Sensitivity analysis oversimplifies the real world

Dealing with uncertainty is not a new issue in business, but coupled with the COVID-19 crisis, it has intensified it threefold. A common approach to thinking about uncertainty in the realms of business is to perform a sensitivity analysis. This generally involves taking one business parameter at a time and flexing assumptions up and down around a base case to identify its impact. In some areas of advanced project and investment analysis, this approach to sensitivity analysis extended to Monte Carlo simulations where value drivers were varied and integrated randomly in financial modelling.

However, as COVID-19 has reinforced, sensitivity analysis tends to oversimply the real world where outcomes are more often the interplay between a range of risks. While foresight is unlikely to ever match the certainty of 20/20 hindsight, with the right business processes, there is no reason why events such as COVID-19 can’t be anticipated, thereby buying valuable time to make decisions that improve business resilience.

From linear planning to scenario planning

While it is very common for management to present linear financial forecasts, uncertainty determines that outcomes are rarely linear. Ideally, strategy and planning are iterative processes that absorb new internal and external knowledge to test key financial assumptions (e.g., demand, margins, market share) and key choices around strategic positioning (e.g., differentiation, technology, products, segments). Naturally, when in a crisis, this process occurs with haste in management teams undertaking ‘triage’ on their key strategies/programs/projects/initiatives to determine which remain robust (to continue), are no longer valid (discontinue), and are most shrouded in uncertainty and their relative sensitivity to future outcomes.

However, there should be no reason to wait for a crisis to undertake such activity. Increasingly, businesses are adopting an approach to scenario planning that helps stress test their business under a range of different ‘future world scenarios’. Importantly, unlike sensitivity analysis, scenario planning is seeking to understand better the interplay between different areas of risk and uncertainty. Experience suggests that success in scenario planning is getting the balance right between scenario complexity and business relevance. In seeking this balance, there are five guiding principles for effective use of scenario planning.

Five guiding principles for effective scenario planning

Breadth of scenarios – A sound guide for scenario planning is to work with more than 3 scenarios to protect against the normal human bias of considering them as low and high case sensitivities around a base case plan; a normal outcome of a sensitivity approach to thinking about uncertainty. Equally any more than 5 scenarios are likely to result in process complexity that won’t necessarily be re-paid with improved business insight. In most cases, four discrete scenarios appear to be a good balance.

Scenario construction – The foundation logic underpinning scenario construction is generally achieved by flexing assumptions along two axes. One axis is generally dedicated to flexing macro demand drivers (i.e. drivers largely outside of the business direct control – political, social, economic, environmental, regulatory); and ranging from low to high demand. The second axis generally focuses on factors that affect competition between firms and generally are more capable of being influenced by the actions and strategic position of the firm (i.e. business and operating model, technology, digital, distribution, etc). This axis scales between less and more disruption. While customising/personalising this approach to scenario construction to the relevant conditions of the industry and company is important, this approach generally delivers a framework for four discrete scenarios ranging from low demand and low disruption to high demand and high disruption and the variants in-between.

Identifying an overall strategic preference – Once plausible scenarios have been created and the impact on key business and strategic drivers has been assessed, it is helpful to frame an overall strategic preference for the firm. This involves positioning amongst a range of alternatives including; sustain and reinforce, restructure, create new business models, or create new businesses. The objective of defining the overall strategic positioning is to align all stakeholders around an overall positioning of the business and a view around its relative resilience under the current state.

Define actions – The next step is defining actions that are robust across a range of scenarios and where there is some alignment with the overall strategic preference. This may include a range of actions that are robust under all/most scenarios and are effectively ‘no regrets’ moves. Some actions may have a high differential impact under different scenarios and require ‘big call’ decisions (ideally seeking options to phase decision making around these big calls). And there may be other actions that ‘provide optionality’ that can be kept in a portfolio of options and adjusted over time based on trigger points.

Identify trigger points – The ideal outcome of scenario planning is not a definite and locked in workplan, but rather, an agile and flexible forecast, where a management team is positioned to be quick and insightful learners and adaptors and make better decisions as uncertainty evolves over time. Some actions only make sense under certain scenarios so being clear about ‘trigger points’ is helpful to drive progressive decision making.

Success is not knowing the future

The benefit of effective scenario planning is to enhance decision making in the face of uncertainty and thereby improving business robustness/resilience. Success is not measured by knowing the future but rather being the first (amongst peers) to identify the emerging pathway and act appropriately.