Transcript MAR_6648_Lecture_4_Observation_Notes

Observation and Archival
Techniques
MAR 6648: Marketing Research
January 13, 2010
Overview
• What is observational research?
– What can it do? What can’t it do?
• What is archival research?
– What can it do? What can’t it do?
• What are modern methods of each?
The Marketing Research Process
Goal formulation
• Define research purpose and objectives
Research design
• Exploratory, descriptive, causal
Data collection
• Databases, survey design, sampling, target respondents
Data analysis and interpretation
Final report
• Communicate findings and recommendations
Observational Research
• Infant mortality and illness
rates stopped improving
• Doctors thought they had
simply hit the limit
• Dr. Virginia Apgar
suggested that they simply
report the physical
condition of newborns
• Adoption of the score
revealed variation in
hospital performance,
identified at-risk babies,
and improved infant
health
Observation and Grocery Stores
• Why are frequently bought items (milk, bread) in the back?
– How far shoppers walk (not time spend)
– Greater incidental purchases
• Why the grocery cart?
– Often determines the amount you shop
• Begin with the produce or bakery section
– Shoppers who are physiologically sensitized buy more
• Store brands are placed right of Name brands
– Easier to reach (most people are right handed)
• High-profit items on right side of aisle
– Predominant traffic direction
• Limit # per shopper
– People tend to buy the limit, or at least buy more
• Froot Loops, Cocoa Puffs, etc., on lower shelf than Shredded
Wheat, etc.
– Children’s eye level
Why We Buy
• Taking anthropology
into the store
• Hidden cameras,
stopwatches,
sophisticated data
analysis…
• Most important: The
Tracking Sheet
– Simply a structured way
to record what people
actually do
Why We Buy
• What can be learned?
– Shoppers go to the right
– The “butt-brush” effect
– Who actually buys the dog biscuits?
– When do people read signs?
– “Conversion Rate”
Ethnography
• From anthropology
– Immerse yourself in a culture
• Usually done by trained anthropologists or
sociologists
– Requires extended videotaping, sophisticated
coding and editing
• Objectives
– Direct observation of customer
– Experience ins and outs of customers’ daily lives
Ethnography
• Good for…
– Noting nonstandard uses
– Understanding how consumers interact with their
environment
• In complex situations
• In low involvement situations
– Figuring out unarticulated needs
Observational Research
• Conducting observational research:
– Keep your goals in mind: Descriptive vs. Inferential
vs. Evaluative
– Identify a behavior (or behaviors) you want to
study
– Define them extraordinarily carefully
– Determine the unit of analysis (e.g., person, time,
product, etc.)
– Unobtrusively count
• Ideally, check your count against someone else’s
Modernizing Observation
• Measuring media consumption
• Radio moves to the web, the web moves to TV,
and TV to VOD, cellphones, and iPods
– What is the right measure?
•Originally people
completed paper diaries
•Then the “People
Meter”—an electronic
device attached to the TV
set
•Portable People Meters
•Small device clipped to belt
•Measures radio and TV usage
•Automatically uploads data to
the central office at the end of
the day
Arbitron’s Portable People Meters
• Radio stations encoded
with a unique signal
identifiable by the PPM
– 10 years, $80 million
• The PPM captures any
media exposure in the
vicinity of the wearer
– What are the shortcomings
of this measure?
• Consider adding GPS to
identify proximity to
billboards
PPM in Action
PPM in Action
• AFC Championship Game: NY Jets vs. Indianapolis
Colts (Jan. 31, 2010)
– 40,000 panelists
– Findings:
• Many men watched the game away from home.
• So do women.
• The game was particularly popular in New York.
– Are these novel insights?
– Are they actionable?
– What other information should Arbitron try to
observe?
Consumer Generated Media
• “Consumer Generated Media” (CGM)
encompasses the millions of consumer-generated
comments, opinions, and personal experiences
posted in publicly available online sources on a
wide range of issues, topics, products, and
brands”
• Why care about CGM?
– Marketers can easily lose control of the message
– Consumers can talk to each other at an
unprecendented speed
Case: Kryptonite
Case: Kryptonite
Case: FedEx Furniture
Case: FedEx Furniture
Keep in Mind…
• The Streisand Effect
• “The Net interprets censorship as damage and
routes around it.”
Advantages of Observational Research
•
•
•
•
Conclusions are externally valid
Inferences are broadly unassailable
Data collection is (generally) inexpensive
Plausibly unbiased
Disadvantages of Observational Research
•
•
•
•
All but impossible to discern cause
Limited ability to manipulate critical variables
Concerns about reactivity
Can only measure overt behavior
What Can Archival Research Do?
• Do people become more aggressive the hotter
it gets?
– Police and weather reports
– Rioting and high temperature
– Hit by pitch in baseball
Archival Research
• General well-being and trends
– Content analysis
• Song lyrics, blogs, etc.
• ngrams
Industrial Archival Data
• Available, focused, but often expensive
• Different industries collect different info in
different ways
– Packaged goods
– Retail
– Direct Mail
– E-commerce
– Entertainment
Packaged Goods
• Household scanner data
– Consumer behavior
– Trends
• POS scanner data
• Nielsen BASES
– Product development
– Line extensions
– E-Concept testing
Retail
• Loyalty card databases
– Frequent shopper programs
– Allows the retailer to track households’ in-store
behavior over time
– Uses:
• Who are my most valuable customers?
• Optimal coupon design and targeting
• Effects of competition—who do I lose?
Case: Wal-Mart vs. …
• Frequent shopper data from a retailer covering about
2 years
• A Wal-Mart Supercenter entered the market about
halfway through the period
Daily Store Sales for local retailer
1
75000
$ Sales
65000
55000
45000
35000
25000
15000
• Why the dip?
0
1
41
81 121 161 201 241 281 321 361 401 441 481 521 561 601
Day
– Fewer customers, different customers, or fewer
purchases?
– Loyalty card data can answer the question
Direct Mail
• Catalog companies
– “The house file”: Database on past mailings and
purchase history for each customer
• Non-profit solicitations
– “The house file”: Database on past contacts and
donation history
• Uses:
– Design optimal contact strategies
– Customer Lifetime Value calculations
E-commerce
• comScore Networks
– This is similar to Nielsen/IRI home scanner
panels—only for the web
– Monitor web activity for a large sample of
households
• Clickstream
• Online purchase behavior
Entertainment Industry
• Companies in the media and entertainment industry
operate in “fast” markets:
– Lifespan of product is short
– Consumers are constantly exposed to and are trying new
products
– Need to know: “Where is the market now and where is it
going?”
• Traditional and modern approaches
– Traditional: Nielsen TV ratings, BookScan, SoundScan, etc.
– Modern: iTunes sales, Big Champagne, Buzzmetrics, CGM
Summary
• Observational research can let you know what
happens out in the world
– Our beliefs and the data often don’t match;
observation can tell you when, and give you ideas
for what to do
• Archival research can give you data on smallscale and large-scale trends