Sie sind auf Seite 1von 8

SVKMs Narsee Monjee Institute of Management Studies, Mumbai

MARKET RESEARCH
Assignment

DURR: Clustering and Conjoint Analysis



Date: 10 July 2014

Submitted By:
Sujatha Gnanasekher (D018)
Sarthak Mohanty (D039)
Tushar Nayanam (E034)
Vineet Jeyendranath (E062)






Overview
Durr Environment Controls is a German conglomerate which had existing industrial operations
in the US. It was considering an option to launch one more product i.e. Air Emission Control
System. It had existing well structured distribution system in the US which was an edge for
them to grow their business.
Market & Competition
The firm found that there were four main dimensions which govern the supplier selections.
a) Efficiency b) Delivery Time c) Price d) Delivery Terms
There were already three existing players in the market with following offerings
Competitor Efficiency Delivery Time Price Delivery Terms
Waste Watch Exceed by 5% 9 Months $800K FOB Seller
Thermatrix Exceed by 9% 9 Months $900K FOB Seller
Advanced Air Meets EPA
Target
6 Months $700K Installed, Service
Contract
Customer survey and study
To find out the viability of the idea the firm conducted few surveys and a conjoint analysis study
to find out the weightage of the factors involved and the tradeoffs that the customers make in
selecting the suppliers. A proper conjoint study was made from the data collected by 31
companies. It was designed to help them identify the target for the new offerings and position
them for the particular segment.
Efficiency Delivery Time Price Delivery Items
Exceed by 9%(99%) 6 months $600k Installed, 2 yr
Warranty
Exceed by 5%(95%) 9 months $700k Installed, 1 yr
Warranty
Meets EPA
Target(90%)
12 months $800k Installed, Service
Contract
Short by 5%(85%) 15 months $900k FOB Seller, Service
Contract
Each firm filled out one such form using a group decision process involving key purchase
influencers for these products. Based on all the discussions between product and Marketing
team the firm had these three options to look for.
Model Efficiency Delivery Time Price Terms
Servair_DX Meet Specs 15 Months $900,000 Installed with 2 yr warranty
Premier LX Exceed specs by 9% 12 Months $900,000 Installed with service
agreement
Base Product Exceed specs by 5% 12 Months $700,000 Installed with service
agreement
Analysis
Based on the survey data a hierarchical cluster analysis was carried out to find out the number
of clusters the sample market should be divided into. The results are as follows:
The following table shows the agglomeration schedule:

Agglomeration Schedule
Stage Cluster Combined Coefficients Stage Cluster First Appears Next Stage
Cluster 1 Cluster 2 Cluster 1 Cluster 2
1 11 12 195.000 0 0 10
2 10 22 201.000 0 0 18
3 24 26 203.000 0 0 13
4 29 30 235.000 0 0 9
5 1 9 248.000 0 0 12
6 23 25 312.000 0 0 17
7 4 27 350.000 0 0 14
8 14 16 426.000 0 0 14
9 7 29 429.500 0 4 16
10 11 20 442.500 1 0 20
11 6 28 445.000 0 0 15
12 1 15 464.000 5 0 25
13 24 31 469.500 3 0 20
14 4 14 557.500 7 8 21
15 6 18 570.500 11 0 23
16 3 7 637.667 0 9 19
17 17 23 664.000 0 6 21
18 5 10 686.500 0 2 24
19 3 13 763.250 16 0 28
20 11 24 767.667 10 13 23
21 4 17 788.833 14 17 24
22 2 19 923.000 0 0 26
23 6 11 935.278 15 20 26
24 4 5 1077.762 21 18 27
25 1 21 1312.000 12 0 28
26 2 6 1601.167 22 23 27
27 2 4 1869.645 26 24 29
28 1 3 2138.700 25 19 30
29 2 8 3132.429 27 0 30
30 1 2 3462.419 28 29 0
Theres a stark change in coefficients between stages 28 and 29 which implies the set of two
groups together has changed the dynamic of the group since they are dissimilar. Hence we
fixed on the number of clusters as 3. Using this output we perform a K-Means clustering
analysis to segment the market into three different clusters with similar preferences each.
Below are the number of cases in each cluster and the list of companies/cases under each
cluster.
Number of Cases in each Cluster
Cluster
1 12.000
2 11.000
3 8.000
Valid 31.000
Missing .000
Now that we have arrived at the number of clusters, we perform a conjoint study for each
cluster for a set of 256 possible product offerings arising from 16 attributes. To do the conjoint
study, the cumulative mean scores (K-mean) of each of the 4 attributes for each of the 256
product offerings was calculated. Then they were ranked after sorting this score from largest to
smallest. This rank formed the dependent variable while the 16 attributes formed the
independent variables. A regression analysis was carried out for each cluster and the beta
coefficients for each attribute in each cluster were determined. The utility values can then be
used to calculate the tradeoffs for the customer and thus determine how each cluster values an
attribute. Thus once the order of preference of the attributes for each cluster have been
established a customized product offering for each cluster can be designed in such a way that it
Cluster 1 Cluster 2 Cluster 3
GenMills TXU Chev
KrispyKreme Dean AEP
Pepsi Molson Duke
Mohawk BrownForman EnergyEast
Fab Xerium PG&E
AveryDenn Champion GAPac
SealedAir ConGraph Bridgestone
Citgo Monsanto UTC
HollyCorp Lubrizol
Coffeyville Nucor
Scotts Timken
Carlisle
maximizes utility for them, thus enhancing the value of the product offering. Based on the beta
coefficients from regression analysis, the alpha value or the utility was calculated. The utility
values thus calculated for each cluster are given below.

Attribute Utility
Cluster 1 Cluster 2 Cluster 3
Exceeds 9% 40.52 73.88 33.38
Exceeds 5% 7.37 32.38 -0.65
Meets specifications -12.38 -30.16 -8.92
Short by 5% -35.51 -76.10 -23.81
6 months 79.48 52.82 33.93
9 months 27.46 15.24 4.41
12 months -33.41 -16.96 -9.34
15 months -73.52 -51.10 -29
600000 17.6 27.05 64.47
700000 5.37 6.35 30.89
800000 -5.99 -9.99 -22.76
900000 -16.98 -23.41 -72.61
Installed, with 2-year warranty 40.8 20.02 56.66
Installed, with 1-year warranty 19.05 6.12 18.96
Installed, with service contract -18.73 -4.49 -28.61
FOB, with service contract -41.13 -21.65 -47.01

Recommendation Based on the Companies Proposed Products
Customer trade off is calculated based on the difference between attributes with the highest
and lowest utility. Based on the above analysis, a model is recommended for each of the
clusters according to their attribute preference as given in the below table.
Attribute Cluster 1 Cluster 2 Cluster 3
Efficiency 76.03 149.98 57.19
Delivery Time 153 103.92 62.93
Price 34.58 50.46 137.08
Delivery Terms 81.93 41.67 103.67
Attribute
Preference
Time > Terms >
Efficiency > Price
Efficiency > Time >
Price > Terms
Price > Terms > Time >
Efficiency
Recommended
Model
Premier LX Premier LX Base Product
Delivery Time = Time; Delivery Terms = Terms; Efficiency= Efficiency; Price=Price
Recommendation Based on the Competitors Products
Cluster 1
Delivery Time > Delivery Terms > Efficiency > Price
So cluster 1 is the least price sensitive of the lot but they need quick delivery and excellent
service which means that they would not mind paying a premium if those two attributes are
satisfied. With respect to delivery time Advanced Air has the least delivery time of 6 months.
Here additional utility required to charge a premium of $200,000 is 23 units. Hence based on
the utility values and the trade off values the following product offering is suggested for Cluster
1. The offering is referred Cluster 1 customization as described in the below table.

CLUSTER 1 - Customization based on Competition, Utility and Profit
Advanced Air Cluster 1 customization
Attribute
Attribute Level
Utility
Attribute Level
Utility
Additional
Revenue
($)
Time 6 months 79.48 6 months 79.48 -40000
Terms Installed, with service
contract -18.7
Installed, with 1-year
warranty 19.05 -100000
Efficiency Meets specifications -12.4 Exceeds 5% 7.37 0
Price 700000 5.37 900000 -17 200000
Total

53.74

88.92 60000

In this case by offering the above mentioned attribute level for cluster 1 cases, Durr can achieve
the following advantages:
Additional utility for the customers
Competitive advantage
Additional revenue of $60,000
Cluster 2
Efficiency > Delivery Time > Price > Delivery Terms
Cluster 2 considers efficiency and delivery time as the most important attributes. Price and
delivery terms are the next important ones with price edging delivery term slightly. Since
delivery term is the least priority we can provide FOB service while providing the highest
efficiency and delivery time. With respect to efficiency Thermatrix exceeds 9%. Hence based on
the utility values and the trade off values the following product offering is suggested for Cluster
2. The offering is referred Cluster 2 customization as described in the below table.

CLUSTER 2 - Customization based on Competition, Utility and Profit

Thermatrix Cluster 2 customization
Attribute
Attribute Level Utilit
y
Attribute Level Utilit
y
Additional
Revenue ($)
Efficienc
y
Exceeds 9%
73.88
Exceeds 9%
73.88 -170000
Time 9 months 15.24 6 months 52.82 -40000
Price 900000 -23.4 900000 -23.4 200000
Terms FOB, with service
contract -21.7
FOB, with service
contract -21.7 100000
Total

44.06

81.64 90000

In this case by offering the above mentioned attribute level for cluster 2 cases, Durr can achieve
the following advantages:
Additional utility for the customers
Competitive advantage
Additional revenue of $90,000
Cluster 3
Price > Delivery Terms > Delivery Time > Efficiency
Cluster 3 is highly price conscious and at the same time wants high service levels too. Its not
that conscious about delivery time and efficiency. With respect to price Advanced Air has the
least price of $700,000. Hence based on the utility values and the trade off values the following
product offering is suggested for Cluster 3. The offering is referred Cluster 3 customization as
described in the below table.

CLUSTER 3 - Customization based on Competition, Utility and Profit

Advanced Air Cluster 3 customization
Attribute
Attribute Level Utilit
y
Attribute Level Utilit
y
Additional
Revenue($)
Price 700000 30.89 700000 30.89 0
Terms Installed, with service
contract
-28.6 Installed, with service
contract
-28.6
0
Time 6 months 33.93 6 months 33.93 -40000
Efficiency Meets specifications -8.92 Meets specifications -8.92 70000
Total

27.29

27.29 30000

In this case by offering the above mentioned attribute level for cluster 3 cases, Durr can achieve
the following advantages:
Can be on par with the competitors offering. Hence less chances of losing this customer
segment/cluster
Additional revenue of $30,000
Conclusion
Thus after creating the product offerings for each cluster, it can be concluded that Clusters 1
and 2 look the most attractive in terms of both earning revenues and providing value to the
customer.

Das könnte Ihnen auch gefallen