Safety stock and reorder point under uncertainty

Reorder point and safety stock under uncertain demand and lead time, cycle service level and fill rate, combining variances, and the periodic-review order-up-to level.

Drafted with Aria, reviewed by the AiCanCode.org team. Spotted an error? Use Give Feedback at the bottom of the page.

Why it matters

Demand is never exactly the forecast, and suppliers are not always on time. If an order is placed when stock just covers average lead-time demand, the plant runs out in roughly half of all cycles. Safety stock is the extra inventory that buys a chosen level of protection, and the reorder point tells the stores clerk exactly when to order. Setting them statistically avoids both line stoppages and shelves full of dead stock.

Key ideas

Continuous-review (Q, ROP) system. Inventory position (on hand + on order − backorders) is watched continuously; when it falls to the reorder point ROP, a fixed quantity Q (often the EOQ) is ordered. Uncertainty matters only during the lead time, because that is the only time stock cannot be replenished.

Safety stock (SS). ROP = expected lead-time demand + SS. SS is the average stock left when the replenishment arrives; it is carried all the time and costs SS × H per year.

Service level – two common definitions.

  • Cycle service level (CSL, type 1) – probability of not running out during a replenishment cycle. With normally distributed lead-time demand, SS = z·σ_L where z is the standard normal value for CSL. Standard values (from the normal table): 90 % → z = 1.28, 95 % → 1.645, 97.5 % → 1.96, 99 % → 2.33.
  • Fill rate (type 2) – fraction of demand met directly from stock. It uses the normal loss function (from tables) and usually gives a much higher percentage than CSL for the same SS, because even in a stock-out cycle most demand is filled. A "95 % service level" in a GATE question normally means cycle service level unless stated otherwise.

Sources of variability during lead time.

  • Demand varies, lead time fixed: σ_L = σ_d·√L (daily variances add over L independent days).
  • Lead time varies, demand constant: σ_L = d̄·σ_LT.
  • Both vary independently: σ_L = √(L̄·σ_d² + d̄²·σ_LT²). Note that standard deviations do not add over time – variances do. This is why a 4-week lead time needs only twice (√4) the safety stock of a 1-week lead time for the same daily demand variability.

Choosing the service level. Higher service needs disproportionately more stock (z rises steeply above 95 %). The economic service level balances holding cost of SS against shortage cost; when the cost per stock-out occasion is known, the optimal probability of stock-out per cycle is P(stock-out) = H·Q/(D·B_s), where B_s is the shortage cost per occasion.

Periodic-review (order-up-to) system. Stock is reviewed every T periods and an order is placed to bring the inventory position up to a target level. Protection is needed over the review period plus the lead time, so a periodic system needs more safety stock than a continuous one for the same service.

Other practical points. Reduce SS by cutting lead time and its variability, improving forecasts (smaller σ) and pooling stock in fewer locations. Safety stock levels for items in MRP are usually set only at end-item or purchased-part level, not at every BOM level.

Formulas

ROP = d̄·L + SS SS = z·σ_L

  • d̄ = mean demand rate (units/day or units/week), L = lead time (same time unit), z = standard normal deviate for the cycle service level (take from the normal table), σ_L = standard deviation of lead-time demand (units).

σ_L = σ_d·√L (demand variable, lead time constant) σ_L = d̄·σ_LT (lead time variable, demand constant) σ_L = √(L̄·σ_d² + d̄²·σ_LT²) (both variable, independent)

  • σ_d = standard deviation of demand per period (units/period), L̄ = mean lead time (periods), σ_LT = standard deviation of lead time (periods).

Annual cost of safety stock = SS·H

  • H = holding cost (₹/unit·year).

Order-up-to level S = d̄·(T + L) + z·σ_d·√(T + L) Order quantity = S − inventory position at review

  • T = review period (same time unit as L).

Worked examples

Example 1 (standard) – reorder point with variable demand. Mean demand 40 units/day, σ_d = 8 units/day, lead time 9 days (constant), cycle service level 95 % (z = 1.645 from the normal table).

  1. Lead-time demand = 40 × 9 = 360 units.
  2. σ_L = 8 × √9 = 24 units.
  3. SS = 1.645 × 24 = 39.5 units (keep 40).
  4. ROP = 360 + 39.5 ≈ 400 units. In about 5 % of cycles demand during the lead time will exceed 400 units.

Example 2 (GATE level) – demand and lead time both variable. Weekly demand: mean 200 units, σ_d = 30 units. Lead time: mean 4 weeks, σ_LT = 1 week. Cycle service level 97.5 % (z = 1.96). Holding cost ₹12/unit·year.

  1. Expected lead-time demand = 200 × 4 = 800 units.
  2. σ_L = √(4 × 30² + 200² × 1²) = √(3600 + 40,000) = √43,600 = 208.8 units.
  3. SS = 1.96 × 208.8 = 409.3 units.
  4. ROP = 800 + 409.3 ≈ 1209 units.
  5. Annual cost of SS = 409.3 × 12 = ₹4911.
  6. Insight: the lead-time term (40,000) is more than ten times the demand term (3600). Making the supplier reliable would cut safety stock far more than better forecasting.

Example 3 – periodic review. Same demand (200 ± 30 units/week), review every T = 2 weeks, constant lead time L = 1 week, z = 1.645. At a review the inventory position is 250 units.

  1. S = 200 × 3 + 1.645 × 30 × √3 = 600 + 85.5 = 685.5 units.
  2. Order = 685.5 − 250 ≈ 436 units.

Common mistakes

  • Multiplying σ_d by L instead of √L (adding standard deviations instead of variances).
  • Using the z for a two-sided interval (e.g. 1.96 for 95 %) instead of the one-sided value (1.645).
  • Mixing periods: σ per week with lead time in days.
  • Forgetting the d̄·L term and quoting SS as the reorder point, or vice versa.
  • Comparing ROP with on-hand stock when orders are outstanding; use inventory position.
  • Treating cycle service level and fill rate as the same thing.

For GATE PI

Expect reorder-point questions with normally distributed demand (z values usually given), the effect of lead-time variability, the probability of stock-out for a given ROP, and the annual cost of safety stock. Practise converting time units, combining variances, and reading z from the given service level both ways (z from probability and probability from z).

Quick check

  1. d̄ = 50 units/day, L = 4 days, SS = 30 units. What is the ROP?
  2. σ_d = 10 units/day and L = 16 days (constant). What is σ_L?
  3. A ROP equals mean lead-time demand plus 1.28σ_L. What is the cycle service level?
  4. Why does a periodic-review system need more safety stock than a continuous one?

Answers: 1. 230 units. 2. 10 × 4 = 40 units. 3. About 90 %. 4. It must cover uncertainty over the review period plus the lead time, not just the lead time.

Interview questions

All Production Planning and Operations Management interview questions

Try answering each one aloud before you open it.

  1. 1.What is safety stock and why is it important in production planning?Concept

    Safety stock is an additional quantity of inventory kept on hand to mitigate the risk of stockouts caused by uncertainties in demand and supply. It is important because it acts as a buffer against unexpected fluctuations, ensuring that production processes are not interrupted and customer demand is met consistently.

  2. 2.Define reorder point and explain its significance in inventory management.Concept

    The reorder point is the inventory level at which a new order should be placed to replenish stock before it runs out. It is significant because it helps maintain optimal inventory levels, preventing both stockouts and overstock situations, thus ensuring smooth operations and cost efficiency.

  3. 3.How do you calculate the reorder point under uncertainty?Concept

    The reorder point under uncertainty is calculated using the formula: Reorder Point = (Average Demand during Lead Time) + Safety Stock. This formula accounts for both the expected demand during the lead time and the additional safety stock to cover demand variability.

  4. 4.Explain how safety stock levels are determined in a production environment.Concept

    Safety stock levels are determined by analyzing demand variability, lead time variability, and the desired service level. Statistical methods, such as calculating the standard deviation of demand and lead time, are often used to quantify these uncertainties and set appropriate safety stock levels.

  5. 5.Why is it important to consider lead time variability when calculating safety stock?Application

    Lead time variability can significantly impact inventory levels because it affects the timing of replenishment orders. If lead times are longer or more variable than expected, it can lead to stockouts. Therefore, considering lead time variability is crucial to ensure that safety stock levels are sufficient to cover these uncertainties.

  6. 6.What happens if safety stock is set too high or too low?Application

    If safety stock is set too high, it can lead to excessive inventory holding costs and potential obsolescence. Conversely, if it is set too low, it increases the risk of stockouts, which can disrupt production and lead to lost sales. Therefore, finding the right balance is essential for cost-effective and efficient operations.

  7. 7.How does demand variability affect the calculation of safety stock?Application

    Demand variability affects safety stock calculations by determining the level of uncertainty in demand forecasts. Higher demand variability requires more safety stock to buffer against potential stockouts, while lower variability allows for reduced safety stock levels, optimizing inventory costs.

  8. 8.A company experiences an average demand of 100 units per week with a lead time of 2 weeks. If the standard deviation of demand is 10 units, calculate the reorder point assuming a safety stock of 20 units.Numerical

    Reorder Point = (Average Demand during Lead Time) + Safety Stock = (100 units/week * 2 weeks) + 20 units = 200 units + 20 units = 220 units.

  9. 9.If a company wants to maintain a 95 % service level, how does this affect the safety stock calculation?Application

    For a 95 % cycle service level the safety stock is z·σ_L with the one-sided z = 1.645 from the normal table, so SS = 1.645 × the standard deviation of lead-time demand. This means a stock-out is expected in about 5 % of replenishment cycles, not on 5 % of days or 5 % of demand. Moving from 95 % to 99 % (z = 2.33) raises safety stock by about 42 %, so very high service levels become expensive.

  10. 10.A product has a lead time of 3 weeks and an average weekly demand of 50 units. If the standard deviation of weekly demand is 5 units, calculate the safety stock required for a 90 % cycle service level.Numerical

    Standard deviation of lead-time demand σ_L = σ_d × √L = 5 × √3 = 8.66 units. For 90 % the one-sided z from the normal table is 1.28, so SS = 1.28 × 8.66 = 11.1 units, rounded up to 12 units. The reorder point would then be 50 × 3 + 12 = 162 units.

Finished this topic? Mark it so your progress, study plan and readiness keep up.

Stuck on something here?

PreviousEPQ, quantity discounts and backordersNext ABC and VED analysis