How Precise Forecasting Tames Item Proliferation
Accurate forecasts enable better planning, lower inventory and manufacturing costs, and improved service levels. By limiting line item proliferation, precise forecasting reduces errors significantly for new product launches.
New products lack the history required for traditional demand planning methods. In the same way that the financial industry uses algorithms to sort through masses of market data to determine what stocks to buy or sell, leading multinational manufacturers have turned to automated systems that can accurately sense what demand will be.
The difference is that demand-sensing algorithms sort through the masses of supply chain data available to manufacturers to find predictive patterns and create more accurate forecasts.
Since 2010, the number of active items grew by 31 percent compared to only a 6-percent increase in sales, according to E2open’s 2016 Forecasting Benchmark Study. As a result, average sales per item dropped 19 percent.
The study analyzed demand planning performance in consumer goods based on more than $250 billion in annual sales from the global businesses of 17 multinational companies, with 9 billion cases and more than 1 million SKUs. Findings reveal that growth through innovation strategies drives supply chain complexity instead of sales, with the rate of proliferation outpacing sales by more than five times. This additional complexity creates a challenging planning environment and increases workload.
Even more concerning is the rapid rate of product introductions and discontinuations. The number of total items offered for sale has more than tripled over the 5-year period, and many have since been discontinued—for every 100 items introduced, 86 were discontinued. Only one in 1,000 new items becomes a top seller within the first year, and more than 90 percent of new items fall to the bottom.
The introduction cycle’s sheer scale and pace raises questions about the financial advantages of current innovation strategies. Each phase-in and phase-out is associated with various supply costs, including setup changes to manufacturing, inventory of raw materials, packaging, and finished goods, as well as write-downs for obsolescence.
Not surprisingly, error and bias are much higher for new products than existing ones. Weekly error for items in the first year of sales was 66 percent, 1.4 times greater than for existing items. Bias for introductions is even more pronounced: 12 percent for new items, three times greater than for existing products.
The culture of high internal rates and competition for limited budgets likely contribute to a continued belief that new products will be winners, even when faced with contradictory sales evidence. The pervasive over-optimism is costly in that it commits limited corporate resources to underperforming items.
Looking at your network and customers is crucial for supporting product launches. Sensing demand reduces forecast error for new items by 31 percent, according to data from the Forecasting Benchmark Study. Are new products moving faster than expected? Is pickup faster in one market than another? Analyzing this data daily for every product in every stocking location allows companies to better predict demand and serve customers at a lower cost.