These figures show that efforts at inventory optimization aren't bearing fruit as retailers grapple with the conflicting demands of omni channel retailing and, more recently, the unpredictability posed by trade wars and other global disruptions.
The importance of effective inventory management is shown by the Deloitte Global Chief Procurement Officer Survey of 2018 where supply chain managers rated cost reduction and managing risk as two of the top three business strategies going forward. It's clear a rethink of inventory optimization practices is needed as conventional processes aren't yielding the results organizations need.
The underlying concept of inventory optimization is clear: to manage inventory levels to meet predetermined service levels while minimizing the amount of working capital that's unproductively tied up in inventory.
The difficulty is that future demand is difficult to predict, and the risks of stock-outs include:
It's for this reason that many supply chain managers feel carrying higher levels of stock is a lesser evil than risking the adverse consequences of a stock-out.
Factors that are typically considered include lead time, shipping, daily usage and economic reorder quantities. Supply chain managers use this information to calculate optimal stock levels and reorder points to ensure adequate stock availability. They may also selectively categorize inventory using tools such as the ABC analysis method, which determines inventory levels based on sales volume and value.
According to Deloitte's survey, only one-third of procurement managers use advanced analytics and other modern technologies to guide their procurement and inventory management policies. This may possibly explain why inventory levels are still climbing. The truth is that the pace of business is such that few organizations can succeed if they persist with manual and legacy inventory optimization practices
Various software solutions and techniques have been developed to aid supply chain managers to optimize inventory. Most supply chain and ERP solutions include analytical tools for automatically calculating optimal inventory levels and determining the best replenishment practices. These solutions include features such as the ability to identify safety stock levels, automatically calculate reorder points, and highlight potential stock-outs. Many solutions can track inventory in real time, minimizing delays and reducing the risk of stock-outs of fast-moving goods.
No inventory optimization solution can hope to succeed unless it includes demand planning capabilities. Demand planning acts as a link between sales forecasts, production plans and the supply chain strategy. It fills the gap between historical turnover and current reality by allowing for future demand.
Demand forecasting allows supply chain mangers to track metrics and adjust stock levels to cater for different demand patterns, such as:
Demand planning helps supply chain mangers to better control inventory levels to ensure inventory policies are in line with those of the organization.
Working out how to deal with unpredictability is the greatest challenge facing supply chain managers. It's a requirement that simple inventory optimization algorithms don't handle effectively because they're based on fixed parameters that don't change with time. While these conventional algorithms work fine when demand is steady and predictable, erratic demand causes stock-outs or, if parameters are set to compensate, excessive stock. The same applies to ABC stock holding methods that are relatively inflexible and not suited to inventories containing thousands of stock items.
Another stock category that defies conventional logic is long-tail stock items that typically sell in low numbers, yet their combined off-take is such that they can represent a significant percentage of sales. Conventional inventory strategies, as well as ABC stock holding methods, set stock levels of long-tail items low as to ensure these items are never available. This phenomenon is seen in many chain stores that stock a range of almost identical fast-moving, low-cost products, and is a strategy that drives some discerning customers away.
The main weakness of most inventory optimization algorithms is that they use historical data as their primary input for calculating stock levels. Additionally, inventory planning processes are often carried out in functional silos that don't take into account other functions nor consider the challenges of omni channel retailing that demands a different inventory management approach. Traditional multi-echelon inventory optimization goes some way in overcoming these limitations because it considers stock in different locations as well as in different stages in the supply chain. This includes counting components in their raw form, items in manufacture, and stock that's on the water or in transit.
Perhaps the most exciting inventory optimization possibilities include the use of machine learning, as well as predictive and prescriptive analytics. These model the environment, taking into account a range of external factors to predict demand and determine the best inventory policies that satisfy demand at the lowest overall cost.
Two factors have created an environment for next generation inventory management solutions to flourish using machine learning and artificial intelligence. These are the explosion in the amount of data available to organizations and the increase in accessible computing power. Taken together, these factors make it possible to analyze large volumes of data quickly and model different scenarios to determine the best inventory management strategies in any situation.
This technique, known as prescriptive analytics, draws on the wealth of information available within the organization's inventory management, ERP, financial and CRM systems as well as its market intelligence. Using powerful nonlinear solver programs, prescriptive analytics help organizations determine inventory optimization strategies, taking into consideration factors such as:
Using prescriptive analytics, it's possible to create a model that accurately represents the organization's supply chain and determine effective inventory strategies that:
Many inventory optimization software solutions incorporate sophisticated algorithms and mathematical formulae for calculating inventory levels. However, relatively few have the capability to model the environment and consider the effect that external factors have on supply and demand and ultimately the right inventory level. This is made more difficult by the need to satisfy the often-conflicting demands of omni channel commerce. Additionally, because the business environment changes so quickly, there's the need to be able to proactively revise inventory strategies.
Most inventory management solutions will optimize inventory based on a predetermined management strategy. What they can't do is determine what that strategy should be. This is the benefit of a prescriptive analytics inventory optimization solution because of its ability to analyze the data and determine the best inventory strategy to meet corporate goals such as service levels, working capital and profitability.