Supply Chain Management Week 3 Balancing Supply Demand

Supply Chain Management Week 3 Balancing Supply & Demand - Inventory

1. Why we have inventory 2. Inventory Carrying Cost 3. Inventory Classification 4. Economic Order Quantity 5. Safety Stock 7. Risk Pooling 2

1. Why we have inventory 2. Inventory Carrying Cost 3. Inventory Classification 4. Economic Order Quantity 5. Safety Stock 7. Risk Pooling 3

Why do we have inventory? 1. Economies of Scale 2. Uncertainty 3. Time 4. Seasons 5. Anticipation 4

Why do we have inventory? Order Size Inventory Economies of Scale = EOQ its subsequent cycle inventory = cycle inventory is ½ of the order size assuming it sell at a consistent rate until the next order arrives Uncertainty = Safety Stock to cover uncertainty of demand during lead time and uncertainty of supply (will supplier be on time or late; will he supplier ship all that I ordered? ) Time = Lead Time = in-transit inventory = Re-order point = number of days of lead time X the average daily demand Seasonality not shown here is closely tied to a season. Either no stock in the off season, or less stock in the off season as in the season. Anticipation = stocking up expecting a shortage or price increases Re-order Point Cycle Inventory Safety Stock Time 5

Where is the Inventory cost? The value of the inventory is on the balance sheet, but what about my costs to hold the inventory and my cost to acquire the inventory. The inventory carrying cost is comprised on the firms’ weighted average cost of capital as well as several non-capital expense accounts. The non-capital expense items include: Inventory taxes: The amount of taxes paid to state and local governments as tax upon the value of the inventory. Insurance: Premiums paid to insure the inventory. Warehousing: The fixed space expenses of owning and operating a distribution center. This includes lease, utilities, insurance, janitorial, and property taxes. Administrative and Inventory control: This is the administrative expense of maintaining accurate inventory records. The cost of annual, quarterly, or on going cycle count activities is included. Obsolescence: Is the value of the inventory disposed of during a year due to being idle or obsolete. This tends to be higher for firms with short product life cycles due to newly developing technology such as personal computers and telecommunications. Shrinkage: Is the amount of inventory written off because it cannot be found or accounted for. This is normally due to theft.

Which inventory carrying cost is correct? A B C D Cost of money 3 10% 15% 20% Inventory tax 1% 1% Insurance 1% 1% Warehousing 2% 2% Administrative / Inv Control 1% 1% Obsolescence 4% 4% Shrinkage / Pilferage 1% 1% Total Inventory Carrying Cost 13% 20% 25% 30% Answer 7

B, C, D are correct. And A is the only one that is incorrrect. A is based upon a firm who has great credit and always borrows at or below prime rate. They are not considering their weighted average cost of capital which for public firms ranges from 10% to 15%. This includes a mix of debt 30% and equity 70%. Next page for more a table and then discussion afterwards

Weighted Average Cost of Capital = Min Cost of Equity Average Max 7% 11% 15% 70% 70% 4. 9% 7. 7% 10. 5% Borrowing Rate 7. 0% Marginal Tax rate 40. 0% 1 - marginal tax rate 60. 0% After Tax cost of debt 4. 2% 30% 30% 1. 26% 6. 2% 9. 0% 11. 8% Marginal Tax Rate 40. 0% 1 - Marginal Tax Rate 60. 0% 15% 20% Equity / Total Cost of Capital Debt / Total Cost of Capital After Tax Weighted Average Cost of Capital Before Tax Weighted Average Cost of Capital Discusssion 9

The cost of capital should include the cost of debt as well as the cost of equity. The cost of equity includes dividends paid to shareholders plus the growth in stock price. The cost of equity represents the opportunity cost of being able to invest money back into the firm for other purposes other than holding inventory. The weighted average cost of capital calculation for most manufacturing and distribution companies is 70% equity and 30% debt. A study of firms over last 5 years showed the cost of equity ranging from 7% to 15% depending upon the firm’s operating risk. The average cost of equity was 11%. It should be noted that the cost of equity an after-tax cost. The cost of debt is typically the prime rate or the borrowing rate of the firm. Stated borrowing rates are before tax rates, and should be adjusted to an after tax rate to adjust the cost of debt to recognize that interest is tax deductible. For example a 7% borrowing rate is 4. 2% after realizing the tax deduction at a 40% marginal tax rate (7% X (1 -40%) = 4. 2%). Calculating a weighted average cost of capital assuming that 70% is cost of equity and 30% is cost of debt yields a weighted average cost of capital of 9% (. 7 X 11%) + (. 3 x 4. 2%) = 9% This 9% weighted average cost of capital is an after-tax cost, and needs to be adjusted to a before tax cost so that it can be used with the non-capital inventory carrying cost items which are before tax costs. This is also needed because analysis and decision making of reducing or increasing inventory is traded off against other before tax expenses; such as transportation. Converting the after tax weighted average e cost of capital to a before tax weighted average cost of capital is performed by multiply the after tax cost times one minus the marginal tax rate. In this case the before tax weighted average cost of capital is 15%. (9% X (1 -40%) = 15%

Classifying inventory leads to inventory strategy We can also look at how to classify our inventory by comparing the value to the risk. Here are managerial implication of this type of classification. Value is measured as the value contribution to profit (margin) Risk is the negative impact of not having the product available when it is needed Distinctives High Safety Stocks More than 1 stock location Produce to Inventory Criticals High Safety Stocks Multiple Stocking locations Produce to Inventory Generics Low/No Safety Stock Produce to Order Commodities Adequate Safety Stock More than 1 stock location Produce to Inventory or Produce to Order Risk High Low Value High 11

THEN 80% Of Sales 40% Of SKU’s 20% Of SKU’s 90% of Sales 5% of Sales 15% Of Sales NOW 3% of Sales 50% Of SKU’s 7% Of Sales 40% Of SKU’s 10% Of SKU’s In 1906, Italian economist Vilfredo Pareto created a mathematical formula to describe the unequal distribution of wealth in his country, observing that twenty percent of the people owned eighty percent of the wealth. In the late 1940 s, Dr. Joseph M. Juran applied this to make the 80/20. ABC analysis has been an effective tool at categorizing fast and slow moving products. Traditionally 80% of a firm’s sales come from 20% of it’s products. These are desinated as A’s. The next 15% of sales comes from the next 40% of it’s products designated as B’s. The last 40% of the products are required to deliver the last 5% of sales are designated as C’s. ABC analysis has been an effective tool at categorizing fast and slow moving products. Traditionally 80% of a firm’s sales come from 20% of it’s products, the next 15% of sales comes from the next 40% of it’s products, and 40% of the products are required to deliver the last 5% of sales. This phenomenon is becoming more skewed with companies reporting 90% of sales from 10% of their products, the next 40 % of items delivering the next 7% of sales, and over 50% of their products being required to deliver the last 3% of sales. It is in this lengthening tail of the ABC curve that the complications of product proliferation lay. What are the implications. In warehousing put all the slow movers in back and fast movers in front. In sourcing, higher volume products may be outsourced while slower products needs closer more reactive sources of supply. I may have high volume in all my dc’s but consolidate the slow movers into one plant. I may even move slow movers to make-to-order while leaving fast movers as make to stock.

Volume AND Variability Volume Variability Profile Variability (Standard Deviation) 8. 0 Discussion 7. 0 6. 0 5. 0 4. 0 3. 0 2. 0 1. 0 0 100 200 500 1000 2000 Average Weekly Volume 3000 10000 13

In addition to looking at classifying inventory by volume, you can also look at it by volume and variability. Higher volume products tend to have less variability make sales easier to predict. Lower volume products tend to have higher variability. SKU stratification logic: A SKUs (green) High volume SKUs Stable order rate/low variability = easy to plan Longer term lifecycle means little risk of obsolescence B SKUs (Yellow) Medium volume SKUs Order rate low / high variability = harder to plan Longer term lifecycle means little risk of obsolescence C SKUs (Red) All low volume SKUs regardless of lifecycle Order rate low/high variability = hard to plan Short to medium term lifecycle SKUs have higher risk of obsolescent D SKUs (Blue) Any volume SKU with very short (e. g. primarily launch) lifecycles Any high volume SKU with high variability

The goal is to match the manufacturing & distribution strategy to the volume and variability of each SKU Discussion 8. 0 7. 0 6. 0 5. 0 “A”s - Build to Stock 4. 0 3. 0 2. 0 1. 0 0 100 200 500 1000 2000 3000 10000 “B”s – Build to Order “C”s - Make to Order “D”s – Make to Order 15

Category Color Operation Strategy A Green Manufacturing Assemble Line or Automation; Build to Forecast; Make to Stock; Rate Based & Level Loading A Green Distribution Fill from Finished Goods Stock; Rate based replenishment trigger (Re-Order Point) B Yellow Manufacturing Build to Order; Assembly Line or Cell; Component Inventory on hand B Yellow Distribution Fill from Finished Goods Stock; Kanban or Min-Max C Red Manufacturing Make to Order; Cellular Manufacturing; On Demand Manufacturing if available C Red Distribution Non Stocking SKU; Make to Order from manufacturing D Blue Manufacturing Make to Order; Assembly Line or Automation D Blue Distribution Finished Goods Inventory for maximum order quantity projections only.

MAQ Corporation Exercise 17

MAQ Corporation, a major producer of consumer electronics equipment, is currently faced with a rapidly growing product line and its associated inventory problems. MAQ’s President, Mary Semerod, has decided to initiate a program to analyze the company’s inventory requirements utilizing different inventory techniques. The first phase of the program consists of an ABC analysis of the companies product line (shown on following slide). Ms. Semerod has encountered difficulties in deciding on the appropriate criteria to use in the classification and in developing appropriate cutoff levels for each class of inventory. To solve her dilemma, Ms. Semerod has contracted the services of a logistics consulting firm to perform the inventory analysis 18

MAQ Corporation 1. If you were employed by the consulting firm, how would you construct your method of analysis? 2. What criteria would you use? 3. What would be the cutoff levels? Data 19

MAQ Corporation Use ABC Analysis. xlsx Step 1 rank order from fastest moving to slowest Step two: calculate each SKU’s percent of total sales Step three: cumulative percent of sales Step four: draw lines at 20%, next 40%; and final 50% of items Note how close it comes to 80%; 95%; and last 5% of sales 20

Let’s take up the questions of: 1. How much should I order? 2. When should I order? 21

How much should I order? Total Inventory Cost includes = Holding Cost + Ordering Cost The Economic Order Quantity (EOQ) is the lowest total cost recognizing the trade off of: 1. As order size increases, holding cost increases 2. As order size increases, order cost decreases 22

Economic Order Quantity Costs Total Cost Order Cost ICC Order Size in 100’s of Units Larger order sizes bring a steady decline in ordering cost per unit purchased. A very typical fixed cost curve. I have a great big one time fixed cost so the more I order the lower that fixed cost is per unit. Larger order sizes bring steadily increasing Inventory Carrying Costs (ICC). Larger orders = higher inventory = higher ICC Add the two curves together and you get a total cost curve. With increasing order sizes Total cost declines then later rises. The optimum is the low point directly above where the odering cost = the ICC

Fixed Ordering Costs • Buyer time – Time to review inventory – Time to place an additional order • System cost to place order • Receiving costs – Receiving – Inspection – Putaway – Purchase order matching – Inventory update • Manufacturing Set-up Costs In automated inventory replenishment systems a lot of the buyer time has been greatly minimized if not eliminated. Yet for items controlled by these system that are new, have erratic demand, or have extremely high value, or extremely long lead times; buyers review system recommendations. 24

Optimal order quantity is where: Ordering costs per unit of time = Holding cost per unit of time Order Cost RS Q = KQ Holding Cost 2 S = Fixed ordering or set-up cost R = Annual Demand Q = Order quantity K = Cost to hold inventory $ per unit Solving for Q results in the EOQ formula: EOQ = 2 RS K Holding Cost = ICC (Inventory cost) 2 indicates the average inventory from an order quantity is sold over time at a constant rate until the next order arrives (shown on next slide) 25

Cycle Inventory EOQ Inventory Cycle inventory is the EOQ (Economic Order Quantity) divided by two to represent the average inventory during the cycle of receiving one order (EOQ) till receiving the next. = EOQ 2 Time 26

EOQ Assumptions 1. Demand is known and constant 2. Order quantities are fixed at Q items per order 3. A fixed cost incurred on each order 4. Holding cost accrued per unit of time 5. Lead time is known and constant at 0 6. Initial inventory is 0 7. Infinite planning horizon Do all of these exist in all situations? No But EOQ does force you to balance acquisition costs with holding costs. Lets look at how to counter number 5 (Lead time is known and constant at 0). Meaning the new order always arrives just when we sell the last unit of the last order. Next slide has our 27 answer.

When should I Order? A Min – Max System allows for lead time & answers that question EOQ Cycle Inventory EOQ ÷ 2 Lead Time Inventory S Re-order Point = L x AVG Time Lead Time= lead time (L) X Average Demand per unit of time (AVG) The maximum is the amount that receiving an order sized based upon the EOQ. The trick is when to order. If I am selling 10 units per day and my lead time is 10 days, I must order when inventory reaches 100 units. This is my re-order point. This also referred to as Lead Time Inventory. Not to be confused with In-Transit inventory. In-transit inventory is the last order you placed (EOQ) which has not arrived yet. It is in-transit 28

Nittany Fans 29

Nittany Fans of Lewistown, Pennsylvania, is a distributor of industrial fans used in plants, warehouses, and other industrial facilities. The fans are manufactured in Neenah, Wisconsin, and currently shipped to Lewistown via rail transportation. Kenny Craig, vice president of logistics, has asked his staff to evaluate using motor carrier service to ship the fans. Nick Gringher, director of distribution, has collected the following information on the next slide 30

Nittany Fans: Data 31

Nittany Fans: Questions 1. 2. 3. 4. 5. What is the economic order quantity for Nittany Fans in units? In pounds? What is the total cost (w/o transportation) of the EOQ? What is the total cost for using rail transportation? What is the total cost for using motor carrier transportation? Which alternative should Nittany Fans use? 32

Let’s take up the questions of: 1. Do I have enough inventory? 2. How do I deal with demand uncertainty? 3. How do I deal with supply uncertainty? 33

Safety Stock protects from: 1) Uncertainty of demand 2) Uncertainty of supply EOQ Inventory S So now lets answer the question: What do I do when things go wrong: 1) If I sell more than I planned; and or 2) The shipment from my supplier is late. To mitigate this risk we have safety stock, Some refer to it as just-in-case inventory. Re-order Point s Cycle Inventory Safety Stock Time 34

Protects from Uncertainty Of Supply Protects from Uncertainty Of Demand Safety Stock = Z x (AVGL x STD²) + (AVG² x STDL²) Where: AVG = Average demand per unit of time AVGL = Average lead time z = Statistical service level safety factor STD = Standard deviation of demand STDL = Standard deviation of lead time Protection from uncertainty of demand during lead time. AVGL is the average lead time (lets say in days) and STD is the standard deviation of demand for one day. So this represent the inventory I need to protect me from variance in demand during lead time causing me to be out of stock. Protection from uncertainity of supply. AVG is average demand per day. STDL is standard deviation lead time. So this represents the inventory I need to protect me from being out-ofstock because the shipment is late. 35

Let’s Look at an example Use Safety Stock Calculator. xls posted in Week 3 Document Sharing for the example. AVGL = Average lead time and is 15 days in this example The variability of demand is relatively low at 2 units. Average demand per day is 10 units The variability of lead time is measured by the standard deviation of lead time and is 2 days here. Try changing any of: Average lead time; standard deviation of demand; average demand per day; and or stnadard deviation of lead time. What happens to safety stock? Why? 36

Shown here is the difference between central standard deviations and Z the area under the curve from the far left to the standard deviation you want on the right. Relate this to example and the same curve in the safety stock calculator. xls

Statistical z values required by desired service level Service Level 90% 91% 92% 1. 29 1. 34 1. 41 93% 94% 95% Z = 148 Service Level 96% 1. 56 97% 1. 65 98% Z = 1. 75 Service Level 99% 1. 88 99. 9 2. 05 Z= Service Level Z= 2. 33 3. 08 Z factors are multiples of the standard deviation to get the service level you desire. Relate to example or use Safety Stock Calculator. xls 38

Average Inventory = S EOQ + SS 2 EOQ = 2 RS K s Cycle Inventory SS = Z x Re-order Point = L x AVG (AVGL x STD²) + (AVG² x STDL²) So here is all of our inventory. We can see our order size (EOQ); our cycle inventory (1/2 of the EOQ assuming we sell the EOQ uniformly over time); our re-order point or lead time inventory (remember it is different than in-transit inventory; intransit is last order –EOQ- that has not arrived yet); and we can see our safety stock. Note our average inventory is the cycle inventory plus our safety stock. 39

Maxim 1: Higher service levels increase safety stock Go back to the Safety Stock Calculator and see how much safety stock increases by increasing the desired service level from 90% to 99. 9% 40

Maxim 2: Higher Variability of demand increases safety stock SKU A SKU B SKU C s 1 s 1 By looking at the 3 demand curves above: Which SKU has the highest variance? Which SKU has the largest standard deviation? Which SKU should have the largest safety stock? Answer 41

Maxim 2: Higher Variability of demand increases safety stock (Answer) SKU A SKU B SKU C s 1 s 1 By looking at the 3 demand curves above: Which SKU has the highest variance? Which SKU has the largest standard deviation? Which SKU should have the largest safety stock? The answer to all 3 questions is “A”. It is broader meaning more variance and a higher standard deviation. This means more risk of a stock out so more safety stock is needed. 42

Maxim 3: Increased lead time increases safety stock Protects from Uncertainty Of Demand Safety Stock = Z x Protects from Uncertainty Of Supply (AVGL x STD²) + (AVG² x STDL²) Lead time (L) appears twice in the safety stock formula. First in the protection from uncertainty of demand during lead time (the longer the lead time the more risk of incurring a variation of demand during that lead time. Second in the protection from uncertainty of supply (the longer the lead time the greater the odds of a missed shipment, late shipment, errors, problems encountered on the way)43

Maxim 4: Periodic review replenishment requires more safety stock than continuous review replenishment Safety Stock = Z x (AVGL x STD²) + (AVG² x STDL²) Periodic review says I will review this product line once a week or maybe one a month and place an order with my vendor at that time. If I do that, I just added a month to my lead time from when the sku broke order point. Continuous review with an inventory control, MRP, or DRP system looks at each sku (stock keeping unit) daily and issues an order if the sku broke order point. So longer review for whatever reason add to lead time which requires more safety stock 44

Inventory Level Maxim 5: Disaggregating demand increases safety stock The change in average inventory Equals the Square Root of the ratio of Number of Locations New Locations Old Locations Applies to stocking locations, SKU’s, Components, and Channel Partners If I have only one distribution center and 1000 retailer stores I am shipping to, I only have to predict what I am going to sell to each retail store. If I divide the country in half and have an eastern dc and a western dc while assigning all the retail stores in the east to the eastern dc and the western retail stores to the western dc; I now have to predict two things. First that I am going sell the product to a retail store and which dc I am going to ship it from. Having to predict two thing increases my forecast error because I have disaggregated demand (split between the east and the west). Therefore, I need more safety stock. Likewise if I consolidate 6 regional dc’s into one centralized dc I only have to predict 1 thing (I will sell it) rather than 6 things (which dc I am going to ship from). Forecast error drops. Safety Stock drops. This is called risk pooling which is much like what insurance companies do when they can insure because many-many people are paying premiums a very-very few get sick. Risk Pooling is also called the portfolio effect. It can be very accurately estimated using the square root formula shown here. It applies to number of stocking locations (DC’s); SKU’s; Components; and channel partners; even time it is the driver of what supply chain professionals call Postponement. Examples next -> 45

Brand Auto Parts Manufacturer 1 FG Inventory P GBrandes 10 FG Inventories Brand PBL Sleeve & ship 10 labels PBL PBL PBL P P Bin PBL PBL PBL 1 bin inventory PBL PBL PBL An automotive parts manufacturer selling automotive fan belts and radiator hoses had one finished goods inventory of fan belts to sell to their distributors. Nine of their biggest distributors started their marketing program and brand; therefore wanted the belts to be shipped with their brand’s label (APS, NAPA, Auto. Zone, Bumper to Bumper etc) That meant they now had 9 unique inventories of the same SKU. The manufacturer saw inventory triple. So they started a system keeping the belt as the manufacturers’ brand until an order was received; then labeled it with the customers’ brand shipped. Inventory went back to where it stated.

National Bicycles - Japan New Push-Pull Boundary Parts and Components Standardized parts for 11 million combinations Weld Frame Traditional Push-Pull Boundary Paint Assemble One day to weld, paint, & assemble. Next day delivery Retail Fitting Machine Electronic order to the plant National Bicycles (sub of Matsushita) sells Panasonic and National brand in japan. Stagnant sales due to not being able to predict and satisfy varying customer demand. 20% of bicycles from the previous year remained unsold adding to overstock burden. Sport bikes, 10 speed, and mountain bikes had become fashion items sold partly on intricate colors and patterns changing every year. So National introduced a postponement based mass customization strategy. They standardized parts that could give up to 11 million combinations. A lot of choice.

Benetton: knitting first & market decides color Spin Yarn Dye Yarn Knitting Traditional Push-Pull Boundary Package Retail New Push-Pull Boundary Spin Yarn Knitting Dye Major supplier of knitwear; Europes largest clothing manufacturer; world’s largest consumer of wool in the garment sector. The nature of the fashion industry is that consumer preferences change rapidly. Due to long manufacturing leads times, retailers had to place orders for wool sweaters up to 7 months in advance of when the sweaters would arrive at their stores. The wool sweater manufacturing process typically consisted of : acquire the yarn; spin the yarn; dye the yarn; knit the dyed yarn into a sweater; package; and ship to retailers. The new method was to acquire; spin; knit (now before dye); dye; package; ship. This allowed for delayed differentiation of the sweater into its final form. Thus moving the push-pull boundary back. Color choices could be delayed until more forecasting and sales information were received. Postponing the dying process, yarn purchasing and manufacturing could be based upon aggregate forecasts for product families rather than forecasts for specific sweater color combinations. Traing and changing the manufacturing process was needed. Manufacturing costs increased 10% but this was more than gained back by less inventory; less overstock and under stock penalties; and increased sales due to always having the right product at the right time. Due to the postponement of the dyeing process; 48

Postponement Requires 1. 2. 3. 4. Modular Product Design Modular Processes Faster Information Fast execution 1. 2. 3. 4. Postponement Results Increase manufacturing costs Decrease inventory by aggregating demand Decreased logistics costs Decreased stock-outs & markdowns Step 1 Step 2 Step 3 Step 4 Modular product design so that you have components (building blocks) that can be assembled into many different end products. Processes also need to be modular so that differentiation can be delayed; and the full range of options are open until you have an order. Fast information and execution is mandatory because you are waiting until you have an order before you produce. Customers still have delivery expectations. Manufacturing costs typically increase but are more than offset by reduced logistics costs; decreased stock outs & markdowns. All due to better forecasting based upon aggregated demand.