475. Supply Chain Analytics

475.1. Inventory Turns

Inventory Turns: How many times inventory is turned over per year

Inventory Turns = Cost of Goods Sold / Average Inventory Value

Days of Supply: Average days the item is held in inventory

Days of Supply = Number of Days in Period / Inventory Turns

Efficiency Benchmark

Supply chain strategies

• Push: Forecast-driven, build inventory based on demand forecasts
• Pull: Customer-order-driven, only produce when there is a demand

Factories -Shipping Inbound-> Distribution Centers -Shipping Outbound> Retail Stores <- Customers

Factory -> DCs -Direct Ship-> Customer

Question: Use stores or direct shipping?

Shipping v. Inventory Trade-off

1. Classical (Store) Model

DCs -Truck Load (TL) -> Stores (Large Inventory) <- Customers

Transport v. Inventory Trade-off

Save shipping cost by TL but require large inventory at stores

Low Value, Long Shelf Life, High Volume Items

2. Continuous Replenishment Model

DCs -Less Than Truck Load (LTL) -> Stores (Small Inventory) <- Customers

Increase Shipping Cost but Reduce Inventory Cost at Stores

3. Direct Shipping Model (Showroom)

Factory -> DCs -Direct Ship-> Customer

No Stores, Direct Shipping

Almost Zero Inventory, High Shipping Cost

High Value, Slow Moving Items

Inventory Risk Pooling Effect

Consolidate inventory into fewer locations, and total safety stock drops — even while serving the same demand

Simple Example Say you sell umbrellas in 2 cities. Each city has:

Average weekly demand: 100 units Std deviation: 50 units You hold safety stock (95% service level)

Decentralized (2 separate warehouses):

Safety stock per city = 1.65 × 50 = 82.5 units Total safety stock = 82.5 × 2 = 165 units

Centralized (1 warehouse serving both cities):

Combined std deviation = √(50² + 50²) = √5000 ≈ 70.7 units Safety stock = 1.65 × 70.7 ≈ 117 units

Savings: 165 → 117 units ( 29% less inventory)

475.1.1. Push Pull Strategies

Push — you push product into the supply chain speculatively. You bet on your forecast being right. Pull — you wait for a customer order, then pull production/replenishment. No forecast needed, but you need fast response.

The Push-Pull Boundary Most real supply chains are hybrid. The boundary (called the decoupling point) is where you switch from push to pull. Example — car manufacturing:

Steel → stamping → painting → [decoupling point] → final assembly per order Everything upstream is pushed (economies of scale); downstream is pulled (customer config)

Tradeoff:

• Shipping, picking / packing
• Customer wait time

vs.

• Inventory investment
• New product introduction

Categorize Products

Data Collection

• Inventory Holding Cost
• Shipping Cost
• Picking / Packing Cost
• Sales Volume (by product)
• Inventory Levels (by product)

Calculate:

• Inventory Cost
• Shipping Cost
• Warehouse Cost

Under Push / Pull by Product

1. Inventory Cost

Estimating Inventory Cost Rates (Per Week Per Unit) At Stores

Capital cost per week per unit + Depreciation per week per unit

Inventory cost per week per unit =

E.g.,

• Smartphones

  • Product Value: $500
  • Liquidation value: 0%
  • Capital return rate: 8%
  • Inventory cost / unit: $20
  • Life cycle: 26 weeks

• Feature Phone

  • Product Value: $200
  • Liquidation value: 0%
  • Capital return rate: 8%
  • Inventory cost / unit: $8

2. Picking / Packing Cost

• Push: Weekly

Cost of N units = Cost of 1st pick/pack cost + subsequent pick/pack cost * (N-1)

• Pull: Unit

1st pick/pack cost * N

475.1.2. Inventory cost rates (per week per unit)

Inventory cost per week (per unit) = Capital cost per week (per unit) + Depreciation per week (per unit)

Capital cost per week (per unit) = Product value * Annual capital return rate / 52

Depreciation per week (per unit) = (Product value - Liquidation value) / Product life-cycle (in weeks)

Example 1: Smart phones

The value of a smart phone is $500 on average. Suppose the annual capital return rate is 8% (if the $500 is invested elsewhere, one may get a return of 8% annually), the liquidation value at the end of the product life-cycle is $0, and the product life-cycle is 26 weeks (half a year). Then for smart phones,

Inventory holding cost per week (per unit) = $500 * 8% / 52 + ($500 - $0) / 26 = $20.00 / unit.

Example 2: Feature phones

The value of a feature phone is $200 on average. Suppose the annual capital return rate is 8%, the liquidation value at the end of the product life-cycle is $0, and the product life-cycle is 26 weeks (half a year). Then for features phones,

Inventory holding cost per week (per unit) = $200 * 8% / 52 + ($200 - $0) / 26 = $8.00 / unit.

475.1.3. Shipping cost rates (per unit)

Overnight express shipping rate (per unit): $12.00 / unit (FedEx benchmark)

Standard 2-day shipping with batch discount:

The 2-day standard shipping cost is usually much cheaper than overnight express, and batch shipping (shipping more than 10 units at a time to the same destination) usually enjoy a discount depending on the carriers and distance.

Using FedEx as a benchmark, the ratio between overnight and 2-day shipping rates is 2.5 / 1, and shipping batch can enjoy a 50% discount per unit. Then the unit shipping cost for 2-day batch shipping is

$12 / 2.5 * 50% = $2.40 / unit.

475.1.4. Warehouse picking / packing cost rates

To pick / pack N (>1) identical products in one order:

1st pick / pack cost = $1.00.

subsequent pick / pack cost = $0.10