THINK about some of the predictions you have made recently which haven’t quite gone to plan. It could be anything from the value of a stock to an optimistic weather forecast. It could be the World Cup game where Brazil fans watched in horror as they were knocked out 7-1 by Germany.
If there’s one thing about predicting the future that always comes true, it’s that everyone, from time to time, gets things wrong. The same applies to financial forecasts. A whole host of unpredictable factors can affect the accuracy of the financial planning and analysis (FP&A) team and alter the whole strategy or direction of the company.
But what separates best-in-class companies from those that struggle with accuracy is how they root out (and learn from) forecasting errors. While FP&A teams can control and eliminate factors that cause forecasting discrepancies ahead of time, others are simply uncontrollable. In those cases, teams should concentrate on learning from their mistakes.
The key is identifying why forecasts go wrong and separate those causes into two camps. One is controllable drivers, such as process errors, miscalculations or managerial biases that make forecasts too conservative or optimistic. The other is uncontrollable factors, including everything from natural disasters to interest rate rises to a supplier suddenly raising his prices. Finance teams that can separate these drivers are better able to increase accuracy over time and drive accountability for the process.
So how can you identify these variables and improve your chances of accuracy?
The best (and simplest) way is often to ask the front line business leaders themselves. One FP&A team in CEB’s network interviewed c-suite executives to ask about the operational drivers that affect business performance. After identifying the most common causes of variance, the function published its findings in a summary designed to help inform and improve future iterations. It’s a bit like isolating the biggest variables that affect sports results – whether penalties, injuries or red cards – and learning from experience about which of these is most likely to happen and should be accounted for.
Through our research we have also seen that personal bias can play a huge role. Poorly structured incentive schemes often lead managers to game forecasts in a way that artificially boost the payouts they receive. Companies like Unilever have countered this by tracking forecast accuracy over time to isolate patterns and trends that drive managerial bias and change behaviours.
Another good example comes from the finance team at pharmaceutical company Eli Lilly. This team ties forecast accuracy metrics to each business unit’s incentive scheme to spur FDs and CFOs to focus on accuracy, rather than their instincts, which might in turn lead them to set low targets or get side-tracked by their own personal goals.
Other companies improve accuracy by holding stakeholders accountable for their forecasts and encouraging better performance. Stanley Black & Decker’s finance team measures forecasting success according to five metrics that measure both quantitative and qualitative performance. To isolate controllable inputs like process flaws and managerial bias, the team compares these forecast performance scores across business units and takes action based on what it finds.
Being proactive about learning from past mistakes is also critical. Too many FP&A teams churn out reports and variance analyses each forecasting cycle without incorporating learnings from past reports. There are companies that do this well, however. One global industrial company we work with speaks with business managers to identify and share the top variance drivers with forecasting stakeholders in an actual scorecard. This in turn is used to inform future forecast iterations.
The best thing any FD or CFO can do is to implement both after-the-fact learning processes and pre-emptive procedures that limit inaccuracy as much as possible. Once you accept that perfection is impossible, it will be easier for your team to concentrate on areas that improve future forecasts and inform better business decisions going forward.
The golden rule comes down to this: control what you can and learn from what you can’t. If you follow this formula, you’re sure to see the accuracy of your forecasts improve.
Paul Dennis and Peter Young, CEB