conjoint analysis for smart phones
-
date post
13-Sep-2014 -
Category
Business
-
view
1.746 -
download
4
description
Transcript of conjoint analysis for smart phones
1 | P a g e
The Smartphones in
India – A Conjoint
Analysis and Simulation
CFPP Project Report Date of submission: 17-Feb-2012
Group 16 Dhruv Dhruv Dhruv Dhruv Anand Anand Anand Anand ((((FT12425FT12425FT12425FT12425)))) SudhanSudhanSudhanSudhanvavavava DDDD VVVV ((((FT12264FT12264FT12264FT12264)))) Anamika Anamika Anamika Anamika Roy Roy Roy Roy ((((FT12477FT12477FT12477FT12477)))) BikramBikramBikramBikram Satapathy Satapathy Satapathy Satapathy ((((FT12417FT12417FT12417FT12417)))) SrinivasSrinivasSrinivasSrinivas Dhenuvukonda Dhenuvukonda Dhenuvukonda Dhenuvukonda ((((FT12467FT12467FT12467FT12467)
2 | P a g e
Contents
Introduction ....................................................................................................................................................... 3
Objectives ........................................................................................................................................................... 4
Methodology ...................................................................................................................................................... 5
Pre-study and selection of attributes ............................................................................................................. 5
RESULTS AND ANALYSIS ..................................................................................................................................... 7
Participation Level .......................................................................................................................................... 7
Conjoint Analysis ................................................................................................................................................ 8
Introduction ................................................................................................................................................... 8
Part worth utility curves ............................................................................................................................... 10
Demographic Analysis ...................................................................................................................................... 11
Benefit segment Analysis ............................................................................................................................. 13
Market simulation ............................................................................................................................................ 16
Sensitivity Analysis ....................................................................................................................................... 17
Issues and recommendations to full-scale study ......................................................................................... 19
Conclusions .................................................................................................................................................. 19
Exhibit-1 ........................................................................................................................................................... 20
3 | P a g e
Introduction
A smart phone is can be found in every second hand nowadays – from a white collared
professional to a student! This pervasive entry of smart phones into our lives is due to two
primary reasons. One reason being the rapid advancement in technology and R&D which
makes the present technology redundant within weeks. The second is the drastic drop in
prices which occur every few days.
So the question now is why is the Smart Phone such an in demand product when compared
with our traditional Feature phone; for a multitude of reasons. A smart phone is a mobile
phone built on a mobile computing platform, with more advanced computing ability and
connectivity than a feature phone. The presence of Application Programming Interface (APIs)
on the smart phones is used for running the third party applications which bring in life to the
mobile phone.
The first smartphones were devices that mainly combined the functions of a personal digital
assistant (PDA) and a mobile phone or camera phone. Smart phones now have well-developed
touchscreens, web browsers that can access any page on the web and not just sites designed
specifically for mobile phones, and high-speed data access via Wi-Fi and mobile broadband.
We were immensely interested in understanding the consumer’s preferences in this ever
changing dynamic market. The cell phone from being a product of utility at the beginning
turned into an accessory and a hand held device with multiple features. We are in a very
crucial phase for a country like India where the purchasing power of the consumer is
increasing and they don’t think too much about spending a little more.
We designed our project with the intention to understand how the different attributes and
features provided by the manufacturers hold how much value to the consumer. When there is
a tradeoff between different attributes the consumer makes a choice based on the attributes
he considers the most favorable to his taste. This process is as much a scientific process as is
psychological.
The study is conducted among urban individuals who are part of the workforce and among
students who are old enough to own mobile phones. The Smart phone manufacturers can
enhance their products and better position their phones to the consumers. After all, the
customer is the king!
4 | P a g e
Objectives
The objective of this study is to understand the consumers preference in the purchase of
Smart Phones and the attributes he/she thinks are of importance during the time of purchase.
We are trying to understand how the five attributes – price, design, brand, shape and user
friendliness, interact with each other to shape the purchase decision of the consumer.
We hope this conjoint project will help prioritize the most desired attributes of the Smart
Phone so as to maximize revenue by understanding the consumer utility. By using this
conjoint analysis, we can conclude what are the most significant attributes and what is each
attributes’ relative value. This study gives us insights into what are the consumer’s
preferences in a Smart phone and how changes in each attribute effect the likelihood of
purchase.
5 | P a g e
Methodology
Pre-study and selection of attributes
In order to select the attributes of the service, we conducted qualitative research. Two Focus
group discussion and 6 interviews were conducted to understand the participants’
preferences. This formed as a good base to arrive at attributes and further frame the levels of
the attributes. As the product is smart phones so the Participants were enthusiastic in
expressing about the latest trends and style of the smart phone.
Attributes and Levels
Price
• Rs 40000
• Rs 30000
• Rs 20000
User-friendliness
• Low
• Medium
• High
Brand
• Nokia
• Apple
• HTC
• Samsung
Design
• Trendy
• Sleek
• Changeable Skin
Shape
• Touch screen
• Qwerty
• Slider
6 | P a g e
Survey Development and Design
In order to make the survey effective, a two part conjoint survey was designed. The survey asked
the survey takers to first rate the importance of various attributes. The survey takers were then
given various options to choose between mock purchase scenarios, in this case job offerings. This
was developed using ASEMAP which is a computer adaptive survey generator. This survey was
linked to the demographic survey which was developed using Survey gizmo. The purpose of linking
both was to have consistency in data and also not to break the flow of the survey for the survey
takers. The survey was then tested by the team and a few participants to gauge the user-
friendliness and usability. Based on the feedback, changes were made in the survey design.
Survey Administration.
Our target response size was 75 in order (15 per member) to complete our survey. The
challenge was to get the respondents fill all the four sections of the survey. We tracked the
validity of the responses based on the Rank order correlation, Adjusted R-Square, Logit
Adjusted R conversation and made sure that 50 valid responses are available for analysis.
7 | P a g e
RESULTS AND ANALYSIS
Participation Level
Validity of Data We got 65 responses for the survey. On analysis of the data, it was found that few of the
responses were not valid. Following criteria was used to determine the validity of the data
• Rank order should be greater than 0.5
• Adjusted R square should be greater than 0.25
• Logit Adjusted R square should be greater than 0.25
Only those responses were considered which satisfied all the three conditions mentioned
above. It was found that 23 responses did not satisfy at least one of the conditions mentioned
above and hence they were removed. The demographic data for these respondents was also
removed. Analysis was done for the remaining 54 valid responses.
8 | P a g e
Conjoint Analysis
Introduction
Based on the ASEMAP output, we calculated the mean utilities and importance levels for the
various attributes and individual levels. All 5 attributes were then ranked according to their
importance levels. Table 1 shows the importance rankings of different attributes as well as
the mean part worth utilities of a given attribute level.
A review of Table 1 shows that price is the most important attribute, across all participants
with its importance level being 37.3%. Brand, User-friendliness, shape & design formed the
next set of important attributes with the important percentages being 25%, 21.7%, 8.7% and
7.3% respectively.
It is also prevailed and common observation that brand and price are the two most important
attributes for buyers in the consumer market before purchasing any smart phone.
Table 1
Sl No Rank Attributes Levels Mean utilities
Mean importance
1 2 Brand
Apple 10.83
25.0% Nokia -1.00
Samsung 1.14
HTC -10.97
2 1 Price
20000 17.12
37.3% 30000 -1.67
40000 -15.45
3 5 Design
Sleek 2.93
7.3% Trend 0.49
Changeable skins -3.42
4 4 Shape
Slider -3.52
8.7% Qwerty -0.58
Touch screen 4.10
5 3 User
friendliness
Low -9.56
21.7% Medium 0.15
High 9.40
9 | P a g e
From the conjoint analysis one can says that consumers are more interested in price of the
smartphone and followed by brand. This could be the reason that Samsung captured market
from Nokia in smartphone segment due to Samsung decreased prices for the same phone
segment.
The next best attribute is brand with nearly 25% importance for consumers in the market. It
will also tell that consumers will prefer brands to other attributes while purchasing a mobile.
User-friendliness is the third most important for consumers of softphone segment. As the
most of the consumers go for smartphones only for the sake of sophistication, versatility and
fully functional. Only user-friendly mobiles will have all those qualities.
The utility curves of conjoint analysis are given in the figures
10 | P a g e
Part worth utility curves
The mean part worth utility curves are drwan for the 5 attributes and as given above
11 | P a g e
Demographic Analysis
We collected demographic data in 12 categories:
• Age
• Gender
• Family income
• Profession
• Usage Level per day (hours)
• Location
• Brand
• Persons influencing brand selection
• Price
• Style
• Features
• User friendliness
Gender based preferences for Brands:
Of all female respondents 78% preferred Apple brand of smart phones. For males the
preferences vary with Apple being preferred by 46% of respondents, 33% preferring
Samsung.
Preferred Brand
Apple Samsung
Female 78% 22%
Male 46% 33%
12 | P a g e
Gender based Usage level patterns:
The level of usage per day for females suggests that 67% use it for more than 4 hours a day.
The percentage of males that use it for less than 4 hours is 67 %. Of these 50% preferred
Apple brands of smart phones.
Usage Pattern per day
3-4
hours
4-6
hours
Female 30% 67%
Male 67% 21%
Inference from Importance Ratings:
The importance rating show the contrast in the preference when it comes to Price as a factor,
almost equal percentage of respondents have preferred Price as most and Least important
criteria.
This shows that Price is a critical factor and has extreme reactions from respondents.
Similarly, more than 50% of respondents have chosen Style as 2nd or 3rd in ratings indicating
that relative unimportance compared to Price.
Features comes across as a criteria in which there is almost equal distribution whereby the
respondents are divided and hence there may be a clarity required in terms of explaining or
understanding the meaning attached to what all is covered in features when respondents rate
it.
User-friendliness as criteria seems to the 1st preference for minimum number of respondents
and hence gives insight that even for a complex product like smart phones the user attaches
less importance to it.
13 | P a g e
Importance (1 - high, 4 - Low Priority
1 2 3 4
Price 30% 12% 6% 33%
Style 12% 24% 30% 18%
Features 21% 15% 24% 24%
User-friendliness 10% 24% 27% 24%
Benefit segment Analysis
Table 2: Product category table
Demo
variable
1
2
3
4
5
6
Sex 0-female
1-male
Location NCR/DELHI Mumbai Chennai Bangalore Hyderabad Others
Income
monthly
<50000 50000-
100000
>100000
Profession Private
employed
Govt
employed
Others
We have Performed Cluster Analysis (Benefit Segmentation) and Pseudo-F calculations. The
value of Pseudo-F is max for 3 segment levels. The value of maximum Pseudo-F is 8.3704.
Cluster No Pseudo-F
2 clusters 7.662908
3 clusters 8.373043
4 clusters 8.119052
14 | P a g e
Table 3: part worth utilities segment wise
Final Cluster Centers
Cluster Cluster(importance)
1 2 3 1 2 3
Rs 40000 -10.44 -9.28 -19.05
22.70 21.26 48.00 Rs 30000 -.46 -1.88 -2.03
Rs 20000 10.91 11.16 21.08
Slider -1.37 -4.32 -4.04
8.33 7.63 9.92 Qwerty -3.23 1.31 -.21
Touch screen
4.60 3.01 4.25
Low -7.22 -25.37 -5.63
17.33 49.41 13.56 Medium -1.84 3.25 -.08
High 9.07 22.12 5.71
Sleek .08 1.46 4.37
4.13 4.10 10.26 Trendy 1.90 1.02 -.17
Changeable skin
-1.98 -2.48 -4.20
Apple 19.08 6.65 9.19
47.51 17.59 18.26 Nokia -6.18 4.66 -.89
HTC -25.59 -10.25 -6.07
Samsung 12.68 -1.06 -2.24
Segment 3 is a of price sensitive than segment than segments 1&2 as its part worth utility importance is
more for price. Segment 1 is more of brand oriented. Mostly all segments are not much worried about
either design or shape of the smartphone. It is widely assumed statement that any normally available
smart phone must be a good shaped phone and will have touch screen compulsorily. Hence
consumers may not feel those two attributes are not as important as others such as brand, price and
user-friendliness.
15 | P a g e
Table 4 : Anova table for segmentation
ANOVA
Cluster Error
F Sig. Mean Square df Mean Square df
Rs 40000 331.756 2 32.959 30 10.066 .000
Rs 30000 6.522 2 28.210 30 .231 .795
Rs 20000 398.522 2 53.657 30 7.427 .002
Slider 20.861 2 22.022 30 .947 .399
Qwerty 36.612 2 44.667 30 .820 .450
Touch screen 4.653 2 31.434 30 .148 .863
Low 923.569 2 33.384 30 27.665 .000
Medium 43.339 2 7.427 30 5.835 .007
High 621.743 2 36.544 30 17.014 .000
Sleek 55.791 2 8.424 30 6.623 .004
Trendy 12.172 2 12.170 30 1.000 .380
Changeable skin 16.034 2 12.186 30 1.316 .283
Apple 317.433 2 32.432 30 9.788 .001
Nokia 190.143 2 74.542 30 2.551 .095
HTC 989.509 2 57.211 30 17.296 .000
Samsung 594.829 2 57.754 30 10.299 .000
The F tests should be used only for descriptive purposes because the clusters have been chosen to
maximize the differences among cases in different clusters. The observed significance levels are not
corrected for this and thus cannot be interpreted as tests of the hypothesis that the cluster means are
equal.
From Anova table of cluster analysis, it is also found that, changing price from Rs 40000 to Rs 30000 is
a significant rather changing to Rs 30000. Also, Apple has been chosen by consumers as a significant
player as a brand in the smartphone market.
16 | P a g e
Market simulation
Our primary objective of this market simulation is to find the product attributes that
maximizes the share of the product. Profitability cannot be an objective in this simulation
because we are not known of costs of features of the product.
The best segments from the benefit segment analysis is given below Table 5: Best features of the 3 segments
Segment Price Design User-friendliness shape Brand
Segment1 20000 T/s High Trend Apple
Segment2 20000 T/s High sleek Apple
Segment 3 20000 T/s High sleek Apple The following table gives the various attributes of various brands in the market. Table 6
Segment Price Design User-friendly shape Brand
Product 1 40000 Touch screen High Trend Apple
Product 2 35000 Touch screen High sleek Nokia
Product 3 25000 Touch screen Medium sleek Samsung
The choice shares of the products are calculated by conjoint simulator using the principles of
maximum choice or log it choice rules. If the three brands Apple, Nokia and Samsung launch
those products in the market, then the choice shares of the above products by the consumers
can be calculated as follow.
Table 7
Segment Brand
Choice share (Max choice rule)
Choice share (Logit choice rule)
Product 1 Apple 26.94% 24.34%
product 2 Nokia 14.18% 19.86%
Product 3 Samsung 30.56% 27.47%
No purchase None 28.32% 28.32%
17 | P a g e
Samsung would emerge as a most preferable consumer choice, if the same configurations are
competing in the market. It is purely due to price advantage over the others. There is slight
change in choice shares from the both choice rules. It is most preferable to use log it choice
rule for FMCG kind of products.
Sensitivity Analysis
What if suddenly Google launches a phone with the following configuration? Let us see the
calculations of the choice shares of the consumers. It is assumed that, since Google is new to
smartphone market, it is brand is perceived as in between Apple and Nokia. Now let us see
how the dynamics of the smartphone market changes with the launch of Google smartphone.
Table 8
Segment Price Design Uf shape Brand
Product 1 40000 Touch screen High Trend Apple
product 2 35000 Touch screen High sleek Nokia
Product 3 25000 Touch screen Medium sleek Samsung
Product 3 36000 Touch screen High sleek (middle of Apple and Nokia)
The following table will give us the consumer choice shares of the various smartphones.
Table 9: Choice shares
Segment Brand
Choice share (Logit choice rule)
Product 1 Apple 18.54%
product 2 Nokia 14.88%
Product 3 Samsung 21.44%
Product4 Google 17.09%
No purchase None 28.05%
So, with the introduction of Google with the above specified configuration, there is huge
damage to Both Apple and Samsung by 8% and 13% approximately assuming no reaction of
competitors with the new launch of the smartphone by Google.
18 | P a g e
Suppose, in reaction to the Google’s launch, if Samsung changed its one of the attributes price
to Rs 22000/-
Hence, the following table will give us the choice shares post Samsung’s reaction to Google.
Table 9: choice shares
Segment Brand
Choice share (Logit choice rule)
Product 1 Apple 17.89%
product 2 Nokia 14.27%
Product 3 Samsung 25.77%
Product4 Google 16.41%
No purchase None 25.66%
There is slight gain of choice share for Samsung, but Google loses slightly. Whereas Apple and
Nokia don’t get any affect out of Samsung’s change of price. The entire usage conjoint
simulator is captured in screenshot as an Exhibit1.
19 | P a g e
Issues and recommendations to full-scale study
There are several limitations in the study including limited sample set. The sample size is
small and the lack of details regarding the cost the sample is limited only to regular students
which should have been extended to staff, weekend batch students the more diverse sample
would have given more robust and better results. One more concern in the survey is many
respondents did not complete the ASEMAP survey. Also we need to go for heterogeneous
samples to get perfect utility values.
Conclusions
1. The five attributes we chose as key features for a Smart Phone gave us insights into the
purchasing behavior of the consumers. The tradeoffs they had to make while making a
choice of a Smart Phone among the attributes force them to choose few and leave out
others. For certain individuals certain attributes were more important than others. The
sequence of questioning ensured the consumer’s picked the attributes which really
mattered to them.
2. Conjoint analysis helps us to understand the part worth utilities, there by importance
levels and choice shares of any products.
3. Conjoint simulator can be used to strategically position the product or introduce the
product into the existing market.
4. Apple remained undisputed leader in the smartphone market according to consumer
preferences. Samsung is emerging as an alternative to Nokia.
20 | P a g e
Exhibit-1
Conjoint simulator for calculating choice shares