# You are the GaTech Barnes & Noble Merchandise Buyer MANAGING INVENTORY IN THE FACE OF UNCERTAINTY The Newsvendor Problem MGT3501 You are the GaTech Barnes & Noble Merchandise Buyer Its Sunday, March 28, 2004. The Jackets just made it to the Final Four. Before heading out to celebrate, you need to call your T-shirt supplier and order Final Four merchandise. #BB12 - Final Four tee How many do you order? 100% cotton -shirt with Final Four logo \$21.98. The Newsvendor Framework One chance to decide on the stocking quantity for the product youre selling Demand for the product is uncertain Known marginal profit for each unit sold and known marginal loss for the ones that are bought and not sold Goal: Maximize expected profit Examples where the Newsvendor framework is appropriate Perishable goods Meals in a cafeteria Dairy foods Short selling season Christmas trees, toys Flowers on Valentines day Fashion clothes, seasonal clothes Newspapers The Newsvendor Problem

Newsvendor selling the AJC Sells the papers for \$1.00 Buys them for 70 cents Leftover papers sold at discount: 20 cents each. He will definitely have at least 35 customers, but no more than 40: 35 customers with probability 0.10 36 customers with probability 0.15 37 customers with probability 0.25 38 customers with probability 0.25 39 customers with probability 0.15 40 customers with probability 0.1 How Many Papers Should He Buy To Maximize His Expected Profit? Marginal analysis x Probability that demand = x <35 0 35 0.1 36 0.15 37

0.25 38 0.25 39 0.15 40 0.1 >40 0 P = Prob. of selling the xth unit (1-P)=Prob. of NOT selling the xth unit P x \$0.3 =Expected profit from selling xth unit (1-P) x \$0.5 =Expected loss from NOT selling xth unit Expected NET profit from stocking xth unit

1.0 0 \$0.30 \$0 \$0.30 Marginal analysis x Probability that demand = x P = Prob. of selling the xth unit (1-P)=Prob. of NOT selling the xth unit P x \$0.3 =Expected profit from selling xth unit (1-P) x \$0.5 =Expected loss from NOT selling xth unit Expected NET profit from stocking xth unit <35

0 35 0.1 1.0 0 \$0.30 \$0 \$0.30 36 0.15 0.9 0.1 \$0.27 \$0.05 \$0.22 37 0.25 38 0.25 39 0.15 40 0.1 >40

0 Marginal analysis x Probability that demand = x P = Prob. of selling the xth unit (1-P)=Prob. of NOT selling the xth unit P x \$0.3 =Expected profit from selling xth unit (1-P) x \$0.5 =Expected loss from NOT selling xth unit Expected NET profit from stocking xth unit <35 0 35 0.1 1.0

0 \$0.30 \$0 \$0.30 36 0.15 0.9 0.1 \$0.27 \$0.05 \$0.22 37 0.25 0.75 0.25 \$0.225 \$0.125 \$0.10 38 0.25 39 0.15 40 0.1

>40 0 Marginal analysis x Probability that demand = x P= Prob. of selling the xth unit (1-P)=Prob. of NOT selling the xth unit P x \$0.3 =Expected profit from selling xth unit (1-P) x \$0.5 =Expected loss from NOT selling xth unit Expected NET profit from stocking xth unit <35 0 35 0.1

1.0 0 \$0.30 \$0 \$0.30 36 0.15 0.9 0.1 \$0.27 \$0.05 \$0.22 37 0.25 0.75 0.25 \$0.225 \$0.125 \$0.10 38 0.25 0.50 0.50 \$0.15

\$0.25 \$-0.10 39 0.15 0.25 0.75 \$0.075 \$0.375 \$-0.30 40 0.1 0.1 0.9 \$0.03 \$0.45 \$-0.42 >40 0 0 1 \$0.0 \$0.50 \$-0.50 Marginal analysis x

Probability that demand =x P= Prob. of selling the xth unit (1-P)=Prob. of NOT selling the xth unit P x \$0.3 =Expected profit from selling xth unit (1-P) x \$0.5 =Expected loss from NOT selling xth unit Expected NET profit from stocking xth unit <35 0 35 0.1 1.0 0 \$0.30 \$0 \$0.30

36 X*=37 0.15 0.9 0.1 \$0.27 \$0.05 \$0.22 37 0.25 0.75 0.25 \$0.225 \$0.125 \$0.10 38 0.25 0.50 0.50 \$0.15 \$0.25 \$-0.10 39 0.15 0.25

0.75 \$0.075 \$0.375 \$-0.30 40 0.1 0.1 0.9 \$0.03 \$0.45 \$-0.42 >40 0 0 1 \$0.0 \$0.50 \$-0.50 Decreasing marginal returns to each additional unit. Solution Start from 35, buy an additional paper only if you expect to make extra profits. Lets generalize! c: cost r: selling price s: salvage value MP: marginal profit from selling a stocked unit = r-c ML: marginal loss from NOT selling a stocked unit = c-s x: the number of newspapers you buy.

P(x): the probability that the xth newspaper is sold = P(D x) Solution Buy one more unit only if the expected net profit of doing so is positive. Buying x is profitable if P(x) MP - (1-P(x)) ML 0 P(x) ML / (MP + ML) Critical ratio: Pc=ML / (MP + ML) Optimal solution: Buy largest x such that P(x) Pc Solution In this problem: MP = \$1-\$0.70 = \$0.30 ML = \$0.70 -\$0.20 = \$0.50 Pc=ML / (MP+ML) = 50 / (50+30) = 0.625 Buy largest x such that P(selling xth unit) Pc Marginal analysis x Probability that demand =x P= Prob. of selling the xth unit (1-P)=Prob. of NOT selling the xth unit P x \$0.3 =Expected profit from selling xth unit (1-P) x \$0.5 =Expected loss from NOT selling xth unit

Expected NET profit from stocking xth unit <35 0 35 0.1 1.0 0 \$0.30 \$0 \$0.30 36 X*=37 0.15 0.9 0.1 \$0.27 \$0.05 \$0.22 37 0.25 0.75 0.25 \$0.225

\$0.125 \$0.10 38 0.25 0.50 0.50 \$0.15 \$0.25 \$-0.10 39 0.15 0.25 0.75 \$0.075 \$0.375 \$-0.30 40 0.1 0.1 0.9 \$0.03 \$0.45 \$-0.42 >40

0 0 1 \$0.0 \$0.50 \$-0.50 Pc = 0.625 Now Lets Assume that Demand is Normally Distributed 0.09 PROBABILITY DENSITY 0.08 MP=0.30 ML=0.50 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 DEMAND How do we find order quantity X*? Remember solution: Buy largest x such that P(selling xth unit) Pc P(selling the xth unit) =P(Demand x) 0.09 0.08 PROBABILITY

0.07 0.06 0.05 0.04 P(Demandx) P(Demand

INCREASE COST OF LOST SALES DECRESE COST OF LOST SALES DECREASE COST DUE TO UNSOLD ITEMS X* INCREASE COST DUE UNSOLD ITEMS Newsvendor: Solution with Normal Distribution Remember: MP = 0.30; ML = 0.50 Then Pc is Pc=ML / (MP+ML)=50/(50+30)=0.625 Look at the z table to find the z corresponding to 1-Pc = 0.375 In this case: z=-0.32 X*= + z = 37.5 + (-0.32) 1.44 = 37.04 What Do We Learn from the Newsvendor? Forecasts are always wrong. A demand estimate that only gives the mean is too simple, you also need the standard deviation. The optimal order quantity depends on the relative cost of stocking too much and stocking too little. The smaller the standard deviation, the closer will be the order to the mean.