Vous êtes sur la page 1sur 6

Segmentation

For the purpose of segmentation, we collected demographic data from all the
respondent to help us identify difference in need of different segments of population.
For clustering the data, we ran K-MEAN clustering technique on the attribute
data collected from the various respondent. No of cluster (i.e. the value of k) was adjusted to
ensure that cluster represent a meaningful set of attributes. Weighted distance of each cluster
from its centroid was calculated to find the most optimum cluster.
We found out that the data collected can be explained by 4 cluster as they
provide the lowest weighted distance. Description of Cluster are as follows

Cluster 1
Description: 21-35 Unmarried Students (Include PhDs, Masters Student majorly)
Cluster 2
Description: 21-35 Unmarried Working
Cluster 3
Description: 16-21 Unmarried Student (Majority Undergraduate Student)
Cluster 4
Description: 21-34 Married Working

Cluster Name Weighted Distance from Centroid


Cluster 1 0.85
Cluster 2 1.34
Cluster 3 0.94
Cluster 4 0.85

Regression Analysis
To evaluate how much customers value different attribute of the VoD
services we tried to run a regression model on the overall attitude of consumer toward VoD
services and using rating given to various attributes as dependent variable.

As different clusters were expected to show different dependence we run a regression


model on each of the cluster and calculated the coefficient to observe the dependence of overall
attitude on various attribute data collected.

Cluster 1
Description: 21-35 Unmarried Students (Include PhDs, Masters and Undergraduate Student)

While running the regression we observed that there is significant correlation


[Exhibit 1] between Originality of Shows and Stream Quality & Device on which service is
offered and User allowed. To ensure that these correlations doesn’t affect our results we have
used residual methods while running the regression.
Coefficients SE T-Value P-Value
Intercept 8.71 3.44 2.53 0.01
Content 0.30 0.56 0.54 0.19
Plan 0.09 0.26 0.35 0.72
Variety -0.23 0.29 -0.81 0.42
Original Service -0.12 0.20 -0.58 0.56
Stream Quality 0.45 0.26 1.71 0.09
Interface 0.03 0.35 0.09 0.92
Device -0.19 0.22 -0.85 0.40
User 0.38 0.20 1.87 0.06
Network -0.06 0.21 -0.30 0.76

Upon running regression on the data corresponding to this cluster we


found out that Coefficient of Stream Quality and No. of Users play a major role in determining
the overall attitude of consumer toward service.

Coefficient of Stream Quality turns out be the highest 0.45 and is


significant with P-Value of 0.09. Coefficient of User is also very high 0.38 and is significant
with P-Value of 0.06.

Apart from these two coefficient for the quality of content is also high,
however this coefficient is significant at confidence level of 20%. Thus raising some doubt
regarding the value of the coefficient.

Inference: From above regression analysis we found out that for Cluster 1 (21-35 Unmarried
Students) Stream Quality, Content and No. of Users allowed are extremely important
parameters for a VoD service.

Cluster 2
Description: 21-35 Unmarried Working

While running the regression on this cluster as well we observed that there is significant
correlation [Exhibit 2] between Originality of Shows and Stream Quality, Device and
Originality of Service and No. of Users and Relatives using the same service. To ensure that
these correlations doesn’t affect our results we have used residual methods while running the
regression.

Coefficients SE T-Value P-Value


Intercept 9.79 4.73 2.07 0.06
Content 0.70 0.39 1.79 0.09
Plan 0.23 0.61 2.01 0.07
Variety 0.27 0.77 -0.17 0.87
Original Service 0.25 0.48 0.51 0.61
Stream Quality 0.21 0.70 0.30 0.77
Interface 0.86 0.74 1.15 0.27
Device -0.20 0.37 -0.53 0.60
User -0.28 0.34 -0.84 0.41
Network 0.05 0.40 0.13 0.90
Result showed that for this cluster Content and Plan played a major role in determining
the overall attitude of consumer toward VoD services.

Coefficients of content is extremely high 0.7 and is significant with 10%


confidence interval, whereas coefficient of Plan turn out to be 0.23 and is significant with P-
value of 0.07.

Inference: From above regression analysis we found out that for Cluster 2 (21-35 Unmarried
Working) Plan and Content are major deciding factor for choosing a VoD service.

Cluster 3
Description: 16-21 Unmarried Student (Majority Undergraduate Student)
While running the regression on this cluster as well we observed that there is significant
correlation [Exhibit 3] between No. of Users and Friends using the same service, Stream
Quality and Originality of Service. To ensure that these correlations doesn’t affect our results
we have used residual methods while running the regression.

Coefficients SE T-Value P-Value


Intercept 8.07 40.35 1.12 0.05
Content 0.73 3.25 1.15 0.07
Plan 0.43 1.04 0.99 0.09
Variety -1.95 1.77 -1.10 0.35
Original Service 2.49 1.83 1.36 0.27
Stream Quality -3.17 3.45 -0.92 0.43
Interface -1.94 2.16 -0.90 0.44
Device -0.48 1.72 -0.28 0.80
User 0.31 1.94 1.10 0.08
Network 1.37 2.23 0.61 0.58

Upon running regression model in the above data we observe that


coefficient for attribute such as Content, Plan and No. of Users are high with significant P-
Value.

Content is one of the biggest criterion for choosing a VoD service among this
segment as its coefficient is extremely high and is significant with 10% confidence interval.
Moreover, attributes like no. of users allowed also play a significant role as it was also revealed
in personal interviews that this segment generally shares a subscription among a group of
people to reduce the cost. Similarly, for plan as well we observed during personal interviews
that this segment prefers monthly plan as they generally prefer using these service during their
vacations.

Inference: From above regression analysis we found out that for Cluster 3 (16-21 Unmarried
Student) Content, No. of Users allowed and plan are extremely important parameters for a
VoD service.
Cluster 4
Description: 21-34 Married Working
While running the regression on this cluster as well we observed that there is significant
correlation [Exhibit 4] between No. of Users and Friends using the same service, Device and
Network. To ensure that these correlations doesn’t affect our results we have used residual
methods while running the regression.

Coefficients SE T-Value P-Value


Intercept 14.04 9.79 1.43 0.18
Content 0.44 1.36 2.05 0.08
Plan 0.89 0.58 1.54 0.15
Variety 0.45 0.53 2.14 0.06
Original Service 0.92 0.80 1.16 0.27
Stream Quality -1.59 0.91 -1.76 0.11
Interface 1.11 1.11 1.00 0.34
Device 0.34 0.44 1.78 0.09
User 0.46 0.30 1.51 0.16
Network 0.59 0.45 1.30 0.22

From the running regression on this cluster we observe that coefficient


for Content, Variety and Device are major criterion for deciding which VoD service to
subscribe for.

Variety is one of the biggest motivator as it was observed in qualitative research


that people of this segment want a VoD service that not only can satisfy their own need but
also can be used by children’s for watching cartoon show and other kids shows. Device is also
a big motivator as it was also revealed in personal interviews that they prefer watching movies
on smart TVs and other devices.

Inference: From above regression analysis we found out that for Cluster 4 (34 Married
Working) Variety, Devices available and Content are major factors in choosing VoD services.
Exhibits

Exhibit 1

Content Plan Variety Original Stream Interface Device User Network


Service Quality
Content 1.00 -0.02 0.11 0.09 0.10 -0.05 0.03 -0.06 -0.03
Plan -0.02 1.00 0.20 0.07 0.18 -0.16 -0.03 0.08 -0.04
Variety 0.11 0.20 1.00 0.25 0.14 0.02 -0.04 0.10 0.03
Original Service 0.09 0.07 0.25 1.00 0.46 0.19 0.26 0.39 0.39
Stream Quality 0.10 0.18 0.14 0.46 1.00 0.42 0.07 0.19 0.28
Interface -0.05 -0.16 0.02 0.19 0.42 1.00 0.29 0.12 -0.03
Device 0.03 -0.03 -0.04 0.26 0.07 0.29 1.00 0.49 0.48
User -0.06 0.08 0.10 0.39 0.19 0.12 0.49 1.00 0.48
Network -0.03 -0.04 0.03 0.39 0.28 -0.03 0.48 0.48 1.00

Exhibit 2

Content Plan Variety Original Stream Interface Device User Network


Service Quality
Content 1.00 0.20 0.28 0.22 0.32 0.17 -0.03 0.06 0.00
Plan 0.20 1.00 0.12 0.06 0.29 0.30 -0.12 0.02 0.17
Variety 0.28 0.12 1.00 0.10 0.23 0.00 -0.25 0.08 0.33
Original Service 0.22 0.06 0.10 1.00 0.46 0.37 0.44 0.34 0.17
Stream Quality 0.32 0.29 0.23 0.46 1.00 0.40 0.25 0.35 0.24
Interface 0.17 0.30 0.00 0.37 0.40 1.00 -0.03 -0.02 -0.14
Device -0.03 -0.12 -0.25 0.44 0.25 -0.03 1.00 0.48 0.30
User 0.06 0.02 0.08 0.34 0.35 -0.02 0.48 1.00 0.49
Network 0.00 0.17 0.33 0.17 0.24 -0.14 0.30 0.49 1.00

Exhibit 3

Content Plan Variety Original Stream Interface Device User Network


Service Quality
Content 1.00 -0.26 0.03 0.00 -0.41 -0.23 -0.42 0.12 -0.22
Plan -0.26 1.00 0.04 0.30 0.31 0.40 -0.25 -0.08 0.07
Variety 0.03 0.04 1.00 0.58 0.51 -0.11 0.13 -0.19 0.35
Original Service 0.00 0.30 0.58 1.00 0.74 0.53 0.00 0.30 0.73
Stream Quality -0.41 0.31 0.51 0.74 1.00 0.39 0.06 -0.04 0.66
Interface -0.23 0.40 -0.11 0.53 0.39 1.00 0.10 0.23 0.40
Device -0.42 -0.25 0.13 0.00 0.06 0.10 1.00 -0.12 0.30
User 0.12 -0.08 -0.19 0.30 -0.04 0.23 -0.12 1.00 0.60
Network -0.22 0.07 0.35 0.73 0.66 0.40 0.30 0.60 1.00
Exhibit 4
Content Plan Variety Original Stream Interface Device User Netwo
Service Quality rk
Plan 1.00 0.13 0.00 -0.31 -0.38 -0.13 -0.23 -0.34 -0.29
Variety 0.13 1.00 0.56 -0.29 0.04 -0.44 0.51 0.15 0.10
Original Service 0.00 0.56 1.00 -0.20 -0.08 -0.38 0.13 0.00 0.12
Stream Quality -0.31 -0.29 -0.20 1.00 0.87 0.09 0.26 0.64 0.60
Interface -0.38 0.04 -0.08 0.87 1.00 0.08 0.63 0.85 0.70
Device -0.13 -0.44 -0.38 0.09 0.08 1.00 -0.23 0.23 -0.29
User -0.23 0.51 0.13 0.26 0.63 -0.23 1.00 0.58 0.78
Network -0.34 0.15 0.00 0.64 0.85 0.23 0.58 1.00 0.57
-0.29 0.10 0.12 0.60 0.70 -0.29 0.78 0.57 1.00

Vous aimerez peut-être aussi