Although it’s highly unlikely that Nobel Prize-winning novelist Anatole France was referring to FP&A when he said, 

“That man is prudent who neither hopes nor fears anything from the uncertain events of the future”,

he captured a truth known by FP&A professionals. With the rapid pace of change and complexity of the business and macroeconomic environment, it can be incredibly challenging for us to prepare accurate financial forecasts in the medium term. Making transparent forecasts and providing clear, actionable insights to business leaders really matters to ensure that our organisations can adapt and thrive regardless of the future that unfolds. What steps can FP&A teams take to ensure that their forecasts and models help leaders navigate uncertainty and enhance their understanding of the business environment?

Four Forecasting Hacks to Enhance Business Value

The good news is that some of the most effective changes to forecasts and their presentation to the business are inexpensive and even potentially free to implement. The downside, however, is that these changes are based on behavioural/culture change, which can be challenging to embed properly. Some stakeholder and Change Management efforts are likely to be necessary to ensure that the changes stick within the organisation.

Before diving into forecasting hacks, I’d love to say a quick word on a fundamental FP&A forecasting practice that’s a prerequisite to optimising the value from these hacks: driver-based modelling. These dynamic models use the key performance drivers identified, like pricing and production efficiency for FMCG businesses and the mathematical relationships between them, to adjust model outputs as input values change, enabling more responsive and accurate planning. However, according to the 2024 FP&A Trends Survey, only 37% of survey respondents adopted driver-based models in their planning despite a clear linkage between the implementation of driver-based modelling, enhanced forecast quality and elevated data-driven decision making. This study also identified that driver-based modelling allows FP&A teams to allocate more time to higher-value activities and improve overall team performance.

1) Be transparent about future uncertainty by adopting probabilistic forecasting

Many FP&A teams use deterministic models. They present one possible outcome after considering available information, potentially with a best-case and worst-case variant of the central case. The problem with these is that they provide stakeholders with greater confidence and certainty about expected future performance than is often achievable. With many different, interconnected factors that can impact an organisation’s performance, FP&A teams need to give business leaders a more nuanced understanding of potential outcomes than is provided by presenting a central case prediction.

Probably the single, most effective change that can be made is to incorporate probability weighting into FP&A forecasts to better capture the uncertainty and variability inherent in predicting future events. By doing so, FP&A teams can communicate their degree of confidence in the forecast and provide a clearer understanding of the risks and uncertainties involved. In turn, this enables decision-makers to consider and prepare for the range of possible future outcomes and have more focused discussions about risk mitigation and course correction. For instance, what actions are required to deliver a higher probability of a particular desired outcome?

An industrial chemical manufacturer has taken this step to better account for the uncertainties and risks associated with their business environment. They wanted to have a more robust approach to handling volatility in commodity prices and other performance drivers. They went through the following three main steps:

1. They compared historic forecasts with actual performance to identify the factors contributing to risk and uncertainty in their forecasts;
2. Based on the factors highlighted, they defined the key model inputs that required probabilistic expression and quantified probabilities based on historical data and
3. Incorporated the probabilities into the forecasting model and ran simulations to produce a range of possible forecast outcomes for presentation to decision-makers, including an explanation of the core risks that drive the forecasting range.

The adoption of probabilistic forecasting directly enhanced risk management at the company. Decision-makers had a clearer understanding of the risks and uncertainties within the forecasts and were able to make more informed, data-led decisions.

2) Don’t just focus on the worst-case scenario

Plans and forecasts need to be changed as the future plays out. Due to the inherent uncertainty of the future, a forecast cannot be 100% probable. Therefore, there will always be a chance that a forecast proves incorrect; a forecast that is 90% likely will still be wrong 10%.

Due to this uncertainty, organisations need to be able to quickly run and consider scenarios to remain agile. However, the 2024 FP&A Trends Survey findings regarding Scenario Management were mixed. Although nearly a quarter of respondents (22%) reported the ability to run scenarios in real time or within a day, a similar proportion could not run scenarios at all (21%) or saw no value in Scenario Planning (4%).

Which scenario(s) should be considered to maximise the value derived by decision-makers and maintain business agility? Frequently, FP&A teams will present a worst-case scenario to decision-makers, but it is unlikely that the worst-case scenario will play out. More likely, it is a “quite bad” scenario where the underperformance of a few key drivers results in poorer-than-expected outcomes. To stay agile and avoid surprises, it's more valuable for a business to forecast a likely "quite bad" case than a worst-case scenario. This allows decision-makers to understand the most probable forecast underperformance and the drivers behind it so that they can discuss mitigations ahead of time. Understanding the impact of a black swan worst-case scenario remains valuable, but given its low probability, it shouldn’t be a major focus for business leaders.

A good example is an oil and gas supermajor that relied on this technique to navigate the collapse in oil prices during the COVID-19 pandemic. They built a range of scenarios based on plausible macroeconomic factors, including a “likely worse case,” a variant from their base case, as well as a worst case. The likely worse case highlighted the top risks for management attention and maintained their ability to react quickly to a rapidly changing environment by providing early visibility and discussion of potential challenges.

3) Be wary of assumptions based on averages

We all know that averages can provide a general sense of a dataset’s overall characteristics. However, they can also mask significant variability within that dataset. When it comes to FP&A models, sometimes model assumptions use average values even when they may not be a good representation of the expected value, resulting in forecast errors.

A universally applicable example is model assumptions on inflation. Very often, inflation will be applied to a model at a single rate. As there can be significant variability in the inflation rates across the basket of goods that constitute the Consumer Prices Index, this approach may not be suitable, especially for major cost drivers. For example, businesses in the aviation sector were exposed to energy price inflation following Russia’s invasion of Ukraine.

Taking a more granular approach to such model assumptions will help us improve forecast accuracy. Model changes should be limited to major drivers only to avoid disproportionate impacts on forecast accuracy from the use of average values without introducing unnecessary complexity and detail.

4) Look externally to align model assumptions with macroeconomic conditions

“No plan of operations reaches with any certainty beyond the first encounter with the enemy's main force.”

Helmuth von Moltke, field marshal and Chief of the Prussian General Staff (1857-88)

Increasing our sense of certainty helps reduce our sense of threat and improve our ability to focus. Nevertheless, this instinct presents a trap when it comes to modelling in FP&A. Some of us might have come across incredibly impressive forecast models that incorporate inputs from multiple, diverse sources and accurately capture intricate relationships between drivers with internal focus because there’s greater certainty about the internal environment than the external one. Typically, these models function well whilst the macroeconomic picture remains stable but become less accurate as the external environment changes.

While it's tough to factor in the macroeconomic environment, it's always worth the effort in forecasting. Whether modelled or not, external factors will impact achievable growth, as well as the cost base and other factors, so model assumptions should be evaluated in that context. Otherwise, the forecast may not survive its first encounter with the real world.

Outsmarting Uncertainty

FP&A teams have a tough job, as accurate forecasting is a major challenge! However, some reasonably uncomplicated and inexpensive tweaks can be made to our forecasting approach to enhance the insights that we deliver to the business. By embracing these hacks, we can transform our forecasts from a daunting task into a strategic advantage.