Dealing with uncertainty in your forecast
Posted by John Stretch
Forecasting is one of an FP&A department’s most important activities. We believe that we can produce forecasts which are reliable, and that our forecasting will improve through repeated analysis, learning, judgement, detailed supporting data and good systems.
Every prediction we make contains uncertainty. A forecast of sales to a big customer over the next three months may be more - or less - reliable than a forecast of the oil price over the same period. As professionals, we have to understand and communicate the risk and reliability in each forecast we produce. A single forecast of sales, costs and profits, although based on choices and probabilities, is only one view of a broad range of possible futures. And yet many managers still ask for a single point forecast which they hope will turn out to be reliable.
How should an FP&A professional deal with and communicate uncertainty in a forecast?
The estimated accuracy of the forecast
We can usually attach greater confidence to a forecast for the month or quarter immediately ahead, than for the month or quarter 12 months ahead. In most cases, to provide a single point figure of the reliability of the forecast (“this forecast is 90% reliable”, for example), will be misleading to the user. Most projects can be broken down into phases and the accuracy and reliability of the forecast should be considered for each phase of the project. Similarly, sales or profit forecasts can be analysed by time period or season.
The forecasting process can be greatly assisted by collecting and analysing historical data of the key forecast drivers, for example units sold, customer counts, capacity utilisation, and number of employees. In addition, we monitor the trends in monetary items such as raw material prices and exchange rates. Fluctuations in the data help us form a conclusion about trends into the future. The assumptions made about these trends should be communicated to the user.
Some trends such as technological change and new product sales in some industries, may be exponential rather than linear and thus more difficult to forecast.
The known risks
Every forecast considers and includes known risks such as strikes, potential bad debts or loss of a major customer. We forecast these risks by forming a view on the probability of occurrence, and the past and likely future frequency.
The known risks and how they have been dealt with in the forecast must be made explicit in the commentary.
The unknown risks
Conversely the forecast excludes risks from situations not thought of, or not considered possible, or are so unlikely that they have not been considered worth forecasting. Mathematician Nassim Taleb uses the term “black swans” to describe these unplanned events with severe impact that, with hindsight, were predictable or explainable.
In some industries, and in different parts of the world, Black Swan events repeat at regular intervals (crop failures, epidemics, tsunamis, stock exchange crashes). The regularity of these events suggests that black swans may be giving way to grey swans.
Possible black and grey swans should be presented as scenarios. Scenario planning combines assumptions, trends, knowns and unknowns, to speculate on alternative views of the future world in which the organisation will operate.
In an uncertain and fast changing world, line managers need to be made aware of the uncertainties and risk inherent in the financial forecasts provided to them. FP&A professionals seek to continually develop their ability to identify, understand and communicate risk in their client businesses. When risk is understood, it can be categorised, mitigated, managed, hedged or even avoided.