Think differently about forecast error

There are some practical guidelines for becoming a more sophisticated consumer and user of forecasts. Let’s me explore some thoughts that could be applied to managing the impact of demand forecast error, specifically in the context of business or supply chain management.

Defining a window of uncertainty emphasizes identifying a broad range of possibilities and acknowledging uncertainty at the outset. The window of uncertainty highlights the importance of considering multiple outcomes and not relying on narrow or overconfident predictions. This is directly useful in managing forecast error. By recognizing that demand forecasts can’t perfectly predict outcomes, decision-makers can incorporate flexibility into their strategies and approaches. The building of buffer stocks or creating adaptable production schedules would allow businesses to adjust to demand variations without being caught off-guard by extreme outliers. The broader the window, the better a company can prepare for unexpected demand spikes or drops.

There is often a launch profile that is impacted by exuberant interest, energy and focus, followed by the lag of reality, a slow start, accelerates, and eventually tapers off. This implies that potential inflexion points, or sudden increases or decreases in demand, can be anticipated by watching for subtle signals. Spotting early signs of demand shifts allows businesses to adjust their forecasts before dramatic changes happen. If a company tracks early indicators of new market trends like consumer shifts to electric vehicles or new tech adoption, it can modify production and inventory strategies before the shift accelerates, reducing forecast errors during rapid demand changes, leveraging forecasting model life cycles.

Embracing things that don’t fit encourages paying attention to outliers and anomalies, as these often signal future trends or disruptions. Outliers in sales data or consumer behaviour, such as unexpected increases in demand for niche products, should be investigated rather than dismissed. Embracing these anomalies helps avoid blind spots in forecasts, particularly in industries prone to volatility.

Holding strong opinions weakly means being willing to change forecasts as new data comes in and avoiding attachment to one specific outcome or prediction. This is critical for businesses where demand forecasts need constant revision. A company forecasts strong demand for a new product, but early sales data contradicts this. Thus, it must quickly adjust its supply and marketing strategies to avoid overproduction. Holding weakly held forecasts helps minimize the impact of forecast errors by allowing frequent recalibration based on real-time data.

Looking back twice as far as you look forward emphasizes that historical data provides valuable insights into potential future trends, but it’s essential to look far enough into the past to recognize long-term patterns. Understanding past demand trends can help mitigate errors. Considering how demand fluctuated during previous recessions or holiday seasons can inform more accurate forecasting for similar future events. However, it’s also important not to over-rely on the recent past, as newer market conditions, such as digital disruption, might not follow older patterns exactly.

Knowing when not to make a forecast is another crucial element. In times of extreme uncertainty, it might be better not to make a forecast or at least be cautious with its reliability. Some market conditions, like during a major economic or political crisis, can make demand forecasting unreliable. In such cases, businesses should invest in flexibility and risk management strategies rather than precise forecasting. During the 2008 financial crisis, many businesses struggled because their demand forecasts were not adjusted for the unexpected severity of the downturn.

There are techniques to manage the impact of demand forecast error. By applying the "cone of uncertainty," businesses are better prepared for demand volatility and can manage forecast errors without major disruptions. The "S-curve" principle helps businesses detect and react to emerging trends earlier, which reduces lag in forecast accuracy adjustments. Adopting the idea of "holding strong opinions weakly" encourages companies to revise forecasts regularly, ensuring they remain adaptable to changing market conditions. Acknowledging when not to forecast helps avoid overconfidence in situations where there is little data to support accurate predictions, shifting the focus to resilience over precision.

There are valuable strategies for managing demand forecast error by emphasizing flexibility, iterative learning, and embracing uncertainty—all key components for reducing the risks of incorrect predictions.