13 THE OUTSIDE-IN PLANNING HANDBOOK | 2023
13
A Closer Look at Demand
Traditional planning processes use customer order and shipment patterns as a proxy for demand. The belief is that future demand can be based on mining the patterns of history. But what happens when history is no longer a good predictor of future demand? Or does the market shift faster
than an organization can make decisions? In these cases, the organization quickly learns that the order falls short as a demand predictor. In Table 2, we share the issues of order- based forecasting.
Table 2. The Fallacy of Order-Based Forecasting
Assumption
Reality
Insight
1. Today’s order patterns can be predicted by historical demand. 2. Collaboration across business teams can improve the forecast.
History is not a good predictor of current demand. Most organizations introduce political bias based on functional incentives and decrease the Forecast Value Added (FVA). The longer the tail of the supply chain, the larger the issue with demand latency.
The pandemic period distorted demand, and current levels of migration further distrupt the patterns. Collaborative forecasting increases process latency and time to action. In process-based industries like beverages, chemicals, consumer electronics, and non-durable goods, the elongation of the tail of the supply chain made demand cycles longer than supply cycles over the last decade. Yet few companies measure and manage demand latency and the bullwhip effect. Monitoring the shifts in flow based on market data improves sensing and time for action.
3. The order pattern represents current demand.
4. Demand is best managed as time-phased transactional data to be consumed by supply. 5. Products are forecastable by conventional planning models.
The average supply chain has multiple supply chains best represented by flow.
With the increase in product complexity and regional assortment, most companies greatly decreased the forecastability.
Today, only 40-50% of products are forecastable by conventional planning models/techniques.
Forecast Value Added (FVA) is a valuable technique to analyze value from the forecasting process.. A positive FVA indicates an improvement over the naïve forecast (prior period sales), whereas a negative FVA means that the methods and technologies are making the forecast worse, not better. In eight
An additional error in the forecast has a multiplier effect on inventory and costs. Of the 154 students in the class, only one was measuring FVA at the start of the course. In the shift from inside-out to outside- in processes, there is a shift from a focus on error to FVA. The reasons are many. Degradation of implementations, turnover of employees, dirty data, wrong models, political bias, and market shifts. Overall, forecastability is worse in most companies than in the days before the pandemic, and few companies are measuring these shifts and testing their
______________________________ Forecast Value Added (FVA) analysis measures the improvement of the output of the forecasting process against the naive forecast. The naive forecast definition varies by demand stream. _____________________________
out of ten companies that we worked with in the 2023 training, we found a negative Forecast Value Added (FVA) result. Most companies that measure FVA soon realize that despite all the money spent on advanced technologies and processes, the methods are making things worse.
Made with FlippingBook - Share PDF online