2 Project Report

8/2/2019 2 Project Report

1/20


2/20

Section:A(2009 - 11) ToyLibraryinHyderabad

Page 2 of 20

Artand

innovation

have

crossed

borders,

when

it

comes

to

educational,

mental

development

and

casual toys. Moreover, everyday innovation has become mandatory to remain in competitive

market has resulted in introduction of huge variety of toys in to the market. So, in order to

understand the consumer psyche and inclination in adopting the concept of Toy Library, thatoffersqualityandbrandedtoysonrent,chargingreasonableamountisthecoreidea,behindthis

projectstudy.Ofcourse,thisbeingabigbusinessidea,itrequiresextensiveresearchfornotonly

consumerperception,butalsotoysprice,brandpreferences,location,services,operationtiming

andbusinessstrategies.Butherewefocusonlyontotheverybasicofthisbusinessventurei.e.

DemandSurveyforToyLibrary;anditsdatacollectionandanalysis.

Factor Analysis and Discriminant Analysis, carried out over the collected data gave us some

insights about Toys Buying Behavior and Crche service adaptation propensity (which may be

includedasaseparateserviceunderToyLibrary).

Executive Summary


3/20


Page 3 of 20

Indias

urban

population

is

the

second

largest

in

the

world,

greater

than

the

combined

urban

populationsofallcountriesexceptChina,theUSandRussia.Overall,consumersspiritsarehigh,

whichhasledtosignificantgrowthinmanysectorsofconsumergoodsandcontinuedgrowthfor

thetoysandgamesmarket.Risingprices,thecostsofchildrenseducationandmedicalcare,and

thedeclineinsharepricesareneverthelessallkeyconcerns.ThetoysandgamesmarketinIndia

stoodatRs29.8billionin2007.Ontheotherhand,thevideogamemarketstoodatRs.4.8Billion

in2007.

Children

are

the

main

focus

of

Indian

families,

and

their

aspirations

in

terms

of

education

and

careerchoicesarerunningveryhigh.Theaveragefamilysize in India isdeclining:asofnow it is

almost4.3whereas inearlieryears itwasmorethan5.Withthereduction intheiraveragesize

andtheincreaseintheirincomes,Indianfamilieshavemoremoneytospend.Theirmainfocusis

their children, and they try their best to fulfill their aspirations. Children are getting more

attentionandparticipatingmoreinthedecisionmakingprocesses.

The US toymaker Mattels recall of thousands of its Chinesemade toys in 2007 had a positive

impact on the Indian market, which includes many manufacturers in the unregistered,

unregulated sector. Very little is known about the manufacturing processes, or even the

manufacturers,asthesedetailsarenotavailabletothepurchasers.Leadingtoycompaniesnow

believe that the increased awareness of quality concerns will prompt buyers to choose better

quality, branded toys. Most toy manufacturers in the unregistered, unregulated sector do not

follow these normsmuch of theoutputof the unbrandedsegment is not marked,according to

industry experts. In 2007, the toy sectors grew at 1820%, with the registered and regulated sector

growingat2225%,comparedwithalowerrateofgrowthintheunregistered,unregulatedsector.

The concept ofToyLibrary is very novice in Indian context,especially from the application andadoptabilitypointofview.Andtherestandsaroomforgrowth,duetogrowingliteratepopulation

andchangingbuyingpatterns,underanimpactofglobalization.AlreadyHyderabadiswellknown

for its Pharmaceutical and HealthCare sector growth. Also it is seen as an emerging IT hub in

southIndia,andnewlydevelopedHiTechcity(whereallITgiantsoperationsareconcentrated)is

Introduction


4/20


Page 4 of 20

growingnotonly intermsofnewentrants,butalso intermsofgeneratingrevenuefrom ITand

ITESservices.Growingskilledandsalariedclasspopulationisaddingconsumerbaseofthecity.

Theoverall

objective

behind

setting

up

aToyLibraryistooffervarietyofeducationaltoys,games,

books, interactive VCDs, CDROMs and computer games to kids of the age 2 to 10 years. Here

overalldevelopmentof kids,onenhancing theirvisual, verbal, intellectual, numericalanalytical,

sensoryandmemoryskillsbyadvancedlearningsystemsandeducationalaidshaskeptonmain

focus. Interactingwitha leading ToyLibrary franchiseprovider,wecame toknow thatat initial

stageitrequiresspaceofaround150Sq.foot;andinlatercourseonecanalsoextendthebusiness

byopeningupKGschool.InthestateofMaharashtra,thismodelhasgainwidespreadpopularity.

Keeping

in

mind

above

scenario,

we

worked

over

the

project

ofToy Library in the city of

Hyderabad.Ofcoursecommenceofbusinessrequiresextensiveresearchandmarketstudy.Butat

thisjuncture, we have restricted our research and analysis to the basic demand survey for Toy

LibraryinHyderabad.


5/20


Page 5 of 20

Aswe

said

earlier,

there

are

number

of

problems,

associated

with

the

business

model;

but

since

wehaverestrictedouranalysis tothebasicdemandsurveyandanalysisourmain focus ison

understanding the toybuying behavior of the consumers in the city. In order to analyze the

feasibility of the business model, we are required to know the consumer perception about the

toys, and their buying behavior. We worked on following aspects of consumer behavior in toy

buying:

Familystructure,earningsourcesandtotalannualhouseholdincome Preferencesinbuyingtoyseducational,mentaldevelopmentandcasual Attributesoftoyscustomerlooksforwhilebuyingtoys Buyingfrequencyandannualspendingontoys Preferenceoverbrandedandnonbrandedtoys Likenessoftheconceptoftheservicesandpropensitytoadopttheservices

In order to find the correlation between the attributes (variables) we covered in our analysis, we will

perform

factor

analysis

and

so

will

form

factors

(clubbing

relevant

variables).

These

factors

will

structure

thebaseforourfurtherbusinessresearchanalysis.

Alsowemadeanattempttodiscriminatetheattributesaffectingthedecisionofavailingservicesofcrche.

SincethebusinessmodelofToyLibrary,isascalablebusiness,ourtheseanalysiswillhelpusinclassifying

customersintogroupswhomaygoforcrcheservicesandwhomaynot.

Problem Analysis


6/20


Page 6 of 20

Sinceunderstanding

the

consumer

preferences

and

forces

in

buying

toys,

is

the

main

concern

of

the project; our overall questionnaire is designed based upon that only. It has already been

discussedindetail,aboutthequestionnairedesignandmodificationsbaseduponpilotsurveys.

WhileconductingtheFactoranalysis,followingresponseascribedrespectivequestionsbuildthe

foundation:

1. Ageofakid2. Annualhouseholdincome3. Importancetotoysfor

a. Educationpurposesb. Mentaldevelopmentpurposesc. Casualpurposes

4. Influenceoftoybuyingdecisionbya. Selfb. Spousec. Kidsdemandd. Neighbore. Colleague/friendofrespondent

5. LikelinessoftheconceptofToyLibrary6. PropensityofadoptionoftheservicesofToyLibrary

While conducting the Discriminant analysis, following questions responses will built the

foundation:

1. Ageofthekid2. Familymembers(adult)3. Annualhouseholdincome4. Whethertheyusecrcheornot5. Annualspendingfortoys

Questionnaire Design


7/20


Page 7 of 20

Thissurvey

requires

the

exercise

to

be

carried

out

in

weekends,

when

people

are

relaxed,

and

visit

shopping malls and gardens. Our focus was the people having kids in the age groupof 2 to 10

years. Further to this, we targeted those areas (shopping malls and gardens) only in order to

assure availability of respondents of required category, in sufficient quantity. Due to political

agitation over the division of the state of Andhra Pradesh, we could not reach to the people

personallyinweekends.Butbaseduponourlocalcontactsinthecity,wecouldapproachpeople

workinginITandHealthCaresectortosomeextentviaemails.Hence,ourentiredatacollection

wasmadeusingtheinternettoolsandGoogleSpreadsheet.

Thisexerciseofferedusfollowingbenefits,overandabovegettinggenuineresponses:

We could approach young and married tech savvy people, who are, in real sense theoriginalconsumersoftheservices.

Preparing spreadsheet on Google Documents is a cakewalk. Thus, it saved not only ourtimeandenergy;butalsoitwascosteffectivesincewehardlyrequiredgettinganyform

printed.

We mailed the link of the spreadsheet to the respondents; so it was purely on theirconveniencetofilltheformandrespondthesurvey.So,weexpectgenuineresponses.

TheresponsesofwereautomaticallygetregisteredintheGoogleResponseSheet,inMSExcel format. Thus, it was quite convenient to convert them into code, and record the

same.Thus,hereitsavedourtimeindataentryaswell.

Field-Work and Data Collection


8/20


Page 8 of 20

FactorAnalysis

ThepurposeofFactorAnalysis inourcase istodoanexploratorystudysincewedonothaveapredefinedideaofthestructureorhowmanydimensionsareinthesetofvariableswehaveviz.

Importanceofeducation,MentalDevelopmentandCasualType,Price,QualityandBrand.

Outputanalysis:

Data Analysis


9/20


Page 9 of 20

At this stage all Variables correlate fairly well and none of the correlation coefficients are

particularly large and we cannot eliminate any of the variables. Had any of the variable been

highlycorrelatedtoothervariableswewouldhaveconvenientlydroppeditasitwouldbedifficult

forustoputitunderonefactorandsuppressingitsdependencyonothersaswell.

Nowlets

have

alook

at

the

determinantvalue.Ifthisvalueisgreaterthan0.00001,wesaymulti

colinearityisnotaproblem.Inourcaseitis0.140whichishigherthan0.00001.Sowecandeduce

thatinourcaseatthisstagethereisnoneedtodropanyvariables.

KMOandBartlettsTest

KMOstatisticvariesbetween0and1.Ifitiscloserto1ourdatacanbeeasilyclassifieddistinctly

intofactorsbecausethepatternsofcorrelationarecompactandfor0wesaythefactoranalysis

would

not

give

appropriate

results

as

the

pattern

of

correlation

is

diffused

and

hence

is

difficult

to

segregatethevariablesintofactors.

Accordingtostandards:

KMOgreaterthan0.5isAcceptable.

0.50.7Good

0.70.9Excellent

0.91.0Desired.

OurCase:

KMOandBartlett'sTest

KaiserMeyerOlkinMeasureofSamplingAdequacy..681

Bartlett's

Test

of

SphericityApprox.

Chi

Square

294.094

df 55

Sig. .000


10/20


Page 10 of 20

Inourcaseitis0.681whichisfairenoughandwhichgivesusthumbsuptogoaheadwithfactor

analysis.

IntheabovetablewecanseeBartlett'sTestofSphericitywhichisactuallyfortestingthefollowingNULLHypothesis.Ho:TheCorrelationmatrixisanIdentitymatrixi.e.thereisnocorrelationbetweenthevariables,

soeachvariableitselfisaFactor.

Since for factoranalysistoworkweshouldhavesomerelationshipsamongthevariables,and if

matrix is Identity we would have all the correlation coefficients zeroes. Hence in order to do a

FactoranalysisweshouldbeabletorejecttheNullHypothesis.

Inourcaseasseenaboveinthetablepvalue(0.000)


11/20


Page 11 of 20

Compone

nt

InitialEigenvalues

ExtractionSumsofSquared

Loadings

RotationSumsofSquared

Loadings

Total

%of

Varianc

e

Cumulati

ve% Total

%of

Varianc

e

Cumulativ

e% Total

%of

Variance

Cumulati

e%

7

.682

6.200

82.393

8 .583 5.296 87.689

9 .523 4.754 92.443

10 .491 4.463 96.906

11 .340 3.094 100.000

ExtractionMethod:PrincipalComponentAnalysis.

NowsincewehaveaskedSPSStoextractfactorswitheigenvalue>1,itpulledthefirstfourhaving

2.697,1.698,1.349and1.046astheirEigenvaluesaregreaterthan1,whichgivesusfourfactors.

The eigenvalue associated with these four factors are again displayed in the after extraction

columnas

well

dropping

all

other

insignificant

factors.

But

in

our

case

we

see

the

cumulative

%

accountstoonly61.732% i.e.theextracted factorscanonlyexplain61.732%variance inthe11

variables.

Ifwegoby theoutputofSPSSwecan reduce thecomplexityof thedatasetby using the four

componentsloosingabout38%oftheinformation.

OneofthetableswhichareofsignificanceistableofCommunalities.

Forourcaseitis:

Communalities

Initial Extraction

Age 1.000 .650

Income 1.000 .523

EducationPurpose 1.000 .688

MentalDevelopment 1.000 .853

Casual 1.000 .482

ImportantPrice

1.000

.480

ImportantQuality 1.000 .348

ImportantBrand 1.000 .623

AnnualSpendingontoys 1.000 .529

Likingoftheservices 1.000 .805

propensitytousetheservices 1.000 .810


12/20


Page 12 of 20


Thetableshowscommunalitiesbeforeandafterextraction.

The

common

variance

of

each

variable

summarized

by

the

factors,

or

the

amount

(percent)

of

eachvariablethatisexplainedbythefactors. Communalityisactuallythepercentageofvariationinvariablethatisexplainedbythefactors(inourcasefour).Wecansaythefourfactorscandefine65%variance inAgevariable.Sinceafterextractionsome factorsaredropped,thisvalue65%

hascomedownowingtolossofinformation.

Nowtodecidewhatthefourcomponentsconsistsofwefollowthebelowsteps:The

rotated

component

matrix

helps

to

determine

what

the

components

represent

i.e.

which

of

thevariablescomeunderthecomponents.

RotatedComponentMatrix(a)

Component

1 2 3 4

ImportantBrand .776 .115

Income .701 .151

ImportantPrice

.669

.144

AnnualSpendingontoys .658 .302

ImportantQuality .508 .147 .208 .156

propensitytousetheservices.899

Likingoftheservices .894

EducationPurpose .135 .141 .778 .212

Age .121 .774 .177

Casual .430 .535

MentalDevelopment .922

ExtractionMethod:

Principal

Component

Analysis.

RotationMethod:VarimaxwithKaiserNormalization.

ARotationconvergedin4iterations.

InthistableweareconsideringonlytheAbsoluteValues,sothe firstcomponent ismosthighly

correlated with ImportantBrand, Income , ImportantPrice, Annual Spending on toys and


13/20


Page 13 of 20

ImportantQuality.SinceImportantBrandisnotwellcorrelatedwith2nd

,3rd

or4th

component,

it better represents the 1st component. The second component is most highly correlated with

propensity to use the services and hence it becomes the representative of 2nd

component.

ThirdcomponentwithhighlycorrelatedwithEducationPurpose,AgeandCasual,outofthesewe

can

select

Education

Purpose

as

the

representative.

Finally

in

the

same

way

Mental

Development

weselectitastherepresentativeofthefourthcomponent.

Ifwehavetoselectonerepresentative fromeachfactorweselectImportantBrand,propensitytouse the services, EducationPurpose andMentalDevelopmentwhichwe canuse in thefutureanalysis.

Component Score Variables: Using the component coefficients given in theComponentScore

CoefficientMatrix and thevaluesofeachvariable foreachobservationwederiveanequation.

Usingthis

equation

we

calculate

Component

Score

for

each

component.

ComponentScoreCoefficientMatrix

Component

1 2 3 4

Age .123 .046 .531 .163

Income .312 .019 .159 .090

EducationPurpose .063 .095 .508 .176

MentalDevelopment .072 .044 .027 .865

Casual.111

.026

.301

.042

ImportantPrice .282 .056 .008 .115

ImportantQuality .209 .077 .069 .182

ImportantBrand .335 .000 .028 .145

AnnualSpendingontoys .272 .023 .106 .231

Likingoftheservices .003 .535 .014 .055

propensitytousetheservices .000 .537 .041 .014


RotationMethod:VarimaxwithKaiserNormalization.

Component

Scores.

Forexample:forComponentScore1= (.123)*Age+(.312)*Income+(.063)*EducationPurpose

+(.072)*MentalDevelopment+ (.111)*Casual+(.282)*ImportantPrice+(.209)*Important

Quality + (.335)* ImportantBrand+ (.272)* Annual Spending on toys + (.003)* Liking of the

services+(.000)*propensitytousetheservices.


14/20


Page 14 of 20

TheResultingfourComponentScoreVariablesarerepresentative of,andcanbeusedinplaceofelevenvariables,withalossof38%oftheinformation.Hencewehaveactuallyreducedthedatafilefrom11variablesto4componentscorevariables.


15/20


Page 15 of 20

DiscriminantAnalysis

AnalysisCaseProcessingSummary

UnweightedCases N Percent

Valid 155 .2

Excluded Missingoroutofrangegroupcodes 0 .0

Atleastonemissingdiscriminating

variable0 .0

Bothmissingoroutofrangegroupcodes

andatleastonemissingdiscriminating

variable65380 99.8

Total

65380 99.8

Total 65535 100.0

"Processed"casesarethosethatweresuccessfullyclassifiedbasedontheanalysis.Thereasons

whyanobservationmaynothavebeenprocessedarelistedhere. Here0.2%ofthecasescould

notbesuccessfullyclassified.

EigenvaluesandMultivariateTests

Eigenvalues

Function Eigenvalue %ofVariance Cumulative% CanonicalCorrelation

1 .131(a) 100.0 100.0 .340

a First1canonicaldiscriminantfunctionswereusedintheanalysis.

Herewehavegotonly2discriminatingvariable.Henceweuseonly1function. Eachfunctionacts

as projections of the data onto a dimension that best separates or discriminates between the

groups.

Eigenvalue The magnitudes of the eigenvalues are indicative of the functions' discriminating

abilities


16/20


Page 16 of 20

% of variance: This is the proportion of discriminating ability of the three continuous variables

found in a given function. This proportion is calculated as the proportion of the function's

eigenvaluetothesumofalltheeigenvalues.

Inourcasewehavegotonly1discriminatingfunction.Hence100%

.

Wilks'Lambda

TestofFunction(s) Wilks'Lambda Chisquare df Sig.

1 .884 18.681 2 .000

WilksLambda

shows

the

proportion

of

the

total

variance

(90%)

in

the

Discriminant

scores

not

explainedbydifferencesamonggroups.AsmallLambdavalue(near0)indicatesthatthegroups

meanDiscriminantscoresdiffer.Wilks'lambdaisadirectmeasureoftheproportionofvariancein

thecombinationofdependentvariablesthatisunaccountedforbytheindependentvariable(the

grouping variable or factor). If a large proportion of the variance is accounted for by the

independentvariablethen itsuggeststhatthereisaneffectfromthegroupingvariableandthat

thegroups havedifferentmeanvalues.Wilks'lambdaperforms,inthemultivariatesetting,witha

combinationofdependentvariables,thesameroleastheFtestperformsinonewayanalysisof

variance.

Chisquare This is the Chisquare statistic testing that the canonical correlation of the given

functionisequaltozero. Inotherwords,thenullhypothesisisthatthefunction,andallfunctions

thatfollow,havenodiscriminatingability. ThishypothesisistestedusingthisChisquarestatistic.

DiscriminantFunctionOutput

StandardizedCanonicalDiscriminantFunctionCoefficients

Function

1

Whoalldostaywithyoucurrently,exceptkid(s)? .836

Totalannualhouseholdincome.764

Inourcase,wefindthatannualhouseholdincomeisgreaterinmagnitudethan,whoallstaywith

kids.Henceitwillhavegreaterimpactonthediscriminantscore.


17/20


Page 17 of 20

ClassificationProcessingSummary

Processed 65535

Excluded Missingoroutofrangegroup

codes

0 Atleastonemissing

discriminatingvariable65380

UsedinOutput 155

ClassificationProcessingSummary ThisissimilartotheAnalysisCaseProcessingSummary,but

inthistable,"Processed"casesarethosethatweresuccessfullyclassifiedbasedontheanalysis.

ClassificationResults

Do

you

use

Crche

foryourkid(s)? Predicted

Group

Membership Total

1 2

Original Count 1 18 8 26

2 40 89 129

% 1 69.23076923 30.76923 100

2 31.00775194 68.99225 100

Thesearethepredictedfrequenciesofgroupsfromtheanalysis. Thenumbersgoingdowneach

column indicate how many were correctly and incorrectly classified. In our case for yes group

66%werecorrectlyclassified.And28wereincorrectlyclassified.

Original These are the frequencies of groups found in the data. Here we see form frequency

tablethatforYeswehadcountof26.Heretoowefindthat18werepredictedcorrectlyand8

werenot.

% Thisportionofthetablepresentsthepercentofobservationsoriginallyinagivengroup(listed

intherows)predictedtobeinagivengroup(listedinthecolumns). Inourcase31%ofYesgroup

is predicted to be in the No group. However 30.76% of No group is predicted to be in the Yes

group

Analysis

From thestructurematrix itcanbe interpretedthat Noofmembersstayingwiththekid isan

important variable in determining whether to have a crche or not. It has got a value of 0.667

whichishigherthananyoftheothervariables.


18/20


Page 18 of 20

AlsotheWilks lambdahere is .884whichtransforms tochisquareof18.681with2degreesof

freedomandlevelofsignificance


19/20


Page 19 of 20

1. Fourfactorsemergedoutofavailable11variables;viza. Socialstatus

i. Using these factors we can work further on Consumer behavior of aparticular class people, Location, Toys choices (Brand) and Marketing

strategies

b. Adoptionoftheservicesi. Sinceourresultssay,thosewholiketheconceptexpressedtheirwillingness

toavailtheservices.Hence,moreemphasiscanbeputonmakingtheidea

moreattractive,

offering

abroad

range

of

services

and

assuring

sufficiently

bigtoyscollection.

c. Intellectualleveli. ItshowsAgeandToyspurchaseforeducation&casualpurposearehighly

correlated with each other. So, for us its essential to form a structure of

variousagegroupsofkids,andbuytoyskeepinginmindtheirAge.

d. Mentaldevelopmenti. It shows irrespective of any age, income, price, quality or brand while

buyingtoys,Mentaldevelopmentisgivenadueimportance.

2. DiscriminantAnalysissaysfollowings:a. WemaygoforofferingCrcheservices,targeting:

i. Nuclearfamiliesii. Familieswithhighdisposableincome

b. BeforewegoformakingavailCrcheservices,weneedtoassurethedemandofthesame.ThedemandsurveytoCrcheservicemayhelpinscalingthebusinessin

furthercourse.

Conclusion


20/20


Page 20 of 20

1. http://www.toysindia.in/indiantoy_industry.html2. www.euromonitor.com/Toys_And_Games_in_India3. http://kidsgurukul.com/4. http://ww.smashits.com/video/snoop/1605/hyderabadstoylibraryisahitwithits

children.html

References