method shares the same name, but instead of working with actual datasets as in the original method, Futures Platform's PCA method works with what we call uncertainties and the correlations between them. In other words, the tool is designed to utilise the scarce knowledge we have about the future to the fullest.
Below, we present our process of building scenarios with the PCA method in detail.
“Compared to deductive methods, PCA provides more comprehensive coverage of the spectrum of possibilities with a minimal number of scenarios. ”
1. Defining the uncertainties
In scenario building, the key assumption is that the future state of any phenomenon depends on the development path of its key driving forces and how they interact. For example, if we want to explore scenarios on the future of the circular economy, we first need to understand the key environmental, political, economic and societal drivers that are pushing this development forward, as well as potential anti-drivers that may hinder it. We call all these factors uncertainties, referring to the uncertainty of their future impact, direction, or pace.
Uncertainties form the backbone of our scenario-building process with the PCA method, so it's important to choose them right. Rather than wading through generalities, we focus on the 6 to 10 most impactful uncertainties that can truly shift the tides on a topic. In the case of the circular economy, some key change drivers are the scarcity of natural resources, government regulations, and the increasing awareness of environmental issues among consumers and businesses.
2. Identifying the correlations between uncertainties
Change drivers behind any phenomena are complex and often interconnected. Hence, in the second stage of our scenario building with the PCA method, we assess the correlations between the identified uncertainties.
We assume each uncertainty pair is correlated with a strength ranging from -1 to 1, depending on whether the uncertainties tend to move in parallel or opposite directions. -1 represents a perfect negative relationship, and 1 represents a perfect positive relationship. However, these extreme values are rare in the real world, and most correlations fall somewhere in between.
Going back to our example of circular economy, the drivers of resource scarcity, government regulations, and increasing environmental awareness are all positively correlated: As natural resources become scarcer, governments are likely to introduce tighter regulations to encourage recycling and reuse. Consumers and businesses will also become more aware of the imminent threat of resource depletion, which will drive them to contribute towards a functioning circular economy.
To assess these correlations, we list all the uncertainty pairs and fill them into a table. Sometimes we can estimate the correlations directly from data, which leads to increased accuracy. In other cases, we rely on qualitative assessments.
Excerpt from an example correlation table
Assessing correlations helps us understand how uncertainties might develop in the future and build more plausible scenarios. However, it's important to note that a strong correlation doesn't tell us whether one uncertainty causes another or whether they have any causal relationship at all. It only tells us that there's a statistical dependency between the two.
3. Computation of principal components
Once we have the correlation table, the next step is to select a subset of sub-scenarios that describe different possible future states in the most plausible and diverse manner, given the correlation structure between uncertainties.
To create plausible and diverse sub-scenarios, we use Principal Component Analysis (PCA) to identify orthogonal directions, called principal components, along which the data has the largest variance. We choose data points along these first two or three directions to maximise plausibility and diversity. Our statistical software ensures that each chosen scenario is roughly equally probable based on the variance of each principal component.
Our computation produces a numerical value for each uncertainty in each scenario, ranging from minimum to maximum. By combining these values, we get an interpretable state of the future.
A two-dimensional example of four scenarios picked by PCA
4. Building scenario narratives and future development paths
The next and final step in the scenario-building process is where our team of foresight analysts takes the numerical values generated by the PCA software and transforms them into coherent storylines.
Spiderweb visualisation of one scenario (left) and four scenarios (right)
For example, how would the circular economy develop in a setting where resources become alarmingly scarce, yet regulations are either too slow in development or inconsistent across the globe? Based on these pointers, we describe how these future states would look and their key implications for future governments, societies, economies, and businesses.
Finally, we map out the relevant events and their timing into a development path through backcasting, where we work backwards from the envisioned future state to the present to identify the series of events that need to occur for this future state to become a reality. This gives us a list of signs and indicators to look out for when monitoring the future development of the scenario.
FUTURE-PROOF YOUR STRATEGIES WITH SCENARIO PLANNING
In short, PCA is an innovative and effective way to build robust scenarios that help organisations minimise risk and maximise opportunity in their strategic planning. Its ability to incorporate a large number of uncertainties while producing diverse and coherent scenarios makes it a powerful tool for any strategic planning process.
Futures Platform's trend and scenario database hosts 1000+ scenario and trend analyses written by our team of professional futurists. By analysing future scenarios across all industries and geographies, we empower our clients to make informed decisions and thrive amidst uncertainty.
To stay ahead, organisations must widen their lens, exploring not just their own industry but what’s happening on the edges.