MARK3054 Lecture 6: MARK3054 Topic 6

39 views2 pages
1 Jun 2018
Department
Course
Professor
Topic 6: Customer Heterogeneity and Segmentation Part 2
Customer Heterogeneity Analysis: modelling the dependence between needs and descriptors (t-test,
ANOVA, linear regression direct impact, moderation)
Cluster Analysis: modelling the interdependence among needs. A convenient method used to
categorise entities into groups that are homogenous along a range of observed characteristics (when
clustering, make circles round, not oval)
Lecture Example: To whom shall we sell this new PDA?
Research Questions:
1. What segments exist? (Features of each segment)
2. Who are the consumers in each segment?
3. Which segment is most attractive?
Solution: Conduct cluster analysis
1. Formulate the problem
2. Select a distance measure (recommended: Euclidean)
3. Conduct hierarchical cluster analysis
4. Conduct k-means cluster analysis (more precise than hierarchical)
5. Interpret and profile clusters
Euclidean distance: works by measuring the distance between 2 respondents over all the questions
or (n amount of dimensions)
When determining number of clusters, draw a horizontal line and determine how many vertical lines
cross it OR via an SAS table, look at where there is a large jump in RMS distance, then go up a
cluster. CLUSTERS SHOULD BE MEANINGFUL!
SSB/(SSW+SSB) = ratio of between group variance to total variance (similar to r2 and is called
approximate expected overall r-squared in SAS).
K-means cluster analysis: a method which aims to partition observations into clusters by assigning an
observation to the closest mean.
1. Select input (we know how many clusters we want, thus assign values to closest centre, then
update cluster centres, then reassign until convergence)
Convergence= when you update cluster centres and reassign, but they do not improve current
results. What is good? Minimise within group variance and maximise between group variance
Interpreting and Profiling Clusters
Look at which group has the highest and lowest cluster average scores for each question.
Then, go along each cluster and start making a customer profile based on the cluster
averages for each question.
THEORY: Determining which segments to serve:
Using Criterion
1. Size and growth
2. Structural characteristics (competition, segment saturation, protectability, environmental
risk)
3. Product/market fit (fit, relationships with existing segments, profitability)
GE portfolio matrix: against competitive strength and market attractiveness. Portfolio analysis also
suggests what steps to take.
Unlock document

This preview shows half of the first page of the document.
Unlock all 2 pages and 3 million more documents.

Already have an account? Log in

Document Summary

Topic 6: customer heterogeneity and segmentation part 2. Customer heterogeneity analysis: modelling the dependence between needs and descriptors (t-test, A convenient method used to categorise entities into groups that are homogenous along a range of observed characteristics (when clustering, make circles round, not oval) Solution: conduct cluster analysis: formulate the problem, select a distance measure (recommended: euclidean, conduct hierarchical cluster analysis, conduct k-means cluster analysis (more precise than hierarchical) Euclidean distance: works by measuring the distance between 2 respondents over all the questions or (n amount of dimensions) When determining number of clusters, draw a horizontal line and determine how many vertical lines cross it or via an sas table, look at where there is a large jump in rms distance, then go up a cluster. Ssb/(ssw+ssb) = ratio of between group variance to total variance (similar to r2 and is called approximate expected overall r-squared in sas).

Get access

Grade+20% off
$8 USD/m$10 USD/m
Billed $96 USD annually
Grade+
Homework Help
Study Guides
Textbook Solutions
Class Notes
Textbook Notes
Booster Class
40 Verified Answers
Class+
$8 USD/m
Billed $96 USD annually
Class+
Homework Help
Study Guides
Textbook Solutions
Class Notes
Textbook Notes
Booster Class
30 Verified Answers