Five Things You Need to Know about Inventory Optimization in the Digital Age
With all the buzz around digital transformation, someone peering in from outside supply chain may assume most businesses have already digitized essential tasks like inventory planning. The reality is that this transformation takes time and a change of mindset. It’s about learning to transition from long-held inventory systems and processes, having new conversations across logistics and sales, and the shifting role of supply chain planners. This article will share five things you need to know about optimizing inventory in the digital age.
1. Long-tail demand complexity has completely changed how we optimize inventory. Items with intermittent, unpredictable or “long tail” demand are a growing part of business for manufacturers, distributors, and retailers, and it’s making demand forecasting and inventory management a headache. There are many drivers of this intermittent demand, including product proliferation, faster replenishment, and extended supply chains (increased collaboration between vendors, distributors and retailers causes demand to be disaggregated into smaller streams where intermittent behavior becomes common). Traditional supply chain solutions were not designed for high variability demand. Inventory mixes and service levels get out of balance across the network—and out of line with business objectives. Trust in the planning process breaks down, and when planners stop trusting forecasts they tend to load up on safety stock. This leads to excessive costs, waste and obsolescence.
2. It’s impossible to optimize inventory in a complex environment using ABC inventory analysis and spreadsheets. As many as 75 percent of companies are still trying to use spreadsheets and/or ABC Classification to optimize inventory. Since many companies have hundreds of thousands, or even millions of combinations, it’s impossible to identify a service level for every individual SKU-Location. ABC classification, using a 3×3 matrix, provides a way to simplify an SKU portfolio to make safety stock calculations more manageable. However, the item classification and aggregated service level are calculated with a ‘trial and error’ process that cannot possibly identify the truly optimal stocking level and service for each SKU-Location combination given the complexity of today’s multi-echelon inventory networks. Some traditional inventory management tools try to address this by providing an 8×8 ABC matrix per location. The workload to define and continuously maintain these matrices becomes very intense. What’s more the odds of optimizing inventory are even lower than with the 3×3 matrix because there are more possible combinations. Planners relying on last century’s solutions–especially with today’s long-tail demand complexity–stand little chance of being able to meet both service level and financial goals in a sustainable way.
3. “Service-driven” inventory optimization is a new and proven way to optimize inventory in the digital age. Unlike ABC, which has an operational perspective, service-driven inventory optimization (IO) centers on sales, marketing and customers. It uses categories known as “service classes” that sales and marketing people can easily relate to, like “own-brands”, and “critical spare parts”. Then it optimizes every SKU-Location against a target service level for each service class. The result is that you arrive at an aggregated service class goal with the lowest possible stock investment. This approach truly optimizes your inventory at every stocking point.
What happens next is very different from traditional inventory management. By applying ”stock-to-service” curves, software designed for this purpose optimizes every single service level and safety stock level by SKU-Location. You can even automatically define different target ranges for each service class. For example the aggregated service level goal for “accessories” could be 93% with a lower limit of 89%. The inventory optimization tools then calculates how to service this range in a way that minimizes stock investment.
4. Probability forecasting is the key to service-driven inventory optimization. Let’s face it: the age of forecasting as the end goal has passed. You can try and try, but you’ll never be able to do better than incremental forecast improvements. Probability forecasting identifies a range of outcomes and the probability of each of those outcomes occurring. A probability forecast that takes uncertainty into account helps you manage risk. It’s not just about improving averagedemand predictions but assessing the entire range of possible outcomes—including demand volatility, which has the biggest impact on service levels. This information is used to calculate the optimal inventory targets. So essentially, probability forecasting helps you reduce risks related to demand volatility–excess and obsolete inventory, expediting costs, and capacity challenges, and more–by decoding that uncertainty into a smarter stock mix.
5. Machine learning improves inventory optimization–and elevates your planners. Gartner predicts that, “by 2020, 95% of SCP vendors will be utilizing supervised and unsupervised machine learning somewhere in their SCP solutions.” Machine learning’s ability to find patterns in huge data sets and also get smarter over time make it the perfect complement to human inventory planning efforts.
Aston Martin’s spare parts operation used a machine learning engine to analyze the vast array of historical data collected over decades and pick out eight completely new categories of behaviour. It used these categories to optimize inventory replenishment plans, raising service levels while reducing inventory costs by 18 percent. According to Nick Wilson, Senior Inventory Planner, Parts Operations, the new focus on seasonality has been transformational: “The great thing about the eight categories is that people can see them. This has not only educated the purchasing team in important new skills, but has also really given them confidence in the planning system.”
While we believe these five points are critical to optimize inventory in the new world of demand uncertainty and supply chain complexity, I’ll add an unofficial #6: The biggest inventory optimization challenge for companies today may be learning to trust highly sophisticated planning systems that aren’t transparent like spreadsheets and other simple tools. As Nucleus Research analyst Seth Lippincot reflected in its 2019 report Answering Inventory Complexity, “changing behaviors can often be more difficult than changing software, and building trust in the recommendations from a tool does not happen overnight.” Until business leaders are more willing to take their hands off the wheel, there will always be a level of productivity and financial benefits that will remain unattainable.
 Gartner: Current Use Cases for Machine Learning in Supply Chain Planning Solutions. Analyst: Tim Payne. 19 May 2018