SCANNER DATA AND ONLINE MARKET RESEARCH

1
SCANNER DATA AND ONLINE MARKET RESEARCH

• Scanner data
• Impact of Internet on market research
• Types of online research
• Challenges in online research
• Excel, Part II

2
Scanner Data

• Panel members in test communities agree to
• Swipe a card prior to each purchase
• Have purchases matched to
• demographic profiles
• media/coupon exposure
• promotional status of competing brands
• past purchases
• Problems
• Aggregation over household
• Aggregation bias--averages of disparate segments
  obscure!

3
Scanner Data Research
Purchase on occasion Yes, no Time since
previous purchase Previous purchases Current
price Previous price Current promotional
status Previous promotional status Current
display status Previous display status Display
status of competing brands Promotional status of
competing brands Coupon used Yes, no Coupon
available Yes, no Coupon available for other
brands? Yes, no Amount of coupon
No. of ads seen by shopper Ads seen for competing
brands Split cable
RECORDED PURCHASES
TELEVISION EXPOSURE
HOUSEHOLD FILE
DEMOGRAPHIC INFORMATION
ANALYSIS
Family size Occupation Family size Income Home
ownership
4
IMPACT OF INTERNET ON MARKET RESEARCH

• Information not previously collectible
• Speed and reduced cost
• Large scale data collection
• Access to new customer groups
• Observation of consumer interaction

5
Information Not Previously Observable

• Traditional questionnaires relied on questions
• Now possible to measure behaviore.g.,
• Time spent
• Number of products browsed

6
Possibilities That Overlap With Scanner Data

• Analysis of purchase history
• Assess impact of
• Price
• Price of competing product
• Featured space
• Promotion

7
Cost and Efficiency

• Reduced interviewer costs
• No travel
• Customer does the entry work
• Increased speed

8
Large-Scale Data Collection

• Lower cost means more cost effective
  opportunities to collect data
• Larger sample size reduces sampling errors
• Data can be briefly collected through online
  visits for other purposes

9
Access to New Customer Groups

• Access to
• Time constrained professionals
• Consumers in remote geographic areas
• Value of new customer groups
• Representativeness of general population studies

10
Observation of Interaction

• Examining word of mouth
• Visits to chat rooms
• Examination of postings on web sites and Usenet
  groups

11
Data Analysis Tools Data Mining

• Use of brute force power to find relationships
• Feasible because of the low cost of data
  processing
• Chance results When much data is examined,
  coincidence may stand out
• Need for replication through hold back samples

12
Data Analysis Other

• Test of hypotheses/assessments of magnitudes
• Logistic regression Flexible data analysis
  method requiring large data sets
• Real-time data
• Immediate feedback on current results
• Opportunities for
• Inventory adjustments
• Price/promotion adjustments
• Analysis of individual-level data

13
Types of Online Research
14
Online Focus Groups

• Methods
• Large set-up cost methods
• Teleconferencing
• Home-based web cam and microphone setup
• Lower cost
• Online chat format
• Impact of non-verbal communication

15
Search Log Analysis

• Unmet demandsearch for product not found on
  siste
• Brand identitysearch for discontinued brand
• Message comprehensioncomparison of search terms
  to media message
• Consumer vocabulary
• Feedback analysis
• Forecasting opportunities based on seasonal
  changes

16
Surveys

• Conditional branchingdirect skip to relevant
  question
• Quality of response
• Time pressures
• Willingness to write out answers or respond to
  multiple closed-ended questions
• Willingness to read and follow instructions
• Reliability and browser compatibility issues

17
Click-Stream Analysis

• Analysis of clicking pathhow does the consumer
  get to a desired page or product?
• Advertising as traffic generator evaluation
• Shopping cart analysis

18
Online Market Research Concerns

• Representativeness of
• Populationare relevant groups reached in desired
  proportions?
• Sampleeven if the desired population is reached,
  do respondents respond in desired proportions?
• Consumer acceptance of researchis this regarded
  as spam?
• Panel recruitment
• Privacy

19
Excel, Part II

• Graphs
• Line
• Bar
• Pie
• Copying graphs to other applications
• Scenario analysis
• Optimization