Inventory Optimization and Cognitive Computing

Stephen DeAngelis

March 10, 2015

Inventory is one of the necessary evils associated with doing business. In a manufacturer’s ideal world, a product would be made, shipped, and sold based on solely demand with no need to warehouse merchandise or watch products linger on store shelves. Unfortunately, that world will never exist. Demand for products will continue fluctuate and, if inventory isn’t available when demand increases, sales opportunities will be lost. As a result, a lot time and effort is spent trying to optimize inventory. According to Will Benton, chief executive officer with GAINSystems Inc., supply chain optimization is a simple concept. “It tells people if, where and how much to stock.”[1] Sounds simple, but we all know that supply chain optimization is challenging. Difficult or not, supply chain optimization is essential for any well run business to maximize its profits. As Rich Sherman (), Principal Essentialist at Trissential, puts it, companies optimize their inventories “because that’s where the money is!”[2]

 

Before going more deeply into the subject of inventory optimization, let’s talk terms. Sometimes “inventory optimization” is referred to as “inventory management” or “inventory control.” Some pundits assert that those terms are not synonymous and they seem willing to fight to the death over their position. I agree with Art van Bodegraven, president of Van Bodegraven Associates, and Kenneth B. Ackerman, president of The Ackerman Company, who wrote, “Is there a conflict between inventory management and inventory control? Well, sort of. Not that most of us care, but the purists can get a bit savage in defense of their position(s). … In our view, this inventory subject must be approached holistically, integrating planning (management) and control (transaction execution) for strategically optimal results. This perspective means that focusing on specific and limiting definitions of elements of either management or control (however they are defined) is a loser’s game.”[3] Whatever definitions you want to use, the end game is making inventory as productive as possible.

 

As beneficial as they have been, Sherman notes that traditional inventory management systems have nonetheless been inadequate for the task at hand. “While inventory optimization tools are intended to support better decisions,” he writes, “most are executed outside the daily operating processes.” He elaborates:

“They set guidelines, determine optimal quantities/run lengths, account for seasonality, events, and various demand and lead-time variability, but they don’t provide the daily decision support tools to enable planners/schedulers to actually manage procurement, production and deployment of inventory; and neither does the enterprise system. The lack of enterprise system support provided to the planner/scheduler to make daily and often hourly adjustments to actual production requirements results in the perpetuation of custom spreadsheets and reliance on tribal knowledge to manage inventory, let alone optimize it. While most inventory optimization tools provide significant ROI (the opportunity is that significant), only a few can provide the real solution: operating decision support with embedded optimization that is integrated across the organization to enable planners/schedulers to adjust their decisions to volatility and variability as they occur.”

At Enterra Solutions®, we believe that cognitive computing can help address the problem identified by Sherman. Because cognitive computing systems, like Enterra’s Cognitive Reasoning Platform™ (CRP), can integrate data and analyze numerous variables to discover relationships and provide insights, they can provide the embedded optimization aides for which Sherman is looking. An article in Supply Chain 247, further highlights why a system that analyzes all essential variables is critical for inventory optimization:[4]

“With out-of-stocks representing both a significant cost and a reputational concern, inventory management continues to be a thorn in the side of Consumer Packaged Goods (CPG) companies. The reason? Traditional inventory systems have always focused on improving forecasting, while the forecast is only a small part of the overall inventory management problem. Most of the issues occur at the retail store shelf, which aren’t included in forecasts. A new paradigm is needed — one which brings all factors into play, and which enables a proactive approach to solving inventory problems before they occur.”

At Enterra® we call this approach the Pre-Inventory Prediction™ (PIP) System. The PIP System can help manufacturers get the right product at the right place at the right time. It allows manufacturers to better manage their inventory (before it is available in the warehouse for allocation) and thus better manage their retailer orders. PIP monitors manufacturers’ inbound shipments, identifies potential Product Not Available (PNA) situations, and provides an advanced allocation workbench (which also provides Available to Promise (ATP) functionality). Benton told the SupplyChainBrain staff, “Companies need to assess the tradeoff between manufacturing efficiency, whereby plants turn out large lots of a single product at one time, and the benefits of keeping inventory low.” He continues:

“In addition, … merchandisers often overlook the cost of a stockout, which results in three possible scenarios: expedited transportation, substitution of a higher-priced item at a lower price, or loss of the sale. None should be considered an acceptable outcome in the highly competitive world of retail. The data that supports inventory optimization comes from multiple sources, including historical performance and customer preference. … Companies also need to factor in the degree to which a given product is available from competitive sources.”

Karin L. Bursa, Vice President of Marketing at Logility, indicates that a lot of companies have been fooling themselves into thinking they are optimizing their inventories. “What is the extent of your commitment to optimize,” she asks. “In a survey conducted by Logility and Consumer Goods Technology, more than 50 percent of companies claim they optimize inventory across several stages of their supply chain. But do they really?”[5] Bursa reports that “the majority of companies are at one of three levels of inventory optimization sophistication”:

“1. Basic Math. Inventory targets are not managed. Instead, the supply chain team nets out all future demand and maintains stock for it. Here, ‘safety stock’ is only the additional amount above what is in the plan beyond the next 18 months. This approach to inventory can only work when demand and supply are both deterministic: a highly unlikely prospect.

“2. Rule of Thumb. This is the oldest and most common method of setting inventory levels. Based on the forecast, a company will hold inventory sufficient to meet a specific number of forward weeks of supply. This approach neglects the primary drivers of excess inventory: demand uncertainty, lead-time variability and product velocity.

“3. Single Stage Inventory Optimization. This is usually the first good step a company takes towards optimization. Single stage models are calculators, perhaps in the form of fancy spreadsheets. Localized optimization can improve service levels and free up working capital, but the single stage approach ignores the fact that excess inventory buffers are produced by interrelationships between stages.”

Bursa insists that a more holistic approach is required — a multi-echelon approach. When implemented properly, she notes, a good inventory optimization system “frees up millions of dollars in working capital by reducing inventory end-to-end without negatively impacting service levels.” That is a goal worth striving to achieve.

 

Footnotes
[1] Will Benton, “What Is Inventory Optimization All About?SupplyChainBrain, 15 November 2013.
[2] Rich Sherman, “Why Optimize Inventory? Because That’s Where the Money Is!SupplyChainBrain, 10 March 2014.
[3] Art van Bodegraven and Kenneth B. Ackerman, “Inventory management vs. inventory control,” DC Velocity, 16 April 2012.
[4] Orchestro, “The Problem with Traditional Inventory Management,” Supply Chain 247, 23 February 2015.
[5] Karin L. Bursa, “What Do We Mean by ‘Inventory Optimization?’,” Supply Chain Digest, 15 December 2011.