MEIO, or multi-echelon inventory optimization, optimizes inventory levels across multiple locations, such as suppliers, manufacturing sites, distribution centers, and customers, by considering their interdependencies, rather than by treating each location in isolation.
MEIO packaged software applications entered mainstream commercial use in the 2000s to improve upon overly simplistic policies for setting safety stock levels. But early MEIO approaches fell short because they were difficult to use and produced unrealistic recommendations.
Today, with the rise of AI and machine learning, MEIO is experiencing a revival. New-generation solutions address many of those early issues, producing impressive results in modern supply chains.
A brief review of inventory optimization will be helpful before we look at where MEIO has been and where it’s going.
Inventory Management 101
Everyone wants high customer service levels with the lowest possible inventory.
But simply setting a 99% service level using the safety stock formula doesn’t give you a 99% service level.
Why not?
The safety stock formula assumes forecast error is normally distributed. But real life never works that way. And it will take you a long time to find out if your stock levels are suboptimal. By that time the damage is done.
So what’s the best way to minimize inventory while still meeting your required customer service levels?
Recall that safety stock is extra stock kept on hand to reduce the risk of stockouts caused by uncertainties in supply and demand. For example, you may forecast that you’ll need 100 cases per week of a particular item over the next month. But you hold 550 cases in case of an unexpected demand spike. Those extra 150 cases are safety stock.
Customer service level is the percent of customer orders you can fulfill in a timely fashion. For example, a 98% service level means that you can deliver 98% of customer orders on time and in full. All things being equal, the higher your service level, the more safety stock you need to keep on hand.
With the traditional approach to inventory optimization, you would define the desired service level, and generate safety stock levels assuming a normal distribution of forecast error and supplier lead times. Then you cross your fingers.
And there’s the rub. You’ll have to wait at least the lead time in order to see how your new policy is working out — often up to a few months, at which point it’s too late to fix any problems.
This process often results in too little inventory, resulting in missed service level targets, or too much inventory, resulting in excessive carrying costs.
Additionally, most inventory policies are set independently by location, without taking advantage of the opportunity to reduce inventory at some locations by leveraging inventory at others.
MEIO emerged at the turn of the century as advances in computing power and optimization algorithms enabled enterprise software to finally handle the demands of modeling multiple stocking locations simultaneously. Prior to 2000, custom MEIO models were used for niche applications like service parts. Later, commercial solutions began gaining traction, promising significant inventory reductions while maintaining or improving service levels through coordinated, network-wide optimization.
2000s: Hype Followed by Disillusionment
First-generation packaged applications, notably those from SmartOps and Optiant, produced mixed results. While they achieved some successes, the solutions of the 2000s were overhyped. They relied on modeling assumptions that, while reasonable given the technology at the time, failed to adequately mirror the complexity of real-world supply chains.
Early MEIO lost favor due to challenging implementations that required levels of data accuracy, integration, and modeling sophistication that most organizations and the tools themselves were not yet equipped to support. Companies faced difficulties with scalability, coordination, and managing complexities across multiple supply chain tiers. These challenges led to disappointing outcomes and high operational costs. The anticipated benefits were often undermined by data silos, insufficient forecast accuracy, and poor integration with existing systems, making MEIO less effective than expected.
Optiant was acquired by Logility in 2010 for a fire-sale price of $3.3 million, and SmartOps was purchased by SAP in 2013 for a modest amount. These acquisitions ultimately served to round out the acquirers’ suites, with limited subsequent product evolution. Throughout the 2010s, MEIO experienced a period of disillusionment (in Gartner Hype Cycle parlance), becoming a niche tool.
2020s Revival
In the 2020s, we’ve seen a renewed interest in MEIO driven by new solutions which address the weaknesses of the earlier generation. Some of these improvements include:
- The use of AI and machine learning to overcome the modeling flaws of earlier products
- Native integration into SCP suites rather than standalone solutions or bolt-on acquired solutions
- Cloud-enabled scalability that allows moving from annual or quarterly target inventory resets to monthly and even weekly resets
New-generation MEIO solutions use more advanced models to determine optimal safety stock quantities at each supply chain node, aiming to reduce costs and improve service levels throughout the whole network. Multi-echelon optimization delivers system-wide efficiency by managing inventory as an interconnected whole.
As a result, on Gartner’s 2025 Hype Cycle for Supply Chain Planning Technologies, MEIO has moved from the Trough of Disillusionment to the Slope of Enlightenment. That means it’s classified as an early mainstream technology with 5% to 20% penetration of its target audience and growing.
Today’s Inventory Optimization Challenges
Real-world supply chains are still messy.
Calculating optimal safety stocks requires taking into account uncertainty on both the demand side and the supply side. The more uncertainty, the more safety stock you must hold to ensure you meet your target service levels.
Mathematically, this is accomplished by incorporating the variability of the demand forecast and the variability of supplier lead time (the time between placing a purchase order and receiving the order).
Most MEIO software applications still rely heavily on assumptions of normally distributed forecast error and supplier lead times — meaning if you plotted the values of these variables, they would form a bell curve around the average value.
In real-world supply chains, this often isn’t the case. And the company supply chain data used to set inventory levels is often plagued by poor quality. The result is garbage in, garbage out.
Companies usually don’t discover these problems until it’s too late, leading to missed service level targets for some SKUs and excess inventory for others.
New Horizon’s New Approach
New Horizon’s new-generation inventory optimization solution takes a different approach. It addresses long-standing flaws in MEIO to deliver accurate, data-driven inventory recommendations, moving your customer service to the next level while minimizing inventory.
See the flow chart below for a high level overview of New Horizon’s MEIO algorithm:
New Horizon MEIO Algorithm
Using historical demand data, the forecasting algorithm, supplier delivery lead time performance, your inventory policy, and your supply plan policy, New Horizon derives an inventory policy for each item and location combination.
Then, our machine learning algorithm tests the MEIO inventory settings on your own historical data, continuously adjusting the settings and fine-tuning safety stock recommendations to maximize the probability that service level targets will be achieved.
We don’t assume a normal distribution of forecast error and supplier lead times.
We also understand that demand uncertainty depends on how the demand forecast was generated. To produce more accurate safety stock recommendations, New Horizon tailors its calculations to the specific forecasting algorithms used to create the demand plan.
The benefits of our approach?
- It yields more accurate recommendations because it empirically tests policies against observed outcomes in your supply chain rather than assumed distributions.
- It simulates potential problems before they happen.
- It allows you to focus on those items most susceptible to problems.
- It automatically detects and alerts you to data quality problems. By improving the quality of input data, the solution improves the quality of the safety stock recommendations.
With New Horizon, it’s possible to reduce safety stock by over 30% while still meeting your service targets.
Improve Service and Minimize Inventory with MEIO from New Horizon
What does this mean for you?
Assuming an average inventory of $100M, a modest savings of 10% means a reduction of $10M in inventory. At a typical 20% carrying cost, that’s $2M per year in savings and improvement to your bottom line.
New Horizon’s optimization is tailored to your forecasting algorithms and supply planning parameters. It simulates what will likely happen in order to identify problems and issues before they occur. It identifies problematic items to focus on in your forecasting and procurement processes.
To Learn More
Ready to unlock the next level of inventory efficiency?
New Horizon’s new-generation MEIO solution empowers supply chain leaders to achieve service excellence while dramatically reducing inventory costs.
Learn how tailored optimization can transform your operations — talk to one of our experts today.