BOEK345 2 Student

Marketing Analytics (BOEK 345)
P R O F. R O B E R T K R E U Z B A U E R
U N I V E R S I T Y O F S TAVA N G E R B U S I N E S S S C H O O L
Good old-fashioned Price-Value Ratios
Price
Value
Don’t think like a consumer
Happy 
A lot of cool
advertising
Company
Products
Think like a manager!
Costs/Benefits
Benefit Costs
(thirst-quenching, experience, etc.)
Marketing Mix = 4Ps (Product, Price, Promotion, Place (i.e. Distribution)
4Ps + STP
WTP (Willingness To Pay)
Customers
A marketing strategy is an integrated set of choices about how we

will create and capture value, over long periods of time.
Collaborators Resources
4Ps + STP
WTP
Customers
Successful brands have successful competitors
4Ps + STP
Competitor(s)
WTP
Customers
4Ps + STP
Competitor(s)
WTP
Customers
Technology
4Ps + STP
DYNAMIC
Competitor(s)
WTP
Customers
Four Cs of marketing strategy
1. Customer Value: What your customers want?
2. Competition: How your competitors react?
3. Collaborators: Will your collaborators collaborate? (benefits

for them ?)
4. Costs: Do I have the money for all this?

Customer Segmentation & Targeting
1. Goal: Company (Netflix) wants to understand what are their

main consumer segments
2. Let’s assume we have only two kinds of information

(variables): Age & Days per week spent to watch Netflix
videos
Age: 42 Age: 18 Age: 23 Age: 49 Age: 37 Age: 51 Age: 40 Age: 20

Time: 7 Time: 3 Time: 2 Time: 1 Time: 7 Time: 1 Time: 6 Time: 4
Source: Serrano, 2019

Days per week Segment 2

watching Netflix
7
Age: 37 Age: 42
6 Time: 7 Time: 7
Segment 1
5 Age: 40
Time: 6
4
3 Age: 20 Segment 3
Time: 4
2
Age: 18 Age: 23
1 Time: 3 Time: 2
Age: 49 Age: 51
Time: 1 Time: 1 Age
20 30 40 50
v
X
v
v
v
X
v
v
v
X
v
v
v
X
v
v
v
X
v
v
v
X
• How do we know the right number and size of segments

(clusters)
 Cluster analysis:
 K-Means clustering method
 Hierarchical clustering method
K-means clustering method
Assume you want to place 3 ice-cream shops in a city
v
v

v
v

v
v

v
v

v
v

v
v

• How do we know how many clusters?
 Elbow-method
1 cluster 2 cluster 3 cluster
v v v v v v
v v v
4 cluster 5 cluster 6 cluster
v v v v v v
v v v

Distance 3 cluster
v v
Clusters
1 2 3 4 5 6
• How do we know the right number and size of segments

(clusters)
 Cluster analysis:
 K-Means clustering method
 Hierarchical clustering method
Hierarchical clustering method













• How do we know how many clusters?
 Dendrogram-method
Dendrogram
Distance
8
v
3 6 7
1
2
4
5
1 2 3 4 5 6 7 8
Dendrogram
Distance
8
v
3 6 7
1
2
4
5
1 2 3 4 5 6 7 8
Dendrogram
Distance
8
v
3 6 7
1
2
4
5
1 2 3 4 5 6 7 8
Dendrogram
Distance
8
v
3 6 7
1
2
4
5
1 2 3 4 5 6 7 8
Dendrogram
Distance
8
v
3 6 7
1
2
4
5
1 2 3 4 5 6 7 8
Dendrogram
Distance
8
v
3 6 7
1
2
4
5
1 2 3 4 5 6 7 8
Dendrogram
Distance
8
v
3 6 7
1
2
4
5
1 2 3 4 5 6 7 8
Dendrogram
Distance
8
v
3 6 7
1
2
4
5
1 2 3 4 5 6 7 8
Dendrogram
Distance
8
v
3 6 7
1
2
4
5
1 2 3 4 5 6 7 8
Dendrogram
Distance
8
v
3 6 7
1
2
4
5
1 2 3 4 5 6 7 8
Dendrogram
Distance
8
v
3 6 7
1
2
4
5
1 2 3 4 5 6 7 8
Dendrogram
Distance
8
v
3 6 7
1
2
4
5
1 2 3 4 5 6 7 8
Dendrogram
Distance
8
v
3 6 7
1
2
4
5
1 2 3 4 5 6 7 8
• What if you have many variables?
A B C D E F G H
Age: 42 Age: 18 Age: 23 Age: 49 Age: 37 Age: 51 Age: 40 Age: 20

Time: 7 Time: 3 Time: 2 Time: 1 Time: 7 Time: 1 Time: 6 Time: 4
Male Female Female Male Female Male Male Male
Step 1: Pick random centroids (represented by the stars).
v
v
v
Step 1: Pick random centroids (represented by the stars).
Buyers income age consume

1 4.40 4.57 2.29
2 3.25 3.92 2.17
3 3.10 4.25 2.40
4 4.83 4.31 2.16
5 v
3.63 3.60 1.67
Centroid1 (random) 6 3.26 1.64 1.48 v
7 4.89 1.33 1.04
8 4.50 2.01 1.28
Centroid2 (random) 9 4.99 2.47 2.60
10 4.12 2.12 1.70
11 2.21 2.51 4.17
12 2.97 4.10 3.92
13 2.40 2.45 4.46 v
14 2.10 3.30 4.90
Centroid3 (random) 15 1.13 2.05 3.28
Source: Fripp MktStudyGuide, 2016

Step 2: Measure distances between each centroid and each case
v
v
v
Buyers income age consume

1 4.40 4.57 2.29
2 3.25 3.92 2.17
3 3.10 4.25 2.40
4 4.83 4.31 2.16
5 v
3.63 3.60 1.67
Centroid1 (random) 6 3.26 1.64 1.48 v
7 4.89 1.33 1.04
Sum8 of Squared
4.50 Distances (Demo
2.01 calculation
1.28 with Buyer 1 and Centroid 6
Centroid2 (random) 9 4.99 Buyers
2.47 income
2.60 age consume
10 4.12 2.12 11.70 4.40 4.57 2.29
Centroid
11 1 (random)
2.21 2.51 64.17 3.26 1.64 1.48
12 2.97 Difference
4.10 3.92 1.14 2.93 0.81
13 2.40 Squared 4.46
2.45 v 1.30 8.58 0.66 10.54
14 2.10 3.30 4.90
Centroid3 (random) 15 1.13 2.05 3.28

Centroid1 Centroid2 Centroid3

Buyers (random) (random) (random)
1 10.54 4.85 18.02
2 5.67 5.32 9.22
3 7.68 6.78 v 9.50 Sum of Squared Distances (Demo calculation with Buyer 1 and Centroid 6
4 10.06 3.60 20.05 Buyers income v
age consume
5 4.01 3.99 11.24 1 4.40 4.57 2.29
6 0.00 4.94 7.95 Centroid 1 (random) 6 3.26 1.64 1.48
7 2.95 3.74 19.67 Difference 1.14 2.93 0.81
8 1.71 2.19 15.36 Squared 1.30 8.58 0.66 10.54
9 4.94 0.00 15.54
10 1.02 1.69 11.44
11 9.10 10.19 2.17 v
12 12.09 8.48 8.00
13 10.28 10.17 3.17
14 15.80 14.33 5.13
15 7.95 15.54 0.00

v
v


Step 3: Allocate cases with smallest distance to respective centroids (as shown below)
Centroid1 Centroid2 Centroid3 Member

Buyers (random) (random) (random) Min of cluster
1 10.54 4.85 18.02 4.85 2
2 5.67 5.32 v 9.22 5.32 2
3 7.68 6.78 9.50 6.78 2 v
4 10.06 3.60 20.05 3.60 2
5 4.01 3.99 11.24 3.99 2
6 0.00 4.94 7.95 0.00 1
7 2.95 3.74 19.67 2.95 1
8 1.71 2.19 15.36 1.71 1
9 4.94 0.00 15.54 0.00 2
10 1.02 1.69 11.44 1.02 1 v
11 9.10 10.19 2.17 2.17 3
12 12.09 8.48 8.00 8.00 3
13 10.28 10.17 3.17 3.17 3
14 15.80 14.33 5.13 5.13 3
15 7.95 15.54 0.00 0.00 3

Step 4: Re-position the centroids to minimize distance between each centroid

and its respective cluster-members (distances measured through Euclidian
distance algorithm)
v
v
v

distance algorithm)
v
v
v

distance algorithm)
Cluster 1 Cluster 2 Cluster 3
Member of cluster Buyers income age consume income age consume income age consume
2 1 4.40 4.57 2.29
2 2 3.25 3.92 2.17
2 3 3.10 4.25 2.40
2 4
v 4.83 4.31 2.16
2 5 3.63 3.60 1.67 v
1 6 3.26 1.64 1.48
1 7 4.89 1.33 1.04
1 8 4.50 2.01 1.28
2 9 4.99 2.47 2.60
1 10 4.12 2.12 1.70
3 11 2.21 2.51 4.17
3 12 v 2.97 4.10 3.92
3 13 2.40 2.45 4.46
3 14 2.10 3.30 4.90
3 15 1.13 2.05 3.28
Mean 4.19 1.78 1.38 4.03 3.85 2.22 2.16 2.88 4.15


distance algorithm) Member
of cluster
Buyers Mean cluster1 Mean cluster2 Mean cluster3 Min (revised)
1 8.69 0.65 11.30 0.65 2
2 6.12 0.62 6.17 0.62 2
3 8.37 1.06
Buyers 5.80
income age consume2
1.06
4 7.45 v 0.85 1 13.10
4.40 0.85
4.57 2.292
5 3.73 0.52
Mean cluster1 8.80
4.19 0.52
1.78 1.382 v
6 0.90 6.04
Difference 9.86
0.21 0.90
2.80 0.921
7 0.80 8.48
Squared 19.50
0.04 0.80
7.81 0.841 8.69
8 0.16 4.49 14.44 0.16 1
9 2.62 2.98 10.56 2.62 1
10 0.23 3.28 10.40 0.23 1
11 12.28 8.95 0.14 0.14 3
12 13.38 4.10 2.19 v2.19 3
13 13.19 9.68 0.34 0.34 3
14 19.13 11.25 0.75 0.75 3
15 13.08 12.82 2.51 2.51 3
SSE 14.34


distance algorithm)
v
v
v
Repeat steps 2-4, that is, re-calculate distances between each (now re-
positioned) centroid and each case and re-allocate cases with smallest distance
to respective centroids
v
v
v
Final result from fourth round of repetition of steps 2-4
v
v
v

distance algorithm) Member Member
of cluster of cluster
Buyers Mean cluster1 Mean cluster2 Mean cluster3 Min (revised) (previous)
1 8.69 0.65 11.30 0.65 2 2
2 6.12 0.62 6.17 0.62 2 2
3 8.37 1.06 5.80 1.06 2 2
4 7.45 v 0.85 13.10 0.85 2 2
5 3.73 0.52 8.80 0.52 2 v2
6 0.90 6.04 9.86 0.90 1 1
7 0.80 8.48 19.50 0.80 1 1
8 0.16 4.49 14.44 0.16 1 1
9 2.62 2.98 10.56 2.62 1 2
10 0.23 3.28 10.40 0.23 1 1
11 12.28 8.95 0.14 0.14 3 3
12 13.38 4.10 2.19 v2.19 3 3
13 13.19 9.68 0.34 0.34 3 3
14 19.13 11.25 0.75 0.75 3 3
15 13.08 12.82 2.51 2.51 3 3
SSE 14.34

BOEK345 2 Student

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

BOEK345 2 Student

Uploaded by

Copyright:

Available Formats

Marketing Analytics (BOEK 345)

WTP (Willingness To Pay)

A marketing strategy is an integrated set of choices about how we

1. Customer Value: What your customers want?

2. Competition: How your competitors react?

3. Collaborators: Will your collaborators collaborate? (benefits

4. Costs: Do I have the money for all this?

1. Goal: Company (Netflix) wants to understand what are their

2. Let’s assume we have only two kinds of information

Age: 42 Age: 18 Age: 23 Age: 49 Age: 37 Age: 51 Age: 40 Age: 20

Source: Serrano, 2019

Days per week Segment 2

• How do we know the right number and size of segments

Assume you want to place 3 ice-cream shops in a city

Source: Serrano, 2019

Assume you want to place 3 ice-cream shops in a city

Source: Serrano, 2019

Assume you want to place 3 ice-cream shops in a city

Source: Serrano, 2019

Assume you want to place 3 ice-cream shops in a city

Source: Serrano, 2019

Assume you want to place 3 ice-cream shops in a city

Source: Serrano, 2019

Assume you want to place 3 ice-cream shops in a city

Source: Serrano, 2019

• How do we know how many clusters?

1 cluster 2 cluster 3 cluster

4 cluster 5 cluster 6 cluster

Source: Serrano, 2019

• How do we know the right number and size of segments

Assume you want to place 3 ice-cream shops in a city

Source: Serrano, 2019

Assume you want to place 3 ice-cream shops in a city

Source: Serrano, 2019

Assume you want to place 3 ice-cream shops in a city

Source: Serrano, 2019

Assume you want to place 3 ice-cream shops in a city

Source: Serrano, 2019

Assume you want to place 3 ice-cream shops in a city

Source: Serrano, 2019

Assume you want to place 3 ice-cream shops in a city

Source: Serrano, 2019

Assume you want to place 3 ice-cream shops in a city

Source: Serrano, 2019

Assume you want to place 3 ice-cream shops in a city

Source: Serrano, 2019

Assume you want to place 3 ice-cream shops in a city

Source: Serrano, 2019

Assume you want to place 3 ice-cream shops in a city

Source: Serrano, 2019

Assume you want to place 3 ice-cream shops in a city

Source: Serrano, 2019

Assume you want to place 3 ice-cream shops in a city

Source: Serrano, 2019

Assume you want to place 3 ice-cream shops in a city

Source: Serrano, 2019

• How do we know how many clusters?

• What if you have many variables?

Age: 42 Age: 18 Age: 23 Age: 49 Age: 37 Age: 51 Age: 40 Age: 20

Step 1: Pick random centroids (represented by the stars).

Step 1: Pick random centroids (represented by the stars).

Buyers income age consume

Source: Fripp MktStudyGuide, 2016

Step 2: Measure distances between each centroid and each case