联系方式

  • QQ:99515681
  • 邮箱:99515681@qq.com
  • 工作时间:8:00-23:00
  • 微信:codinghelp

您当前位置:首页 >> javajava

日期:2018-10-05 09:58

MSBA7003 Quantitative Analysis Methods

This Case Is For Use Only In MSBA7003. Please Do Not Distribute This Case Or Its Content.

Case: Estimation of Jewelry Substitutions and Demand

Company Background

H.K. Jewelry (HKJ) started engaging in the jewelry business from early last century, with an extensive

retail network in Greater China, Korea, Japan, Singapore, Malaysia and the United States. In 2018, there

are more than 2400 retails shops around the world with 59,156 million of revenue contribution. Their

products are classified into three portfolios, high-end luxury, mass luxury and youth line. Within each

portfolio, there are various stock natures (or categories), including ring, necklace, bracelet, earrings,

pendant, chain, figure, bands and others.

Problem Description

Because jewelry products are of high value and slow moving, with mostly one unit for each purchase,

demand forecasting and inventory management are of great importance for the business. The managers

need to forecast the demand during a period for each product first and then decide how many units of

that product to carry in a store. If the inventory level is too high, the costs of capital and custody will be

high. If the inventory level is too low, the cost of lost sales will be high. However, demand forecasting

and inventory management are highly challenging for HKJ.

This is largely due to the uncertainty in customers’ tastes of product designs. For any particular category,

different customers may have different tastes of product designs. If a customer cannot find a design she

likes, the potential sales will be lost. Currently, what HKJ does is trying to offer many different designs

for each product category so as to capture customer demand as much as possible. However, there are

almost infinitely many possible designs for each category and it is costly to carry too many products, so

it is impossible for HKJ to cover all the possibilities.

An important behavior among customers is demand substitutions among different products. Specifically,

when a customer finds that her desired product (or design) is not carried in the store, she might be

willing to switch to a different product (or design) which is available. Such a substitution effect

complicates demand forecasting and inventory management, because it suggests that the actual

demand of a product not only depends on its own stock level but also the stock level of other products

for which the product can be a substitute.

However, at the same time, demand substitutions offer an opportunity for HKJ to reduce demand

uncertainty. This is because it is still possible to capture demand even if a customer’s first choice is not

available as long as the second choice is available and the willingness of switching to the second choice

is high for the customer. Therefore, the HKJ can try to find out which products (or designs) are likely to

be the substitutes for other products. If a product can substitute many others, it can be identified as a

core product. Then HKJ can focus on these core products and products that cannot be easily substituted.

In this way, HKJ can maximize sales while minimizing cost of carrying the products.

To this end, it is important for HKJ to estimate the substitution rates among different products and build

a demand prediction model that incorporates such substitutions. Once these are done, the managers

can proceed to optimize their product offerings.

MSBA7003 Quantitative Analysis Methods

This Case Is For Use Only In MSBA7003. Please Do Not Distribute This Case Or Its Content.

Four Scenarios of Sales Encounter

It is important to start from understanding various sales situations. For a customer whose first choice is

product A, there are four scenarios for her to make a decision.

Scenario 1: A is in-stock and displayed

As shown in the Figure 1, if product A is in-stock and displayed, the customer will definitely purchase the

first choice (i.e. product A).

Figure 1. Scenario 1

Scenario 2: A is in-stock, not displayed, and sold eventually

If product A is in-stock but not displayed, there are two cases to consider about. The first one is that the

customer successfully find product A with the help of sales staff and consequently make the purchase.

Figure 2 illustrates this scenario.

Figure 2. Scenario 2

Scenario 3: A is in-stock, not displayed, and not sold

The second case if product A is in-stock but not displayed is that the sales staff fail to help the customer

and consequently lose the sales of product A. In this case, the customer may consider buying another

product such as product B or C as a substitute, and this will lead to a sales increase of the substitute

product. Figure 3 illustrates this scenario.

MSBA7003 Quantitative Analysis Methods

This Case Is For Use Only In MSBA7003. Please Do Not Distribute This Case Or Its Content.

Figure 3. Scenario 3

Scenario 4: A is out-of-stock

If product A is out-of-stock, the customer will never be able to purchase product A. Similar to Scenario 3,

the customer may consider buying another product such as product B or C as a substitute, and this will

lead to a sales increase of the substitute product. Figure 4 illustrates this scenario.

Figure 4. Scenario 4

Data Set

The training data set to be used for the analysis contains the historical daily inventory and sales

information of 72 products at a single shop in Hong Kong during the period from January 3, 2015 to

October 2, 2016. The products are of the same subcategory (pendants) and exhaustively cover the

subcategory.

Specifically, for each product on each day, the data set indicates whether the product is in-stock,

whether it is on display, and the number of units sold. The data set also offers the number of visitors to

that shop in each day. Figure 5 shows a screen shot of the data set.

Questions

As the appointed data analyst, you are asked to develop a model and estimate the parameters using the

training data. Then your results will be tested against a testing data set, which spans a three-month

period from October 3, 2016 to January 2, 2017. You will be given information such as daily store visits,

whether a product is in-stock, and whether a product is on display in the testing data set. Your predicted

daily sales of each product will be compared against the actual sales.

How could you use the substitution information to improve the business in other ways MSBA7003 Quantitative Analysis Methods

This Case Is For Use Only In MSBA7003. Please Do Not Distribute This Case Or Its Content.

Figure 5. Screen shot of data set.

This Case Is For Use Only In MSBA7003. Please Do Not Distribute This Case Or Its Content.


版权所有:留学生程序网 2020 All Rights Reserved 联系方式:QQ:99515681 电子信箱:99515681@qq.com
免责声明:本站部分内容从网络整理而来,只供参考!如有版权问题可联系本站删除。