The Next Generation of Inventory Optimization has Arrived

Size: px
Start display at page:

Download "The Next Generation of Inventory Optimization has Arrived"

Transcription

1 The Next Generation of Inventory Optimization has Arrived Cutting-edge demand classification technology integrated with network optimization and simulation enables cost reduction and increased inventory availability Executive Summary Finding the balance between overstocks and lost sales is necessary for operational growth and health. With myriad inventory applications available, why are inventory inefficiencies still so prevalent? The answer is that existing inventory applications fall short in three key areas: 1. They assume that all demand is smooth and normally distributed 2. They only consider safety stock for the existing supply chain structure 3. They do not simulate real-world behavior or enable what-if analysis Breakthrough technology for inventory modeling and policy design is now available. LLamasoft Inventory Guru enables companies to analyze and properly categorize demand, factor all aspects of inventory for both existing and new supply chain structures and simulate real-world behavior to enable true what-if capabilities. The result is a prescription for the right levels of working capital across an ever-changing business. Advantage 1: Understanding Demand Simply stated, safety stock is your insurance against variability in the supply chain. One of the biggest sources of variability in the supply chain is demand, and demand can be highly unpredictable, occurring in patterns from fast to slow, smooth to erratic, intermittent to consistent. Despite the fact that there are many widely-varying demand patterns, most inventory optimization tools assume that all demand is normal, leading to either too much inventory or stock-outs and lost sales. This problem is further complicated when the supply chain has multiple tiers, numerous sourcing possibilities, shared components and varying service commitments. Planners must decide how much of each item to keep at each level, or echelon. This is historically an arduous and inexact process, involving gathering statistics from historical demand and determining how to classify outliers and extremely slow-moving demand. A Better Way to Understand Demand for Right-Sized Safety Stock: Adaptive Intelligent Inventory Optimization (AI+IO) Intelligently adapting to widely-varying demand behaviors is the key to better safety stock results, and LLamasoft Inventory Guru is the tool that helps companies do it. Inventory Guru, a comprehensive multi-echelon inventory optimization software solution, acts as a guard against inventory variability, enabling companies to save money while still meeting service requirements. Examples of 10 demand patterns recognized by AI+IO Non-Intermittent Intermittent Smooth (Fast) Erratic Slow Lumpy Clumped Unit-Sized Demand Low Variable High Variable

2 Inventory Guru is powered by Adaptive Intelligent Inventory Optimization (AI+IO) technology, a breakthrough demand classification algorithm created by LLamasoft. The result of over 10 combined years of applied research, AI+IO combines advanced classification and segmentation of demand patterns with dynamic inventory target setting to help companies achieve the lowest possible working capital to meet their target service levels. Through a simple user interface, AI+IO thoroughly analyzes and automatically classifies the underlying demand patterns into one of 10 demand categories. Once the demand patterns are analyzed and classified, the AI+IO technology applies the appropriate inventory optimization algorithm to set the proper reorder points and quantities that will be required to meet the user-defined service targets. Which Is the Better Fit? Inventory Guru vs. Other Tools The better-fitting curves made possible by AI+IO technology result in right-sized inventory. Figure: (Left) Actual Probability Density Function (PDF) of a real world demand pattern. (Middle) PDF AI+IO applies when optimizing safety stock. (Right) PDF common inventory optimization tools apply when optimizing safety stock. (Effect) AI+IO accurately represents actual demand pattern and recommends closer to optimal safety stock levels compared to other tools for the same service level. Actual density Density created by AI+IO Density created by other inventory optimization tools Customer Demand Class Summary Intermittency Variability Demand Class Clumpiness Extremely Slow Intermittent Slow Unit-Sized Highly Variable Slow Low Variable Lumpy Slow Clumped 0.06% 1.87% 12.70% 11.20% 8.21% Baseline 37.07% Demand Class Details Erratic Extremely Slow Lumpy Slow Slow-Highly Variable Non-Intermittent Erratic 16.57% Slow-Low Variable Smooth 11.52% Smooth 0% 10% 20% 30% 40% 50% Figure: (Above) Example of classification of varying demand patterns Case Example: Demand Analysis and Safety Stock Optimization A distributor of convenience store and food service products needed to reduce inventory holding cost and interface inventory optimization with its SAP system for continual strategic inventory optimization. Using LLamasoft Inventory Guru, the company explored trade-offs between cost and service time, replenishment frequency, order size and variability. With Inventory Guru s ability to incorporate bills of materials and solve multi-echelon safety stock optimization for all associated products in one run, results identified a 16 percent reduction in holding costs through safety stock and cycle stock reductions, as well as savings found by postponing production in several cases. The company continues to work with LLamasoft to set up a repeatable process for model inputs for a broader range of products, and plans to implement the solutions in SAP for their global inventory planning LLamasoft, Inc. All rights reserved.

3 Example: How New Demand Classification Technology Can Reduce Cost Below is an example of a historical demand series that exhibits lumpy behavior during lead time. AI+IO correctly identifies the pattern and applies the appropriate probability function for calculating safety stock. Traditional tools, on the other hand, assume normal distribution and thus overestimate the quantity of safety stock necessary to cover the variability. In the example below, the target fill rate for this SKU was 95 percent. Using discrete event simulation to test the results, we can see that AI+IO recommends nearly 30 percent lower inventory while still achieving the target fill rate. Traditional tools that assume normal demand will significantly over-stock this item, leading to increased inventory cost. AI+IO Safety Stock Results Target Fill Rate = 95% Traditional Tool Safety Stock Results Target Fill Rate = 95% AI+IO Recommended Safety Stock 698 Recommended Safety Stock 984 Simulation Average Inventory 702 Simulation Average Inventory 986 Simulation Average Backorders 3 Simulation Average Backorders 1 Simulation Service Achieved 95% Simulation Service Achieved 99% Case Example: Demand Analysis and Safety Stock Optimization A global manufacturer of construction equipment needed to reduce inventory holding costs and improve response times for spare part deliveries to customers. Using LLamasoft Inventory Guru to evaluate alternate supply base strategies and determine the optimal safety stock placement, the company modeled sourcing strategies across its large component supply base. LLamasoft AI+IO technology was used to evaluate daily order history and help the company better understand the relationships of demand variability in individual components and groups of components that make up spare part kits. By modeling various scenarios with differentiated sourcing structures the company was able to understand the trade-offs between inventory holding cost and service in the network. The company was also able to determine the optimal safety stock targets for spare part components and component kits that are sold as finished goods to end customers. In total the modeling effort was able to identify percent of potential cost savings in the supply chain LLamasoft, Inc. All rights reserved.

4 Advantage 2: Covering all Aspects of Inventory for Both Existing and New Supply Chain Structures There are multiple types of inventory held throughout the supply chain, including cycle stock, pre-build stock and work-in-progress (WIP), so why do most inventory applications only focus on safety stock? And why can t most applications consider alternative supply chain structures? Answer: Most inventory applications do not include network optimization capabilities. Network optimization is the technology required to consider the optimal flow of products throughout the supply chain and inherent cycle stock given the trade-offs between a host of variables including transportation modes and facility locations. Multi time-period network optimization is also the technology required to determine when products need to be pre-positioned to accommodate seasonality or capacity constraints. LLamasoft offers both network and inventory optimization within the same data model and user interface, enabling an evaluation of all aspects of inventory including safety stock, cycle stock, pre-build stock and WIP. This integration of network and inventory optimization also enables analysts to consider completely new supply chain structures such as the opening/closing of distribution centers or manufacturing locations, then immediately evaluating the optimal inventory requirements for these new potential networks. Multi time-period network optimization As-is supply chain Case Example: Using Network Optimization to Reduce Risk For a milk and dairy products producer, commodity price fluctuations are a way of life. The company produces different products based on variable market conditions and pricing. Recently, the company began expanding into new regions and utilizing ocean carriers. The challenge was now pricing fluctuations as well as distribution variability. The ocean carrier could arrive at its destination in 45 days but could take as many as 120 days. By building a multi-time period network optimization model and testing different scenarios, the company was able to make more accurate inventory stocking decisions based on unpredictability of new stock arrivals, thus lowering its risk levels in the face of lead time variability. New optimal structure LLamasoft, Inc. All rights reserved.

5 Advantage 3: The Power of Adding Simulation to Inventory Optimization Inventory Guru s optimization capabilities can identify millions of dollars in annual cost savings through right-sizing inventory levels. However, optimization technologies have limitations to how well they can predict operational performance under the stress of real-world day-to-day variability such as transportation delays, production lead times and demand. The only way to truly test the effects of these multiple levels of variability is to use discrete event simulation. Simulation can digitally run the clock to fill each order and schedule each shipment to provide detailed performance metrics such as on-time deliveries, stock-outs or capacity bottlenecks. Once an inventory strategy has been chosen using optimization, simulation is an excellent method to test it further, providing validation that the new inventory policies work as intended. LLamasoft s SimServer engine is integrated as part of the Supply Chain Guru modeling and design software platform. Supply Chain Guru is the industry s only supply chain and inventory optimization solution with an integrated, enterprise-scale simulation engine. This unique Historical inventory pattern integration of optimization and simulation technologies provide a level of confidence not available in typical inventory optimization technologies. Simulated new inventory pattern Case Example: Using Simulation to Smooth Production Planning and Order Variance A consumer goods company wanted to get better performance out of their manufacturing plant. Instability in ordering patterns combined with ad-hoc scheduling led to inefficient use of machines in the plant. Managers thought efficiency could be improved by segmenting fast and slow moving inventory and initiating regular schedules for their manufacture. They wanted to use simulation to test out the improvements from this strategy before implementing them. The simulation suggested that these strategies could result in a 40 percent improvement in stock levels together with dramatic improvements in workcenter change-overs and adherence to production plan LLamasoft, Inc. All rights reserved.

6 Case Example: Inventory Optimization Validated with Simulation A global healthcare and consumer goods company needed to better understand their customer demand patterns, improve customer service and identify more effective postponement and risk-pooling strategies to enable overall reductions in working capital. SAP (including nine levels of BOMs) and forecast data were entered into Inventory Guru, generating a clear picture of demand product classes and associated customer/product demand patterns. Multi-echelon safety stock targets for finished goods, WIP and raw materials were generated providing actionable data well beyond their single echelon finished good safety stock targets. These insights, validated with simulation, showed the company s current safety stock methods were generating significant excess inventory or missing their service metrics for most products. Furthermore, the Inventory Guru model was able to show the network-wide view of cycle stock, WIP, in-transit and safety stock inventories identifying a $2 million potential opportunity for inventory reduction. Simulation graph of on-hand inventory Conclusion Businesses striving for supply chain excellence and the evasive and ever-changing key to customer loyalty must find the balance between overstocks and lost sales. Now, through game-changing inventory modeling technology from LLamasoft, businesses can save money and increase inventory availability while still meeting service requirements and saving customer sales. By demystifying the classification of varying demand patterns driving safety stock levels, LLamasoft enables companies to finally conquer the inventory challenge and achieve a healthier overall business. For more information, visit About LLamasoft, Inc LLamasoft supply chain design software helps organizations worldwide design and improve their supply chain operations. LLamasoft solutions enable companies across a wide range of industries to model, optimize and simulate their supply chain network, leading to major improvements in cost, service, sustainability and risk mitigation. Headquartered in Ann Arbor, Michigan, LLamasoft is a leader in supply chain excellence and innovation, advancing technology focused on continuous improvement of enterprise supply chains for the world s largest organizations LLamasoft, Inc. All rights reserved LLamasoft, Inc. All rights reserved. LLamasoft, Inc. 201 South Main Street, Suite 400 Ann Arbor, Michigan 48104, USA Phone: Fax: LLamasoft.com Info@LLamasoft.com