How to Build an Inventory Replenishment Plan: EOQ, Safety Stock and Reorder Points

A stockout costs more than the lost sale. It costs customer trust, and in industries where competitors are one click away, that trust does not always come back. The primary job of an inventory replenishment plan is to prevent that outcome by ensuring the right products are available at the right time, without tying up excess capital in stock that is not moving.

This guide covers the key concepts, the three core formulas every replenishment plan needs, four replenishment strategies and when to use each, a step-by-step plan structure, and the KPIs to track once the plan is running.

Before building the mechanics of a replenishment plan, you need to choose which strategy governs how and when orders are triggered. The four main options differ in what triggers a reorder and how the order quantity is set.

Most businesses use a mixed approach: Fixed Order Quantity or Min-Max for class A and B items where control matters, and Fixed Order Period or JIT for lower-value class C items. The choice of strategy determines how the reorder point and safety stock are used in practice.

EOQ determines how much to order each time you replenish. It minimises total inventory cost by balancing the cost of placing orders against the cost of holding stock.

D = Annual demand in units

S = Cost of placing one order (admin, freight, receiving)

H = Annual holding cost per unit (storage, insurance, capital; typically, 20-30% of unit cost)

Worked example: A pharmacy stocks a pain reliever with annual demand of 12,000 units. Order cost is £50 per order. Holding cost is £2 per unit per year.

EOQ = square root of (2 x 12,000 x 50 / 2) = square root of 600,000 = approximately 775 units

At this quantity, the pharmacy places roughly 15 orders per year. Ordering more frequently increases admin cost. Ordering larger quantities increases holding cost. EOQ finds the point where those two forces are equal.

Limitations: EOQ assumes stable annual demand, a fixed cost per order, and a constant holding cost. Adjust for seasonal patterns, supplier quantity discounts, and minimum order quantities in practice.

The reorder point tells you when to trigger an order, so that stock arrives before the shelf empties.

Worked example: The pharmacy sells an average of 33 units per day. Supplier lead time is 7 days. Safety stock is 100 units.

ROP = (33 x 7) + 100 = 231 + 100 = 331 units

When stock falls to 331 units, the replenishment order is placed. The order arrives in 7 days, consuming approximately 231 units, leaving 100 units of safety stock as the buffer.

Safety stock is the buffer that absorbs demand spikes and supply delays. The right level depends on how much variability you face and the service level you want to maintain.

Basic method (safety days): Safety stock = Average daily demand x Safety days

If average daily demand is 33 units and you want 5 days of buffer, safety stock = 33 x 5 = 165 units. This method is simple and widely used. It is appropriate for class B and C items.

Service-level method (Z-score): Safety stock = Z x Standard deviation of demand during lead time

Z is the statistical score for your target service level: 1.65 for 95%, 1.96 for 97.5%, 2.05 for 98%. This method accounts for actual demand variability and is more appropriate for class A items where the cost of a stockout is highest.

ABC analysis segments your SKU portfolio using the Pareto principle: a small number of SKUs drive the majority of revenue and deserve proportionally more planning attention and safety stock.

Sort your full SKU list by annual revenue contribution, highest to lowest. Calculate cumulative revenue as a percentage. SKUs in the top 80% of cumulative revenue are class A. The next 15% are class B. The remaining 5% are class C. Apply safety day targets and review frequencies accordingly.

Review the classification at least annually. SKUs migrate between tiers as demand patterns change, new products launch, and others reach end of life.

Read: Annual Recurring Revenue vs Revenue: How Each Metric Impacts Financial Forecasts

Gather data for each SKU and distribution location. The data should include current stock level, location, and status (available, reserved, or damaged). This becomes the baseline for all calculations that follow.

Include quantities already ordered and in transit. Without GIT data, your available inventory calculation will understate your true inventory position and may trigger unnecessary replenishment orders.

Historical sales data by SKU, channel, and time period is the input to your demand forecast. Organise it in the same system or structure as your inventory data so the two can be compared directly. Identify seasonal patterns, promotional uplifts, and any one-off events that distorted the data and should be excluded from the baseline.

Apply a demand forecasting method appropriate to your data. Moving averages are suitable for stable demand with no strong trend. Exponential smoothing gives more weight to recent data and responds faster to trend changes. Statistical models are appropriate when you have sufficient historical data and want to capture seasonality explicitly.

For new products without sales history, use analogues: similar products launched in comparable markets, adjusted for expected differences in price point, distribution reach, and promotional support.

Based on the demand forecast, calculate when each SKU will reach zero stock at current inventory levels.

Worked example: An electronics retailer holds 200 units of a headphone model and sells an average of 20 units per day. Days of cover = 200 / 20 = 10 days. If lead time is 7 days, the reorder point is reached in 3 days (10 – 7 = 3).

Use the inventory level formula to project stock positions across future periods, incorporating GIT and the estimated stock-out pattern.

Worked example: A furniture store holds 200 units of a chair model. 100 units are in transit. Forecasted demand over the next month is 150 units. Future inventory = 200 + 100 – 150 = 150 units.

Apply the safety stock calculation to each SKU, using the method appropriate to its class. Use the service-level (Z-score) method for class A items. Use the safety days method for class B and C items.

For a class A item with average daily demand of 50 units, a lead time standard deviation of 2 days, and a target service level of 95%, safety stock = 1.65 x (50 x 2) = 165 units.

For a class C item with average daily demand of 5 units and 2 safety days, safety stock = 5 x 2 = 10 units.

With safety stock defined and future inventory levels projected, calculate the Stock-In quantity needed for each SKU.

Worked example: A laptop retailer wants to hold 100 units of a popular model to cover next month’s demand, plus 30 units of safety stock (target = 130 units). Current inventory is 50 units with no GIT. Stock-In = 130 – 50 = 80 units to order.

Manual replenishment planning at scale is error-prone and slow. The replenishment trigger, the order quantity calculation, and the plan update should all be automated where possible.

Automation options range from simple reorder alerts in an ERP system to full demand-driven replenishment where the system calculates order quantities and raises purchase orders automatically when the ROP is reached. The appropriate level of automation depends on your SKU count, the variability of your demand, and the reliability of your supply base.

Three KPIs give a complete picture of replenishment plan health.

Compare planned versus actual performance at the SKU level each review period. Identify root causes: was the variance driven by a demand spike, a lead time delay, a forecast error, or a data quality issue? Variances that recur without a clear cause point to a structural problem in the model that needs fixing, not a one-off adjustment.

Update safety stock levels, reorder points, and order quantities when the underlying drivers change. Safety stock that was calibrated for last year’s lead time may be too low if a supplier has lengthened their fulfilment window. Reorder points set for a stable demand product need revisiting when a promotional campaign creates a temporary demand surge.

Sales teams know which products are experiencing customer pressure before the data shows it. Warehouse teams see receiving patterns and supplier reliability issues firsthand. Suppliers provide early warning of lead time changes. Structured feedback from these three groups closes the gap between the plan and what is actually happening in the supply chain.

A replenishment plan that works is built on three things: accurate formulas (EOQ, ROP, safety stock), a strategy that fits your operating environment, and KPIs that tell you quickly when the plan is drifting from reality.

The biggest gains usually come not from sophisticated modeling but from fixing the basics: clean data, correct safety stock by product class, and a consistent review cycle that updates the plan before a stockout occurs rather than after.

Start with your class A items. Classify your top 20% of SKUs by revenue, set reorder points and safety stock for each, implement EOQ-based order quantities, and track fill rate weekly. From that foundation, extend the same rigour to class B and C items over time.