John E. Boylan - Intermittent Demand Forecasting

Intermittent demand for products appears sporadically, with some time periods showing no demand at all. Moreover, when demand occurs, the demand size may be constant or variable, perhaps highly so, leading to what is often termed ‘lumpy demand’. Later in this chapter, we discuss why forecasting sporadic and lumpy demand patterns is a very difficult task. Specific characterisations of intermittent demand series are considered in detail in Chapters 4and 5.

Intermittent demand items dominate service and repair parts inventories in many industries (Boylan and Syntetos 2010). A survey by Deloitte Research (2006) benchmarked the service businesses of many of the world's largest manufacturing companies with combined revenues reaching more than $1.5 trillion; service operations accounted for an average of 25% of revenues. In addition to their contribution to revenues, these items present a distinct opportunity for cost reductions. Maintenance, repair, and operations (MRO) inventories typically account for as much as 40% of the annual procurement budget (Donnelly 2013). Increased revenues and reduced costs naturally lead to increased profits. Many organisations have repeatedly testified to the importance of after‐sales services for their businesses and the profits they generate. Companies such as Beretta, Canon, DAF Trucks, Electrolux, EPTA, GE Oil & Gas, and Lavapiu have reported contributions of the after‐sales services to their total profit of up to 50% (Syntetos 2011). Comparable numbers have been reported by Gaiardelli et al. (2007), Kim et al. (2007), and Glueck et al. (2007), while after‐sales service has been identified as a key profit lever in the manufacturing sector (Manufacturing Management 2018).

Intermittent demand items are at the greatest risk of obsolescence. Many case studies (e.g. Molenaers et al. 2012) have documented large proportions of ‘dead’ (obsolete) stock in a variety of industries, with serious environmental implications. However, under‐stocking situations may be as harmful, given the potentially high criticality of the items involved. In civil aviation, for example, lack of spare parts is one of the major causes of ‘aircraft on ground’ events (problems serious enough to prevent aircraft from flying). Badkook (2016) found that a quarter of the aircraft in an (un‐named) airline's Boeing 777 fleet were affected by such aircraft on ground events over a year.

Defence inventories, which are highly reliant on spare parts, have been repeatedly identified as a high risk area with a direct impact on a nation's economy. In the United States for example, the Department of Defense (DoD) manages around five million secondary items. These include repairable components, subsystems, assemblies, consumable repair parts, and bulk items. They reported that, as of September 2017, the value of the inventory was $93 billion (GAO 2019). Although a matter of concern, there had been no substantial reductions in inventory values over the previous decade (being, for example, $95 billion in 2013 and 2010; GAO 2012, 2015).

A major determinant of the performance of an inventory system is the forecasting method(s) being used to predict demand. Inaccurate forecasts lead to either excess inventory or shortfalls, depending on the direction of the forecast error. Over‐forecasting can lead to holding stocks that are simply not needed. According to the US Government Accountability Office (GAO 2011, p. 11), ‘Our recent work identified demand forecasting as the leading reason why the services and DLA [Defense Logistics Agency] accumulate excess inventory’.

Unfortunately, progress in improving forecasting and inventory management has been slow in many industries, with the defence industry being a case in point. The GAO of the United States reported, ‘Since 1990, we have identified DoD [Department of Defense] supply chain management as a high‐risk area due in part to ineffective and inefficient inventory management practices and procedures, weaknesses in accurately forecasting the demand for spare parts, and other supply chain challenges. Our work has shown that these factors have contributed to the accumulation of billions of dollars in spare parts that are excess to current needs’ (GAO 2015, p. 2). Progress in inventory management has been made since then, especially with regard to the visibility of physical inventories, receipt processing, and cargo tracking (GAO 2019). These improvements in information systems have led to inventory management being removed from the list of high‐risk areas. However, it is notable that no claims have yet been made for corresponding improvements in demand forecasting.

Moving beyond the after‐sales industry, and the defence sector, we now examine the potential benefits that may result from intelligent intermittent demand forecasting for the wider economy. Purchased goods inventories and their management are significant concerns for firms wishing to remain competitive and survive in the marketplace. According to the 26th Annual State of Logistics report (CSCMP 2015, statistics referring to 2014), the United States alone has been sitting on approximately $2 trillion worth of goods held for sale. According to the same report, the inventory carrying costs (taxes, obsolescence, depreciation, and insurance) are estimated to be around $0.5 trillion (i.e. about 25% of the value of the goods). The total value of inventory was equivalent to approximately 14% of the US gross domestic product (GDP) in 2014. Although similar statistics have not been given in subsequent publications, the 30th Annual State of Logistics Report (CSCMP 2019) revealed that inventory carrying costs in the United States increased by 14.8% between 2014 and 2018.

These figures show that a huge amount of capital is tied up in warehouses. They also indicate that small improvements in managing inventories may be translated into considerable cost benefits. We should, therefore, not be surprised to learn that firms, from manufacturing to wholesale to retail, are currently intensifying their search for more efficient and effective inventory management approaches. Their aim is to minimise not only their direct investments in purchased goods inventory but also the indirect cost incurred in managing this inventory. In a make to stock (MTS) environment (discussed in Section 1.4.2), if there is no decoupling in terms of the ownership and location of the inventories, then these indirect costs become more significant the longer the stock remains unsold. The high volumes of stocks of intermittent demand items, and their high risk of obsolescence, should put them very high up the list of priorities for modern businesses.

Obsolescence is a very important topic for supply chain management. The complexity of supply chains, in conjunction with increasingly reduced product life cycles, is resulting in high levels of obsolescence. Molenaers et al. (2012) discussed a case study where 54% of the parts stocked at a large petrochemical company had seen no demand for the last five years. Syntetos et al. (2009b) evaluated the inventory practices employed in the European spare parts logistics network of a Japanese manufacturer. They found one case, reported in Sweden, where some parts in stock had not ‘moved’ at all over the preceding 10 years. The value of the on‐hand excess (spare parts) inventory of the US Air Force, Navy, and Army has been estimated to be of $1.7 billion, $1.4 billion, and $2.5 billion, respectively (GAO 2015). Much of this excess stock is at risk of obsolescence.

When obsolescent (or ‘dead’) stock is created, there is considerable environmental waste. Firstly, there is an environmental cost associated with producing goods that are never used. Secondly, there are environmental costs of transporting these goods to national, regional, or local stocking points. Finally, there are environmental costs of disposing of these stocks. The prevention of the accumulation of dead stock relies on accurate demand forecasts. Consequently, more accurate and robust forecasting methods may be translated to significant reductions in wastage or scrap, with considerable environmental benefits.