Multi-Echelon Inventory Optimization (MEIO) for Branch Networks

Multi-echelon inventory optimization (MEIO) sets the stock of a whole network at once: the central store and every branch it feeds. It works because one shared buffer at the center covers the branches' bad days with less stock than a buffer in every branch. For a business with branches, it decides how much stock to hold back.

This page works that idea through for a central store in Kampala and four branches, in Uganda shillings. It then shows when MEIO pays, and when a reorder point per branch is enough.

What Is Multi-Echelon Inventory Optimization (MEIO)?

An echelon is one level of a supply chain where stock waits. MEIO plans the stock at every level together, because the stock at one level protects the level below it.

Single-echelon planning does the opposite. In Wikipedia's summary, it "forecasts demand and determines required inventory for each echelon separately". MEIO starts from the rule the same page states: "The better the service that is provided upstream, the smaller the protection that is needed downstream."

Each level is a warehouse in its own right, with its own stock and its own wait for new stock. MEIO decides how much safety stock each one holds.

MEIO vs. Traditional Inventory Optimization

Traditional inventory optimization sets each location's reorder point and safety stock as if it stood alone. MEIO sets them for the network, so the central store's stock counts as protection for every branch. What inventory optimization is covers the single-location version.

Why a network that plans as one needs less stock

Two effects shrink the total. Swings in demand partly cancel out when branches share one buffer, and each branch waits days for a van instead of weeks for a ship.

Pooling: one shared buffer is smaller than four separate ones

Branches rarely have their worst day on the same day. When four branches draw on one buffer, a slow day in Jinja offsets a rush in Gulu.

The combined swing grows with the square root of the number of branches, not with the number itself. Four branches with equal, independent swings need a pooled buffer twice one branch's, not four times. Gary Eppen showed this square-root effect for identical, uncorrelated demand in a 1979 paper in Management Science.

Each branch only covers the days until the next van

A branch's buffer must cover the time it waits for new stock, its risk period. Restocked by the supplier, that wait is the import's full lead time. Restocked from the central store, it is a few days.

Safety stock grows with the square root of the risk period. Cutting a branch's wait from 49 days to 4 cuts its buffer by 71%, since √4 ÷ √49 = 2 ÷ 7. The formula itself is worked step by step in how to avoid overstocking and stockouts.

Held-back stock can go to whichever branch sells it

A container split between branches on arrival is committed before anyone knows where it will sell. Stock held at the center goes out on the next van to the branch that actually runs low.

It also ends a costly habit: moving stock back from a slow branch to a fast one, paying for the road twice.

The network's total is easier to forecast than each branch

The same pooling applies to forecasts. The central store forecasts the whole network's sales, whose relative error is smaller than any single branch's.

Each branch then forecasts only a few days ahead, the time until its next van. Inventory forecasting methods covers the formulas.

Where the saving shrinks

The square-root effect assumes the branches' swings are independent. School fees week, Christmas and the harvest move every branch at once, and then pooling saves little.

It also assumes the central store rarely runs out itself. MEIO software models those stockouts, which delay every branch behind the store. A spreadsheet version should keep the central buffer generous.

For example: a central store in Kampala and four branches

Take a phone-accessories business with a central store in Kampala, and branches in Kampala, Jinja, Mbarara and Gulu. It imports chargers through Mombasa. The figures are an illustration, not a real business's.

The inputs, with the same availability target for both options

Both options below aim at the same availability: each branch has stock in 95% of its restocking cycles. The only difference is where the safety stock waits.

The example leaves out the days between import orders to keep the sums short. Adding them raises both options' buffers.

Option A: split every shipment between the branches on arrival

Each branch waits for the next import, so each one covers 49 days of swings on its own.

Per branch: 1.65 × 10 × √49 = 1.65 × 10 × 7 = 115.5, so 116 chargers. For four branches: 464 chargers, worth UGX 6,960,000.

Option B: hold the buffer back in the central store

The central store waits for the import, but it covers the four branches' combined swings. Those grow with the square root of four: 10 × √4 = 20 chargers a day.

Central store: 1.65 × 20 × 7 = 231 chargers. Each branch waits only for the van, which leaves every 3 days and takes 1 day, so 4 days in all.

Per branch: 1.65 × 10 × √4 = 33 chargers, and 132 for four. Option B holds 231 + 132 = 363 chargers, worth UGX 5,445,000.

What holding back saves

Option B needs 101 fewer chargers than option A, 21.8% less safety stock for the same availability. At UGX 15,000 a charger, that frees UGX 1,515,000 for this one product.

The saving repeats for every imported product with similar numbers, and the cash it frees stops costing you rent, interest and spoilage.

The van timetable decides the saving

The branch buffers depend on how often the van runs. The same calculation with three timetables shows the saving falling from 29.5% to 7.1%.

Each branch figure is 1.65 × 10 × √(days the branch waits), rounded up to a whole charger. A weekly van keeps most of the old buffer in the branches, so holding back barely pays. Before you hold stock back, make the van run more often.

When a business with branches needs MEIO, and when it doesn't

A business needs MEIO when a central store feeds its branches and the supplier takes weeks. Without both, a reorder point per branch does the job.

Six signs point to the third or fourth row.

1. Each branch holds weeks of safety stock

If every branch keeps a buffer sized for the import's lead time, the network holds option A's stock. In the example, that is 101 chargers more than it needs with a van every 3 days.

2. Frequent stockouts at one branch while another has plenty

A branch that runs out while another holds the same product has an allocation problem, not a buying problem. Holding stock back fixes allocation; buying more only adds to the slow branch's pile.

3. Dead stock piles up in the slow branches

Stock pushed to a branch that doesn't sell it ages there. If the same product is slow in one branch and fast in another, the split on arrival is the cause. Dead stock covers how to clear what is already stuck.

4. Nobody can see every branch's stock in one place

This sign is a prerequisite, not a reason. MEIO needs each location's stock and sales in one system, and multi-location retail management covers getting there.

5. A central store or warehouse feeds the branches

Without a central level there is nothing to hold back. With one, the choice in the example applies to every product it receives. An online shop shipping from two warehouses faces the same choice (e-commerce inventory management).

6. Supplier lead times run to weeks, and vary

Long lead times are where pooling saves most, because the central buffer covers the long wait once instead of four times. Imports through Mombasa or Dar es Salaam qualify, and so does a factory that allocates stock once a month.

What makes MEIO hard for a growing chain

The math is the easy part. The hard part is records you can trust at every location, and branch managers who accept that they no longer order for themselves.

Managing Data Across Multiple Locations

Every sale, transfer and write-off must reach the same system the day it happens. One branch on paper breaks the numbers for the whole network.

Coordinating with Suppliers

The central buffer depends on how long imports really take, and how much that varies. Record the order date and the arrival date of every purchase, the same way you would measure order cycle time.

Integrating New Technology with Legacy Systems

MEIO software reads from an inventory system or an ERP. Branches on different tills, or on exercise books, give it nothing to read. The first project is one system for every location.

Maintaining Data Security

A system shared by every branch shows every branch's stock and sales. Give each branch manager their own branch only, and keep supplier costs to the people who buy. In ArmPOS, Access warehouses on each user does this (Roles and permissions).

Change Management

Branch managers used to ordering directly lose that right, and the central store must keep to the van timetable. Agree the timetable and how branches ask for stock before the first container is held back.

How to Get Started With MEIO

Start with a spreadsheet and your ten to twenty most important imported products. Software comes later, if the spreadsheet proves the saving.

Review Your Current Inventory Data

Count every location, then pull 8 to 12 weeks of daily sales per product and branch. Note how long each of the last few imports took, from order to arrival.

Measure each branch's daily sales and their swings

For each product and branch, work out the average daily sales and the standard deviation. In Excel or Google Sheets, =AVERAGE() and =STDEV() over the daily column give both.

The standard deviation is the "daily swing" in the example. It is the number that pooling reduces.

Set Clear Inventory Goals

Choose an availability target per product group: 98% for the products customers come in for, 95% for most, 90% for the rest. The safety factors are 2.05, 1.65 and 1.28. The 80/20 rule in inventory sorts your products into those groups.

Choose the Right Tools

A spreadsheet handles a few dozen products across a handful of locations. Dedicated MEIO software earns its cost with hundreds of products, several levels and data clean enough to feed it.

Connect With Suppliers and Fulfillment Partners

Fix two timetables: how often you order from the supplier, and how often the van runs to each branch. Both enter the formula as days of waiting.

Start Small, Then Scale

Hold back stock for your top imported products first, and leave everything else as it is. After one import cycle, compare their stockouts and stock levels with the products you left alone.

Monitor and Adjust

Recalculate every quarter, and after any change in the van timetable or the supplier's lead time. Lower the central buffer only when the records show it rarely runs out.

How to do it in ArmPOS

ArmPOS keeps the records MEIO needs: each location's stock, sales and transfers in one system. It doesn't calculate safety stock or move stock by itself. You set the levels in your spreadsheet; it records, alerts and moves stock when someone approves.

Set up the central store and each branch as a warehouse

In Settings › Warehouse, click Create for each location. Each warehouse keeps its own stock, and your plan sets how many you can have (Your warehouse and its locations, plans).

Count each location before you trust its figures

Products › Count Stock makes a counting sheet for one warehouse, which you download as Excel. Enter the differences as an adjustment in that warehouse (Count and adjust stock).

Export each branch's daily sales

The Product sales report lists each product line with its Date, Warehouse and Qty sold, and EXCEL downloads it (Sales reports). In the spreadsheet, total each day per branch, then work out the average and the standard deviation.

Set each product's Stock Alert to the branch level

A product has one Stock Alert, and ArmPOS checks it in every warehouse. Set it to the branch reorder point. When a branch falls to it, the bell warns the people who work there.

Push and email warnings reach each person who turns them on (Get alerts on your phone). Keep the central store's much higher reorder point in your spreadsheet. Check it weekly against the Stock Report filtered to the central store (Stock reports).

Refill the branches with approved transfers

Transfer › Create Transfer moves stock from the central store to a branch, and nothing moves until someone approves it. Choose Sent while the van is on the road, then edit it to Completed when it arrives (Move stock between warehouses).

The steps are the same in every country ArmPOS serves, including Uganda.

Frequently asked questions

What is an echelon in inventory?

An echelon is one level of a supply chain where stock is held, such as a supplier, a central store or a branch. A business with a central store and branches runs two echelons of its own.

Does MEIO improve customer service?

It lets you choose: the same availability with less stock, or better availability with the same stock. In the example, holding back keeps every branch at 95% availability with 101 fewer chargers.

Does MEIO need special software?

Not to start. The hold-back rule works in a spreadsheet for a few dozen products. Software pays once you plan hundreds of products across several levels.

Should a small business with branches use MEIO?

Use it if a central store feeds your branches and your supplier takes weeks: hold the buffer back and run the van often. With branches that buy locally within a day or two, a reorder point per branch is enough.

Work out option A and option B for your three biggest imported products, with your own van timetable. To keep every branch's stock, sales and transfers in one record, create a free ArmPOS account.