Description
Organize notes based on PPT and make them as concise as possible
Unformatted Attachment Preview
Welcome to
BU.450.740
Retail Analytics
Session 1: Overview
Prof. Mitsukuni Nishida
Johns Hopkins Carey Business School
1
48% of firms have created a Data-Driven
Organization, but…
Source: Wavestone, 2024
2
77%+ report barriers to data-driven:
People, Process, & Culture—not Tech
Source: Wavestone, 2024
3
Marketing is 9th of 47 of jobs
Gen AI can most easily perform
4
Source: indeed 2023
Who Should Take Retail Analytics
• Ability to analyze data for better business decisions
is increasingly important in all level of organization
• Retail Analytics @ Carey…
…is the only course in the United States that
integrates tools on machine learning, including deep
learning, and geographical information system to
marketing topics in retail pricing, entry, and
promotions, using analytical models from Economics
and Marketing
• By taking RA, you will have tools to bring more
value to all industries that hire tech-savvy
executives
– Citigroup announced to train all incoming
analysts in Python
– Goldman Sachs has asked its traders to learn
how to code
5
Welcome to Retail Analytics!
• So excited to teach this course!
• You will learn..
– How to conduct data analysis as data scientist in retailing
– How to conduct market analysis as industry analyst
– ..& hopefully find startup opportunities in retailing
• Learning never stops.. think of this course as a “gateway” to
various ideas & future career paths
6
What is Retail Analytics?
Business needs
Identifies the business need
Data: acquisition &
cleaning
Analysts examine data to uncover
Hidden patterns and correlations
Analytics
Decisions
Based on analytics, retail
firms make corporate-level
decisions
7
3 Pillars for Retail Analytics
1. Retailing
– Overview, pricing, entry & expansion, location choice, revenue predictions,
omnichannel and use of AI
2. Theory (“Analytical model”)
– Marketing: Pricing
– Strategy: Industry analysis, competitive advantage
– Microeconomics: Price theory, location model, and game theory
3. Statistical methods (“Statistical model”)
– Descriptive analysis, regressions, machine learning, deep learning
– Spatial analysis
•
Implementations: R (Python as option) and ArcGIS
•
Nature of the course is applied:
– We use pillars 2 and 3 as needed to examine issues in 1.
•
Nature of the course is gateway:
– We won’t cover the theoretical details– this course provides intro to learn
more on these concepts/methods in future
Theories & Models: Multi-disciplinary
• Marketing
– Pricing
– Marketing strategy
• Economics
– Microeconomics & game theory
– Econometrics
• Statistics
– Statistical analysis, Machine Learning, and Deep Learning
• Operations Research
– Logistics
• Strategy
– Industry analysis
Knowledge and skills are mostly portable to other industries, but our
9
focus is on retailing
How does Economics fit into Retailing?
Retailing Sector
Financial
aspects
Product/Service
aspects
Labor aspects
Common thread? Demand and supply
→ Understanding of how market works applies to all these markets!
10
.. & strong Links to Many Disciplines!
Economics is THE (or one of THE) founding blocks for a number of
B-school disciplines
Economics
Finance
• A sub-field of
economics
• Key ideas
published in
top economics
journals
Strategy
Accounting
Operations
Research
Marketing
• Economics
underlies a big
subset (game
theory, IO)
• Much research
borrows from
economics
(game theory,
principal-agent
models)
• How to “shift
down cost
curves”
• Research uses
game theory &
other econ
tools
• Large overlap
with economic
models of
pricing,
demand and
distribution
… and so graduate work in all these fields usually involves a strong
economics component
11
Statistical Methods: Three Goals
1. Describing and visualizing data
– Method: Summary stats and plots
– “How was advertising related to sales last year?”
2. Finding data patterns
– Method: Plots, correlation, machine leaning
– “Based on past purchases, what would be the
next product that this consumer would
purchase?”
3. Finding causal relationship
– Method: RCT (e.g., A/B testing): Field or lab
experiment
– “Do internet ads increase sales or vice versa?
This course mostly deals with 1 and 2
12
Today’s Agenda
• Introduction
• Course logistics (see syllabus for details)
– I will update the syllabus. Please check.
– The course is ambitious. Please check course load and see if it
fits your need
• Retailing: Overview
– Career in Retailing
– Understand modern marketplace and tech aspects
• Theoretical concepts: Industry analysis
• Intro of RStudio as your development environment
13
Contact Information
• Email: [email protected]
• Office hours: After the session
• Course website: Canvas
• Teaching Assistant: Victor Liang [email protected]
– Grades your submission
14
Brief Introduction of Myself
Univ. of Chicago, PhD in Economics
Since grad
school, I’ve been
fascinated by
retail chains’
behaviors
Tokyo, Japan
Loved Physics and Math
Economist at Economic Outlook Division,
Japanese government
BA in International Relations, Kyoto Univ.
Thesis “On Dean Acheson’s Foreign Policy”
15
In a nutshell…
Always interested in how this world works through different “lens”
• My passion has been always about retailing and service sectors
• Tools I use to analyze reality
– Microeconomics, Quantitative Marketing, Machine Learning
• Topics I work on
– Entry, exit, and expansion of retail chains
– Location of retail chains
– Franchising decision
– Airline and gasoline pricing
– Search costs estimation
– Aggregate and firm-level productivity
• Exposure to “real” business
– Consultant for a S&P 500 Company
– CEO’s secretary at a molding company
16
Course Resources
•
Lectures:
– Students are responsible for everything that is explicitly
covered in class
•
Canvas:
– Depository for lecture slides (ppt files), codes, homework, supplemental
readings, etc.
– Submission of homework and group-work materials
•
Course materials: lecture slides, book chapters, other readings
•
No textbook for retailing due to recent technological progress:
– By the time you read these topics in the textbooks, they are too obsolete
– Instead, we go back to theoretical foundations in Marketing & Economics
– Read newspaper, academic papers instead. Will post some.
•
Having said that, Levy et al.’s Retailing Management 11th edition is a good
handbook for reference
17
Week
Topic
Theory/Model
Statistical methods
Course Schedule
Overview of retailing
1
Understanding modern
retailing marketplace
& technological aspects
Measuring price and
promotion response in
retailing
4
Introduction of development
environment in RStudio
Group presentation
Finalize the group members
Descriptive analysis
Competitive advantage
Linear models
HW1 (individual) on R basics
and “spark city”
Price theory: elasticities,
economies of
scale/scope
Installation of R and RStudio
Retailers’ entry and
expansion in retailing
Game theory: simultaneous Nonlinear models: probit and
and sequential-move games logit regressions
HW2 (individual) on linear
regressions
Retailers’ site location
decision & target
marketing
Spatial analysis: Huff model, Tutorial on ArcGIS
regression model
HW3 (individual) on nonlinear
models
2
3
Industry analysis
Homework, due at the
beginning of class
Phase 1 report by the end of
week 2’s lecture day
Installation of ArcGIS
5
Retailers’ revenue
predictions
Price discrimination
Omnichannel in the age of Yield management
AI
Machine leaning: Tree-based
methods
HW4 (individual) on spatial
analysis
Time series analysis
HW5 (individual) on machine
learning
Difference-in-Difference (DID)
6
Deep learning: Recurrent
neural network
Group presentations
7
8
Slides for presentation (group)
Research report (group)
In-class final exam
18
Peer evaluation (individual)
Evaluation
•
Attendance and class participation: 10%
•
Individual homework (five sets): 30%
– 6% each
•
Group presentation and deliverables: 30%
– Presentation: 20% (Instructor: 10%, floor: 10%)
– Final report: 5%
– Peer evaluation within a group: 5%
•
Final exam: 30%
– Traditional format: questions & answers: 15%
– Practical R implementation: 15%
19
Attendance and Participation
• Attendance: 5%
– Note: Because the course is designed as in-person, attendance
points are obtained via in-person attendance
– You can skip one lecture out of seven lectures without being
penalized for absence
• No need to ask me in advance for permission to be absent
– Please record your attendance on Qwickly Attendance on Canvas
• Participation (a.k.a. asking questions): 5%
– We encourage you to ask questions once in the course by the end
of Week 7
– On Week 7, I will ask you to record your participation on Qwickly
20
Re-grade policy
1. Check the addition of points on individual questions.
2. Check your answers against the answer key posted on the
course website.
•
If you have questions about how your test was graded or
why your answer was incorrect, return your test to the TA
(& cc’ing me) with a short, written description of your
question.
•
TA will check the grading on the entire test. I and TA
return the test to you with an explanation for the grading
and, if appropriate, an adjusted score.
•
If it is still not clear or you are otherwise not satisfied,
make an appointment with me to go over the questions.
21
Grading
• Letter grade
– A: reserved for those who demonstrate extraordinarily
excellent performance as determined by the instructor
– A- is awarded only for excellent performance.
– B+ and B are awarded for good performance.
– B-, C+, C, and C- are awarded for adequate but substandard
performance.
– D+, D, and D- are not awarded at the graduate level
(undergraduate only).
– F indicates the student’s failure to satisfactorily complete the
course work.
• For Elective courses, the grade point average of the class should
not exceed 3.45
• Having said that, I would care about your learning experience
22
How to Succeed?
• Active learning in class:
– Stay focused
– Actively participate, work on coding/questions/problems
during class
• Please be on time
• Missed Classes – please review the recordings
• Please avoid:
– using laptops for other than computing and taking notes
23
Software: R and ArcGIS
• We introduce development environment for computing
• Why R?
– Popular development environment for data analysts
– Easy to use and tailor to your need
– Portable after graduation, unlike MATLAB, Stata, ..etc.
• Why ArcGIS?
– Not portable, but THE gold standard in spatial analysis
• Preparations
– Please check R & RStudio by the second half of this class
– Links and instructions for R are provided in the syllabus
24
– Please utilize your laptop for so you can run in-class exercises
25
Use of Generative AI
• Generative artificial intelligence (AI) tools such as ChatGPT are
widely available, and these technologies present a number of
exciting opportunities in the classroom
• In this course, you may use generative AI tools on homework and
your group project
• You may NOT use generative AI tools for final exam
26
GROUP PRESENTATION
27
Group formation
• Everyone must sign up for a group (Group 1, Group 2, ..)
– See canvas for a self sign-up, which is available until tonight
– The number of students in a group cannot exceed 5 people
– If there are unassigned students, I will be randomly assigning
them by midnight tonight
• Group presentation; Will be conducted in Week 7
– The goal is to force you to think about how you will utilize the
skills and knowledge that you acquire through this class to actual
business in retailing
– Each group has two tracks: RETAIL or ANALYTICS
– Format: 10 minutes (8 mts for presentation + 2mts Q&A from the
instructor and the floor)
28
Group project track 1: Retail
1. Proposing a new business startup project in retailing
– Goal: to come up with an entrepreneurship plan in retailing that
the group can pitch to venture capitalists
– What to do: (1) Identify potential business needs in retailing
and propose a feasible plan to (2) address the needs given the
current state of the business environment
– Presentation includes an analysis of the market for the client’s
products or services, an implementation timeline & budget
– The analytical tools will include, but is not limited to the Porter’s
five-forces model and competitive advantage
• (Past) Example: Use of location for proximity marketing, VR tools
for retailer staff training, AI based store management,…
– Be creative & innovative. Please avoid choosing business
models that already exist
– Extra points (3%) if the project acquire data and applies data
analytics skills (i.e., regression, machine learning) to the data
29
set
Group project track 2: Analytics
2. Conducting a data project in retailing
– Collecting, cleaning, and conducting data analysis in retailing
– Goal: each group identifies and addresses an existing marketing
(& operational) issues in pricing, entry, location, promotion,
products, and logistics via the use of analytics
– The data set to work with can come from:
• The original data set you have access to obtained from network (e.g.,
previous employer)
• Dominick’s scanner data: Most time-consuming to clean up &
understand, but probably most rewarding in terms of experience
• Market-level aggregated store count and revenues for conveniencestore chains in Japan
– Extra points (3%) if the group finds out and utilize its original data
set or utilize Dominick’s
– Be sure to be clear what kind of business needs you are going to
address before you dive into big data
30
Week 2: Phase 1 report
• First, discuss with your group members about what option the group
pursues & possible ideas. I would recommend your group to discuss
this week
• Please submit the report that contains the following by the end of
week 2’s lecture day
– Two possible ideas, specifying either Retail or Analytics track
– Brief discussion on
• Why each idea is relevant for retail business (tracks 1 and 2)
• Timeline of the group work
• Format: One paragraph for each idea, 12-point Times New Roman
font, single spacing, one page in total
• Not for grade; Instructor gives feedback which one is more
promising to pursue
31
Week 4: Phase 2
• No report is necessary for Phase 2
• Expectation
– For track 1, Complete industry analysis and competitive
advantage
– For track 2, Finish cleaning and summarizing data. Conduct
descriptive analyses
– For both tracks, week 4 is a good time to discuss the following
within the group: What the group has done; findings so far;
challenges the group has encountered, and how the group is
going to address the challenges
32
Week 7: Presentation
• Each group is given 10 minutes slot in total (8 minutes
presentation + 2 minutes Q&A)
• Each group submits slides and a research report
• Slide format: 8 slides max
• Report format: 12-point Times New Roman font, single spacing, 8
pages max in total
33
Peer Evaluation for Group Project
• Anonymous peer evaluations within an assigned group will be
completed at the end of Week 7
• You will be giving anonymous evaluation to your colleagues
(excluding yourself) in your group in three dimension:
– Degree of effort
– Quality of output, and
– Communication skill
• You will be allocating 10 points in total for each dimension
– e.g., S4 is giving (S1, S2, S3, S5) = (3, 3, 2, 2) for effort etc.
• Occasionally, there may be issues of free-riding within a group.
Please try to resolve these issues within your group and if this
fails, please contact me.
• If a group member is not pulling their fair share on an assignment,
I encourage you to address the issue immediately. Early
34
intervention can typically resolve the matter and avoid escalation.
OVERVIEW OF
RETAILING (1): CAREER
IN RETAILING
35
Retail Analytics
By the end of the course, you will be able to envision yourself as:
Path 1: Industry analyst in retailing
– Able to analyze the industry for investors
• Entrepreneur in retail and service sectors
– Able to identify a business need and come up with new ideas
how we leverage retail analytics skills to address the need
Path 2: Data scientist in retailing
– Able to collect, clean, & analyze data, and interpret results
• Chief Technology Officer / Chief AI Officer
– Able to navigate the data scientist teams to utilize the findings
to managerial decisions
36
Management Opportunities
• People with a wide range of skills and interests are needed
because retailers’ functions include: Marketing, Finance,
Purchase, Accounting, Information system, Supply-chain
management including warehouse and distribution management,
Design and new product development, Data analytics
• Salary
– Management trainees make $35,000-$65,000
– What about store managers?
37
38
Entrepreneurial & Ladder Opportunities
1. Sam Walton (Walmart)
Started working at JCPenney, became America’s
wealthiest person
John Furner (Current Walmart US CEO)
“Walmart announced that John Furner will become
the President and Chief Executive Officer of
Walmart U.S….
Furner, 45, started with Walmart as an hourly
associate in 1993, working part-time in the garden
center of a supercenter.” (Oct. 10, 2019)
39
Entrepreneurial Opportunities (cont’d)
3. Ingvar Kamprad (IKEA)
Started with a loan from his father
Once the wealthiest person in Europe
4. Howard Schultz (Starbucks)
Began as head of marketing for
Starbucks
Later acquired Starbucks and built it
into 24,000 stores
40
OVERVIEW OF
RETAILING: MODERN
RETAILING
41
42
Retailing: Some history
• 50 years ago, retail was small, independent, local
• Now consists of large, national and international firms
• Active turnover over time (will cover in Week 3)
43
44
Top 10 Global Retailers Revenues
US firm
45
Source: CapitalOne Shopping Research
Retailing: Definition
• Retail Trade: NAICS 44-45
• “The Retail Trade sector comprises establishments engaged in
retailing merchandise, generally without transformation, and
rendering services incidental to the sale of merchandise.”
• Some people add “Accommodation and Food Services (NAICS
72)” as retailing
– Hotels, motels, restaurants, and coffee shops
46
Spectrum of “Retailing”
Merchandise Retailing
NAICS 44-45: Retail Trade
Service Retailing: Sell service rather than merchandise
Both retailing and service sectors can be called as “retailing”
47
Retailing: Category
•
20.0% – Motor vehicle & parts dealers
•
13.0% – Food & beverage stores
•
12.5% – General merchandise stores (hypermarkets, department stores,
discount stores, warehouse clubs): our focus today
•
11.0% – Food services & drinking places
•
10.0% – Gasoline stations (and convenience stores)
•
9.2% – Non-store retailers (Internet shopping, catalog, direct sales, etc.)
•
6.0% – Building material & garden dealers (home improvement)
•
6.0% – Health & personal care stores (pharmacy/drug stores)
•
5.0% – Clothing & clothing accessories stores
•
2.3% – Miscellaneous store retailers (specialty retailers)
•
2.0% – Furniture stores
•
2.0% – Electronics & appliance stores
•
1.7% – Sporting goods, hobby, book & music stores
48
Some Terminology
• Brick-and-Mortar Store Retailers – Those engaged in the sale of
products from physical locations which warehouse and display
merchandise with the intent of attracting customers to make
purchases on site.
• Non-Store Retailers – Those engaged in the sale of products
using marketing methods which do not include a physical location.
Examples of non-store retailing include:
– Mobile-only retailing (m-commerce)
– Internet-only e-commerce
– Infomercials
– Direct Response television advertising
– Catalogue Sales
– In-Home Demonstrations
– Vending Machines
49
Impact of COVID 19 on Retailing
50
51
Ecommerce Penetration:
“10 Years in 3 months”
Source: US Census
52
Retailing: Overview
• Retailing is one of the major economic sectors of the world
– 6.0% of US GDP in 2021: Note this number is value added
– 10.5% of US employees
• Retail “selling of consumer goods or services to the end buyer.”
• Bronnenberg and Ellickson (2015, posted on Canvas) provides
overview of global retailing from Economist and Marketing Scholars
• Fun facts
– Pre covid 19, e-commerce (“shopping online”): 10-11% of retail
sales in 2019 in US, 10% of global retail sales in 2018
• Amazon accounted for 37% of internet retail sales in 2019
– Most retail companies are small businesses
• 99% of them employ 50 people or less
• 95% of retailers have just one location
53
– 20% of annual sales occur between Black Friday & Christmas
Costs of Value-Added Activities in Distribution
Channel for a T-Shirt
Margin: 9.8%
Margin: 8.0%
Margin: 9.1%
54
Day-to-Day Climate Watch
55
Day-to-Day Climate Watch
• WSJ, NYT,
– Google alert (you set up manually)
• Twitter
– https://twitter.com/NRFnews
– https://twitter.com/retailwire
– https://twitter.com/RetailWeek
• Yahoo! Finance
– https://finance.yahoo.com and type in “retail”
56
Getting to know Retailing!
• Obtain first-hand information on retailers’ business
• Subscribe to news (will cover later in slides)
• Become a virtual Walmart staff
57
MODERN RETAILING IN
US
58
E-commerce share by category
59
TREND 1:
IS BIGGER BETTER?
NOT NECESSARILY
60
Major bankruptcies
2008 – 2015
2015-2019
Next one this year?
Source: CBInsights 2018
61
Winners & losers: By category
62
Bigger ≠ Better: More stores than ever..
63
Source: Nielsen, 2016
Bigger ≠ Better: …but size is declining
Source: Nielsen, 2016
64
Bigger ≠ Better: inequality as driving force
Source: CBInsights
65
Leverage a widening income inequality
“The middle-class
continues to go away,
unfortunately, to the lower
end of the economic
scale… so as this
economy continues to
create more of our core
customer, I think there’s
going to more
opportunities for us to build
more stores.”
Source: CBInsights
66
TREND 2:
RETAIL x TECHNOLOGY
x E-Commerce
= OPPORTUNITIES
67
Winners & losers
Common thread governing success & failures? IT Investments
68
Winners & losers in Electronics
Circuit city (bankrupt in 2009): Frequently
conduct promotions, resulted in losses
Bestbuy: invest in Branding, CRM (Customer
Relationship Management), profiling
demographics (e.g., “Jill” store)
69
Technology’s influencing everywhere
• Mobile and online technology
– Personalizing and engaging marketing & shopping experiences
• Product and service recommendation systems
• Scan-bag-go
– Walmart: attention on young customers via social media marketing
• Omnichannel
– Buy online- Pick up store
• Platform business that blurs online vs. brick-and-mortar (offline)
– Amazon acquired Whole foods in 2017
– Online shopping & offline shoppers (e.g., Instacart)
• Store operations
– Inventory management (e.g., replenishing, etc.)
– Digital shelves: changing displays for prices, ads, and navigation to
products
70
Business Opportunities in Retailing
71
Tech startups help retailers
Startups at intersection of retail and technology:
• Location analytics
• Optimized pricing (more on this in Week 5)
• Augmented reality (AR)
• Camera vision technology
• Guest WIFI
• Artificial Intelligence
– US apparel retail chain used ML for online product
recommendations, optimal inventory, & merchandising
.. among other things (see next slide)
Opportunity lies ahead for startups in retail!
72
73
THEORY: INDUSTRY
ANALYSIS
74
How should we analyze supermarkets?
(1/2)
• Retail Trade: NAICS 44-45
• Decompose bu 3 digit
– Motor Vehicle and Parts Dealers: NAICS 441
– Furniture and Home Furnishings Stores: NAICS 442
– Electronics and Appliance Stores: NAICS 443
– Building Material and Garden Equipment and Supplies
Dealers: NAICS 444
– Food and Beverage Stores: NAICS 445
– Health and Personal Care Stores: NAICS 446
– Gasoline Stations: NAICS 447
– Clothing and Clothing Accessories Stores: NAICS 448
– Sporting Goods, Hobby, Book, and Music Stores: NAICS 451
– General Merchandise Stores: NAICS 452
– Miscellaneous Store Retailers: NAICS 453
75
– Nonstore Retailers: NAICS 454
How should we analyze supermarkets? (2/2)
We follow four steps
• Step 1: Narrow down to 4 digit level industry group
– NAICS 4451 Supermarkets and Other Grocery (except
Convenience) Stores
– NAICS 4452 Specialty Food Stores
– NAICS 4453 Beer, Wine, and Liquor Stores
• Step 2: Identify major players within the group
– Kroger Food Stores
– Wegmans Food Markets
– …
• Step 3: Conduct Industry analysis (this week)
• Step 4: Conduct analysis on retail market strategy
– Competitive advantage (next week)
– SWOT
76
Explanation of Profits via Industry Analysis
Source: Ghemawat and Rivkin
Decomposition of Profitability
Source: McGahan and Porter (1991)
Annual effect
2.4
18.7
42.9
4.3
31.7
Industry effect
Parent firm’s
effect
Positioning
Unable to explain
Industry characteristics explain 10-25% of variation in industry profits
Why do we need “framework”?
• Drawbacks of “Excellent Company”?
– Outdated quickly (depends on time and context)
– Often difficult to share common understandings (personal)
– Causality is opaque; survival bias, pseudo-correlation cannot be
ruled out (not an experiment)
• Framework = Analytical framework and tools based on models
(economics)
– General principles behind why companies behave a certain way,
described by mathematical representation
– Model specifies relationship between players, objectives, choices,
choices and outcomes
– Does not get old (economics principle), can determine the
suitability/unsuitability of strategies in different situations
– Common understanding is easy (model)
– Causal relationships are clear (model), but beware of
oversimplification
“Excellent companies”
Marketing Strategy asks 2 questions
Industry
economics
Benefit position
relative to competitors
Added
value
Cost position relative
to competitors
(1) External
environment: What is
the nature of the
markets in which firms
compete? => Industry
Analysis
Economic
returns
(2) Business-level strategy: On what basis
a firms compete? How should firms
compete? => Competitive Advantage
Assets
81
Porter’s Five Forces
• How to analyze the competitive environment of a company?
• 5 forces = 5 determinants that affect industry profitability
• What is “industry”?
– Group of firms producing products of close substitutes
• Professor Porter has a PhD in economics. He applies industrial
organization (a branch of microeconomics) to competitive
strategies for business
• The stronger the pressure from each force, the closer profits to
zero economic profits
– Strong force: “threat”
– Weak force: “opportunity”
Porter’s Five Forces: Summary
•
Barrier To Enter (BTE) force: determines # of firms
•
Other four forces: determine costs and elasticities
83
Minicase:
• Discount retail giant
– Since 1962 (61 years)
– Largest company in the world by revenue: USD 538B in 2022
– Largest private employer: 2.3 million employees (1.7 mil in US)
in 2022
– 10,500 stores world wide in 2022
• McDonald’s: 37,855, 7-Eleven: 68,236
• Runs Sam’s Club (warehouse clubs)
BARRIERS TO ENTRY
(THREAT OF NEW ENTRANT)
Barriers To Entry (BTE): something that allow incumbents to avoid
competition with potential competitors
85
BTE: Economies of scale (EOS)
Economies of scale (i.e., decreasing Average Costs) due to large
fixed cost investment
Fixed cost 1. Distribution: Network of chain stores are more
efficient than single store. Distribution center => Develop outlets
Fixed cost 2. E-commerce and IT investments (next slide)
Conclusion: For Walmart, BTE is high
86
E-commerce & IT Investments
• Website & apps (navi & scan-bag-go)
• Curbside pickup (Buy online, pick up at store)
• Pickup tower (experimented but dropped)
• Training employees via VR glass & apps
• Teaming up with Microsoft’s Azure platform (as Kroger does)
87
E-commerce & IT Investments
• Actively acquiring IT startups in e-commerce, direct-to-consumer
apparel, & last mile logistics
Real-time pricing algorithm based on location
Indie and vintage-inspired women’s clothing
E-commerce package delivery
E-commerce apparel for men’s
88
RIVALRY
89
Rivalry = Product differentiation
Fierce competition can force prices downward toward costs,
thereby eroding profits
• Intense rivalry when product differentiation is low
– Homogeneous goods
– Low switching costs (e.g. little brand loyalty) or search cost
• Key: elasticities: the degree to which consumers change their
demand in response to price or income changes
– More elastic demand = more intense competition
• For Walmart: “High” as it sells commodities
– # and variety of large competitors: Target, Kmart, ..
– Fierce competition from small retail stores in clothing, food, etc..
• Way to mitigate this pressure? Private labels
90
SUBSTITUTES
91
Pressure from Substitutes
“People don’t want to buy a quarter-inch drill. They want a quarterinch hole!” Theodore Levitt
• Key: elasticities: More substitutes = more elastic
• For Walmart: High/Medium
– High factor: selling commodities so competition across
categories (e.g., convenience stores, supermarket chains)
– Medium factor: offering a wide range of service
General merchandise + supermarket + pharmacy + garden
center + tire + optics + hair salons + bank branches + …
=> Substitute to shopping at Walmart Supercenter?
92
BUYER / SUPPLIER POWER
93
Bargaining Power of Buyers
• Buyers = Consumers
• Buyer power: high when
– A few buyers consume a large fraction of industry output (NO)
– The industry’s products are undifferentiated and the buyers
face few switching costs (mostly YES)
– Good substitutes exist (YES)
• Key: elasticities (again!)
• For Walmart: Medium, but can mitigate via customer loyalty and
outlet location
94
Bargaining Power of Suppliers
• Suppliers: Wholesalers, Manufacturers
• Inverse of buyer power. Key is elasticities
• Supplier power: strong when
– Suppliers are concentrated (few and large)
and selling to more fragmented buyers
(NO)
– Products are differentiated and/or the
downstream industry faces high switching
costs (NO)
• For Walmart: Low. They can negotiate with
suppliers to buy in big volume
95
Summary of Walmart
• Pressure from potential entrants: Low (i.e., BTE is high)
• Pressure from rivals: High
• Pressure from substitutes: High/Medium
• Pressure from buyers: High/Medium
• Pressure from suppliers: Low
Takeaway: Industry analysis is a great tool for analyzing retail
industries’ & individual retailers’ source of profitability
=> should be part of the presentation if you pursue option (1):
proposing a new business in retailing
96
Critiques of Porter
• Other players are viewed as “threats”, but can make the potential
industry earnings bigger for everyone
• Static—lacks dynamic strategic interactions or learning or future
• Framework fits matured and concentrated markets with large
companies, but may not work well with new businesses
97
BU.450.740
Retail Analytics
Session 1: Introduction to Analytics
Prof. Mitsukuni Nishida
Johns Hopkins Carey Business School
1
DATA SCIENTIST’S TIPS
2
Three Skills as Data Scientist
Below are appreciated in labor market
1. Understand theory
– What could be the potential mechanisms between A and B?
2. Properly apply them through programming
– Which statistical method to use?
– Performance of the results? (e.g., accuracy of prediction,
statistical significance of the parameter estimates)
3. Interpret and present results
– Economic significance? How are convincing your findings?
Direction of causality? Business implications?
3
GIYF: Tip for success in programming…
Error in Print(2 * 3) : could not find function “Print”
Google
ChatGPT
could not find function
“Print“ RStudio
Would you analyze why I receive this error4
message in R?
A shortest path to cutting-edge is LBD
We will be quickly establishing our development environment & our
command in R
• To get you up-to-speed, I will show you what I think a shortest
path to analyze real-world analytics problems
– “Needs drive methods, not the other way around.”
• I call this approach “Issue driven” or “Kanban-flow”
– We skip some basic stuffs on R, such as data types, etc.
– Everything is LBD (leaning-by-doing) and self-teaching
Disclaimer
• This class is not about learning statistics or R; more about how we
apply statistical methods to marketing problem in retailing
• For concepts, please go back to stats class’ textbook and
references and resources online
5
Further Resources to Learn R
• Introduction to R seminar at UCLA idre
– https://stats.idre.ucla.edu/r/seminars/intro
• Johns Hopkins’s Data Science Course at Coursera
– https://www.coursera.org/specializations/jhu-data-science
• AI for economists by Jesse Lastunen
– https://sites.google.com/view/lastunen/ai-foreconomists?authuser=0
6
KEY CONCEPTS FOR
ANALYTICS
7
Correlation & Causation
Possible causation 1
Possible causation 2
X
Observed correlation
Possible causation 3
Z
Y
Possible causation 3
When X and Y are correlated, 3 possibilities exist
1. X => Y: Ad increase sales
8
2. Y => X: When sales are high, firms increase ad
3. Z => X and Z => Y: firm has been running sales in summer (for some reasons)
Correlation ≠ Causation
X
