Submitted manuscript, 0.98 MB, PDF document
Research output: Working paper
Research output: Working paper
}
TY - UNPB
T1 - Information use in supply chain forecasting
AU - Fildes, Robert
AU - Goodwin, Paul
AU - Onkal, Dilek
PY - 2015/5
Y1 - 2015/5
N2 - Demand forecasting to support supply chain planning is a critical activity, recognized as pivotal in manufacturing and retailing operations where information is shared across functional areas to produce final detailed forecasts. The approach generally encountered is that a baseline statistical forecast is examined in the light of shared information from sales, marketing and logistics and the statistical forecast may then be modified to take these various pieces of information into account. This experimental study explores forecasters’ use of available information when they are faced with the task of adjusting a baseline forecast for a number of retail stock keeping units to take into account a forthcoming promotion. Forecasting demand in advance of promotions carries a particular significance given their intensive supply chain repercussions and financial impact. Both statistical and qualitative information was provided through a forecasting support system typical of those found in practice. Our results show participants responding to the quantity of information made available, though with decreasing scale effects. In addition, various statistical cues (which are themselves extraneous) were illustrated to be particularly important, including the size and timing of the last observed promotion. Overall, participants appeared to use a compensatory strategy when combining information that had either positive or negative implications for the success of the promotions. However, there was a consistent bias towards underestimating the effect of the promotions. These observed biases have important implications for the design of organizational sales and operations planning processes and the forecasting support systems that such processes rely on.
AB - Demand forecasting to support supply chain planning is a critical activity, recognized as pivotal in manufacturing and retailing operations where information is shared across functional areas to produce final detailed forecasts. The approach generally encountered is that a baseline statistical forecast is examined in the light of shared information from sales, marketing and logistics and the statistical forecast may then be modified to take these various pieces of information into account. This experimental study explores forecasters’ use of available information when they are faced with the task of adjusting a baseline forecast for a number of retail stock keeping units to take into account a forthcoming promotion. Forecasting demand in advance of promotions carries a particular significance given their intensive supply chain repercussions and financial impact. Both statistical and qualitative information was provided through a forecasting support system typical of those found in practice. Our results show participants responding to the quantity of information made available, though with decreasing scale effects. In addition, various statistical cues (which are themselves extraneous) were illustrated to be particularly important, including the size and timing of the last observed promotion. Overall, participants appeared to use a compensatory strategy when combining information that had either positive or negative implications for the success of the promotions. However, there was a consistent bias towards underestimating the effect of the promotions. These observed biases have important implications for the design of organizational sales and operations planning processes and the forecasting support systems that such processes rely on.
KW - Demand planning
KW - Sales and Operations Planning
KW - Behavioural operations
KW - Forecasting support systems
KW - Promotional planning
KW - Information effects
M3 - Working paper
T3 - Lancaster University Department of Management Science Working Paper Series
BT - Information use in supply chain forecasting
PB - Department of Management Science, Lancaster University
CY - Lancaster
ER -