Uncertainty Is All Around
FP&A teams do not have a crystal ball… Yet, they have to answer management’s questions about the future every day.
What sales will the company generate next year? What revenue will the launch of a new product bring us? How will the competitor’s move impact our market share? These are only few examples of the uncertainty we face.
Being asked for advice about uncertain future outcomes is indeed flattering, but at the same time it adds much pressure and responsibility. To feel more comfortable and confident with estimates, FP&A teams should embrace uncertainty and incorporate it in their core activities.
First, it is crucial to understand that uncertainty is manifested in many forms, thus should be approached differently. Various types of uncertainty can be well illustrated by the so-called Rumsfeld matrix widely used in risk analysis and risk management.
A former US Secretary of Defense, Donald Rumsfeld mentioned this matrix to answer a question of one of the journalists: “… as we know, there are known knowns; there are things we know we know. We also know there are known unknowns… we know there are some things we do not know. But there are also unknown unknowns—the ones we don't know we don't know… it is the latter category that tends to be the difficult ones.”
The Rumsfeld matrix has two dimensions forming four quadrants of risks: known known, known unknowns, unknown knowns and unknown unknowns.
Depending on the purpose and the field of application these two dimensions of the matrix can vary. From the FP&A’s standpoint the level of awareness and the level of accuracy can be used as these dimensions and then incorporated into forecasting and analysis. In this case the first known/unknown stands for awareness, the second one - for the ability to accurately assess the level of risk.
Known knowns
These are risks we are aware of, and they can be properly measured. Risk identification and assessment are based on knowledge, thus, we are basically operating facts. Known knowns include most of the situations we face under usual or normal conditions.
This is the easiest category of risks both in terms of analytics and management. Regular update of information to check those facts against them ensures accurate forecasts based on the one-point estimate or a range with a high level of confidence. As the source of uncertainty is well-understood, mathematical models contribute to minimization of the probability of occurrence and/or the impact of these risks.
Known unknowns
These are risks we are aware of, but they can’t be properly measured. For instance, either the probability of the risk event can’t be assessed, or the impact can’t be quantified, or both.
Unlike the first category, known unknows include most of the situations we face under unusual or special conditions. Here we deal with hypotheses and assumptions.
Model parametrization plays an important role for this type of risks aiming at better understanding of the risk or improvement of the prediction capacity. Forecast is represented by the range of outcomes.
Unknown knowns
These are risks we intentionally ignore and don’t want to acknowledge. In this quadrant we move to the area of intuition, irrationality and judgmental biases. In the well-built risk-management system this type of risks should not exist.
FP&A teams should be objective in their judgments, avoid wishful thinking and develop non-biased approach to analysis and forecasting.
Unknown unknowns
These are risks we are not aware of, and they can’t be measured. Such risk events are associated with entirely unpredictable uncertainty and are often called “black swans”. Recent pandemic is an example of this type of risks.
Unknown unknowns are unknowable risks or risks that can’t be identified in advance. However, it doesn’t mean they should be ignored. Being low-likelihood high-consequence risks, unknown unknowns are the most dangerous as they threaten the survival of the company.
Though black swans are our blind spots which will never be completely eliminated, there are some techniques that may be helpful in detecting the gaps and having a plan beforehand.
Test implicit assumptions
It sometimes happens that there are assumptions in the models that are considered to be true without saying and no one even thinks of questioning them. They are not discussed, reviewed or tested because that’s “what everyone is doing” or that’s the “the way things have always been done”.
This is a common situation in the companies hiring professionals from the same industry/sector who bring their experience from one company to another and reuse it for years without questioning.
Such implicit assumptions seen as axioms can easily become the source of a blind spot. To overcome it we should avoid the concept “that’s the way it has always been done”, regularly test all the assumptions or/and ask unlikely sources who may see the problem from a different angle and bring a new perspective to the analytical framework.
Premortem analysis
Projects and strategies fail, and when it happens it is rather common to conduct postmortem analysis to understand the underlying reasons for that failure.
Unlike postmortem, premortem is an opposite method which starts with the assumption that the project has failed, or the core value proposition has been considerably damaged or destroyed. The team then works backwards to generate possible reasons for this failure, including those with low frequency of occurrence.
In its underlying backwards way of thinking premortem analysis has some touchpoints with discovery-driven planning. These two techniques are complimentary: while the former aims at figuring out the reasons for failure, the latter is focused on the assumptions that must prove true to achieve success. As result, their combo will help not only identify and understand potential risks but discover creative and unexpected insights which might have been otherwise overlooked.
Stress testing or plan for the worst
Whether of political, economic, societal, technological or environmental origin, overall, black swans may have rather similar severe impact on the firm’s activities and performance.
In this context the ability to predict a specific event is of much less importance than the ability to withstand the consequences of such events. The assessment of the latter can be done by conducting stress-tests and implementing further contingency plans.
The first step is to understand and map the shape of the company which is influenced by numerous factors, including its geographical location and key markets, products and services, supply chain characteristics, sales channels, partners and customers, industry structure and competition to name the most important of them.
Analysts should look for concentrations of revenue and high dependencies in operations in all this variety of factors to identify vulnerabilities of the firm. As we are discussing risks with extremely low likelihood, it is essential not to stop at typical events and first-order relationships while creating a list of potential threats. These weaknesses pushed to their limit values or modeled in the extreme conditions are used to generate stress scenarios.
This exercise usually reveals higher risk concentrations and greater impacts than it was acknowledged before, as well as the weaknesses of the company’s shape. The insights of the analysis serve as a base for generating mitigation plans to address discovered risks.
From FP&A’s standpoint a stress test is a variation of what-if analysis or scenario planning depending on the inputs. If there is no stress-testing in the company, FP&A should initiate this type of analysis to ensure that the firm is ready to face high-magnitude, low-frequency events.
No FP&A team can predict what will happen in the future, as no company can be completely prepared for every possible risk. However, understanding and incorporating various types of uncertainty in FP&A activities ensures proactive position of the function and contributes to the efficiency of risk management in the organization.