factors influencing impulse buying during an online purchase

《factors influencing impulse buying during an online purchase》由会员上传分享，免费在线阅读，更多相关内容在工程资料-天天文库。

ElectronCommerceRes(2007)7:367379DOI10.1007/s10660-007-9011-8FactorsinfluencingimpulsebuyingduringanonlinepurchaseScottA.Jeffrey·RebeccaHodgePublishedonline:24October2007©SpringerScience+BusinessMedia,LLC2007AbstractUsingactualpurchasingbehaviorbyvisitorstoaHighSchoolReunionwebstore,thisstudyexaminesthefactorsthatleadtoanincreasedwillingnessbyon-lineconsumerstopurchaseimpulseitems.Consistentwithmentalaccountingandthepsychophysicsofprices,wefindthatpeoplespurchaseoftheimpulseitemincreaseswiththetotalamountspentonotheritems.WealsofindthatlinkingaUS$1donationtotheimpulseitem,therebyprovidingareasontopurchase,increasesthefrequencyoftheimpulsepurchase.KeywordsImpulsebuying·E-commerce·Reasonbasedchoice·MentalaccountingBetween1945and1959impulsepurchases,thosepurchaseswhichashoppermakesbutdidnotplaninadvance,rosefrom38.2percentto50.9percentofthetotalpur-chasesinsupermarkets[1].Inthelate1970sresearchersfoundthatimpulsepurchasesaccountedforbetween27and62percentofpurchasesindepartmentstores[2].Morerecently,AnnieSeeley,aFoodCommissionnutritioniststatedthatseventypercentofconfectionaryisboughtonimpulse[3,p.23].AmoreextremeresultwasfoundbythePOPAI(pointofpurchaseindustrybody)whichindicatedthat75percentofbuyingdecisionsaremadein-store[4].Sincealargeportionofspendingcanbecat-egorizedasunplannedorimpulse,understandingthepsychologicaldriversofthisbehaviorisofcriticalimportance.S.A.Jeffrey()DepartmentofManagementSciences,UniversityofWaterloo,200UniversityAve.West,Waterloo,ONN2L3G1,Canadae-mail:sajeffre@engmail.uwaterloo.caR.HodgeE-XYNWebSolutions,P.O.Box1042,Guelph,ONN1H6N1,Canadae-mail:becky@e-xyn.com 368S.A.Jeffrey,R.HodgePreviousresearchonimpulsepurchasinghasledtoimprovementsinpoint-of-saledisplaysatthecheckout,betterlocationofproductsonshelves,andanincreaseinunderstandingofproductpackagingandbundlingtoappealtotheimpulsedrive.Whiletherehasbeensignificantresearchonimpulsepurchasingintraditionalbrickandmortarstores,verylittleresearchexistsonimpulsepurchasesine-commerce.Thisisanimportantgaptoaddress,becauseconsumerexpenditureontheworldwideweb(WWW)hasgrownfromUS$15Bin1999toUS$69Bin2004;aCAGRofover40%[5,6].Whilethegrowthisexpectedtoslow,theshorttermforecastsarefor20%to25%increasesoverthenextfewyears.Thisisfarabovetheanticipatedgrowthinoverallretailfigures[6].Ifonlineretailerscouldgainabetterunderstandingofthefactorsthatencourageimpulsepurchasesonline,evena1percentincreaseinsalesfromimpulsepurchaseswouldleadtoanadditionalUS$690millioninrevenues.Withsuchalucrativeopportunityavailabletoonlineretailersitisdifficulttounderstandwhylittlerigorousresearchinthisfieldexists.Wehopethisstudyprovidesafirststeptowardsaddressingthisgapintheliterature,andencouragesotherscholarstostudythisfield.1ImpulsepurchasesStern[1]wrotethatimpulsebuyingissynonymouswithunplannedbuyingandde-fineditasanypurchasewhichashoppermakesbuthasnotplannedinadvance.Thisdefinitionisfairlyconsistentamongotherimpulsepurchasingliterature[7,8].Morerecently,researchershaveextendedthisdefinitionbeyondasimpleunplannedpurchasetoincludeanemotionalelementoranurgetomakethepurchase.Rook[9,p.191]definedimpulsebuyingaswhenaconsumerexperiencesasudden,oftenpowerfulandpersistenturgetobuysomethingimmediately.Piron[10]conductedathoroughsurveyofimpulsepurchasingliteratureandproposedamorespecificandcomprehensivedefinitionforimpulsebuyingthatincludesfourcomponents:thepur-chaseisunplanned,itistheresultofanexposuretostimulus,itisdecidedon-the-spot,anditinvolvesanemotionaland/orcognitivereaction.Thisistheworkingdefinitionthatwillbeusedforthisstudy.Inthispaper,wewillusethetermsimpulsebuyingandimpulsepurchaseinterchangeably.Previousresearchhasclassifiedimpulsepurchasingintofourtypes:pure,sugges-tion,reminder,andplanned[1,4].Pureimpulsebuyingsignifiesasituationwherethepurchaseisoutsideofthenormalpurchasebehaviorandisinitiatedbyemotionalappeal.Forexample,aconsumerwhorarelybuysmagazinesmightseeaPeoplemagazineatthecheckoutwhilewaitinginlineatthegrocerystoreandwantitbasedonthecoverstoryorpictures.Inthiscase,thepurchaseofthemagazinewouldbeconsideredpureimpulsebecauseitisoutsideofthenormalpurchasebehavioranditsatisfiesanimmediatedesireinitiatedbyanemotionalappeal.Suggestionimpulsebuyingoccurswhentheconsumerseestheproduct,visualizesanapplicationforit,anddecidesthattheyneedit.Thisisnotapureimpulsepurchasebecausetheuserhasdeterminedarationalorfunctionalpurposefortheitem,whereasduringthepureimpulsetransactiontheuserisfulfillinganemotionaldesire.Itisthesetwotypesofimpulsebuyingwewillconsiderinthisstudy. Factorsinfluencingimpulsebuyingduringanonlinepurchase369Wewillnotconsiderreminderimpulsebuyingwhichoccurswhenapurchaseismadebaseduponsomethingremindingtheconsumertomakethepurchase.Incontrasttoapureimpulsepurchase,theitemissomethingthattheusernormallypurchasesbutisnotnecessarilyontheircurrentshoppinglist.Forexample,ashoppermightwalkdownthecerealisleandrealizethattheyarealmostoutofCheerios.Ifthisconsumerchosetopurchasecheerios,itwouldtechnicallybeanimpulsepurchaseasitwasunplanned,butitsomethingthatisgenerallypurchasedbytheshopper.Wewillalsonotconsiderplannedimpulsebuyingwhichoccurswhentheconsumerbuysaproductbasedonpriceand/orproductspecials.Forexample,someonewhoconsumesagreatdealoforangejuicemayseeaspecialpriceonitandbuythem,evenifitwasnotanintendedpurchaseforthattriptothemarket.Asmentionedearlier,mostexistingresearchonimpulsebuyinghasfocusedonbrick-and-mortarstores[7,8,1115]andtelevisioninfomercials[16].Previousre-searchone-commercehastraditionallyfocusedonimprovingtheuserexperience[17],advertising[18,19],andidentifyingfactorsthatleadtoaninitialpurchaseonawebsite[20].Onlylimitedworkhasbeendoneonimpulsepurchasesinanonlineenvironmentandanyresultswithrespecttoimpulsepurchaseswereinconclusive[21,22].Inaddition,mostofthepreviouslystudiedantecedentstoimpulsepurchaseshavebeenemotional.Inthispaperwetakeamorecognitiveapproachbyanalyzinghowtheconceptofpurchasejustificationplaysaroleinimpulsedecision.Beforediscussingourcontributiontothisliterature,weanalyzetheotheran-tecedentstoimpulsepurchases.Previousresearchhasmostlydiscussedemotionalantecedentstoimpulsepurchases,inparticular,theeffectofmoodandpositivevs.negativeaffect.RookandGardener[23]lookedatmoodsandfoundthattheydroveovertbehavior.Pleasure,arousal,anddominanceareallassociatedwiththeencour-agingorsuppressionofimpulsepurchases.Positiveaffectcanencourageincreasebrowsingbehaviorwhichwasalsofoundtoleadtohigherlevelsofimpulsepurchases[23,24].ExploratoryresearchbyMadhavaramandLaverie[22]showedanumberofself-reportedstimulithatledtheirsubjectstopurchaseimpulseitemsontheInternet.Consistentwithresearchonbrickandmortarpurchases,manyofthesamevisceralinfluencesledtoimpulsepurchasessuchasappealoftheproduct,advertisements,andcompellingpresentation.Inaddition,theirparticipantsalsoconfirmedthatthosepeopleinagoodmoodweremorelikelytopurchaseimpulseitemsanddoingrecre-ationalbrowsing.Ratherthanexaminethemoreemotionalantecedents,wewanttoexploreamorecognitiveprocessthatmightleadpeopletopurchaseimpulseitems;thatofeaseofjustificationofthepurchase.Whilewearenotdiscountingthefactthatthecreationofanimpulsetobuyisgenerallyemotional,webelievethatonlyifcognitiveforcesimpingeonthedecision,willtheactualpurchasebemade.Inthispaper,wewillanalyzehowmentalaccounting[2527]andprovidingareasontopurchaseincreasethelikelihoodthatsomeonewillmakeanimpulsepurchase. 370S.A.Jeffrey,R.Hodge2CognitivefactorsleadingtoimpulsepurchasesAshasbeenshowninpreviousresearch[28],peopleliketohavereasonstodothings.Thisisparticularlytrueforthepurchaseofitemsviewedasmorefrivolous[29].Inthissection,wewillproposetwopotentialjustificationsforpurchasinganimpulseitemontheweb.Specifically,wewilllookattherolementalaccountingplaysindiminishingpeoplessensitivitytoadditionalpurchasesaswellastheprovisionofaspecificreasontopurchase.2.1MentalaccountingandimpulsepurchasesRatherthantreatingmoneyasafungiblecommodity,consumersoftensegregatedif-ferentsourcesandusesofcashintovariousmentalaccounts[2527,30,31].Theseaccountsgenerallyhaverelativefirmbarriersbetweenthem,makingitanon-trivialprocesstoreassignexpensestoadifferentaccount.Ratherthanchoosingtheitemsthatmaximizeoverallexpectedutility,peopleoftenbasetheirconsumptiondecisionsonthetotalamountbudgetedandtheamountremaininginthesementalaccounts.Thisimpliesthatsomeonemaybuyclothingratherthanticketstoamovieevenifthemoviewouldbemoreenjoyable,simplybecausethereissomebudgetleftforclothingbutnomoremoneylefttospendintheentertainmentaccount.Beyondthecategorizationsystem,mentalaccountinghasitsownspecificarith-meticsystemfortabulatingtheaccounts.ThisarithmeticisbasedontheProspectTheoryvaluefunction[32].Ratherthanlookingatoverallwealthstates,peopleeval-uateincomeandexpendituresasgainsandlossesfromaneutralreferencepoint.Thisneutralreferencepointisoftencurrentwealth,butcanbeatargetamount,oranyothersalientbasisofcomparison.Next,incomeandexpendituresareaffectedbydi-minishingsensitivitytothesechanges.Forexample,the100thdollarspentinastore,feelslikelessofanexpenditurethanthe10thdollarspent.Toillustrate,considerasituationwhereaconsumerismakingaUS$39,000ve-hiclepurchaseandwantstoaddanewstereotothisvehicle.WhentheconsumerisalreadyspendingUS$39,000,anadditionalUS$500forthestereorelativetotheto-talpurchaseseemsinsignificant.IftheconsumerpurchasesthecarstereowiththeUS$39,000car,thentheUS$500isaddedtothecaraccountwhichalreadycontainstheUS$39,000thecarpurchaseamount.Spendingjustanother$500.00or1.3%seemslikeasmallpurchaseduetodiminishingsensitivity.Ifhowever,theconsumerpurchasedthecarstereobyitselfatanelectronicsstore,thespendingofthis$500wouldfeelquitedifferent;itwouldfeellikeabiggeramountofmoneythanwhenitwaspairedwiththepurchaseofthevehicle.Ifthecarstereowerepurchasedonitsown,itwouldbethefirst$500theconsumerspentandintermsofmentalaccounting,theUS$500wouldbeallocatedtoastereoaccountratherthanthecaraccount.Asaresult,uponmakingthispurchase,thevalueofthiscategorystartsat$0andmovesto$500insteadofmovingfrom$39,000to$39,500.Whenconsumersaggre-gatespending,diminishingsensitivitymakesitmuchlesspainfultospendadditionalfundsinthepresenceofamoreexpensiveitem.Webelievethatconsumersaggregateallpurchasesinasingleonlinepurchasingevent,verysimilartoavisittoabrickandmortarstore.Thus,themoresomeoneis Factorsinfluencingimpulsebuyingduringanonlinepurchase371spendingonawebsite,thelesspainfuladditionalspendingwillbe.Thus,higherspendinginadvanceofseeingtheimpulseitemwillleadtoahigherlikelihoodofpurchasingtheimpulseitem.2.2SpecificreasontopurchaseResearchonreasonbasedchoicehasshownthatpeopleliketohavegoodreasonsfortheirpurchasedecisions.Forexample,researchersaskedpeopletoeitherchoosewhichoftwooptionstoreceiveorwhichoftwooptionstoreject.Theyfoundthatanenrichedoption,theitemwithmoreextremeattributevalues,wasmorelikelytobechosenandrejectedthananoptionwithlessextremeattributevalues[28].Theyclaimedthatthisphenomenonoccurredbecausewhendecidingwhichoptiontoaccept,peoplewereseekingreasonstochoosebutwhendecidingwhichoptiontoreject,decisionmakerswereseekingreasonstoreject.Othermarketingresearchhasshownthattheexistenceofacompromisechoicewillleadconsumerstobemorewillingtomakeachoiceratherthanwaitformoreinformation[33].Theexistenceofacompromiseitemprovidesareasontochoosethemiddleoption.Researchonasymmetricdominanceisanotherexampleofhavingareasontopurchase[34].TheseresearchersprovidedachoicebetweenproductX(lowerprice,lowerquality)andproductY(higherprice,higherquality).Participantswerethentoldabouta3rdproduct(thedecoy)thatwouldbedominatedbyoneofthetwoproductsbutnottheother.Theresearchshowedthatthemarketshareofthetarget(theitemwhichdominatedthedecoy)rosedependingonwhichitemthedecoytargeted.Theseresearchersclaimedthattheprovisionofadominatedalternativegaveareasontopurchasetheitemthatwasbetteronbothattributes,sincethenon-targetedoptionstillrequiredtrade-offs.RecentresearchbyStrahilevitzandMyers[35]showedanotherwaytoprovidepeopleareasontopurchaseitemsandtoencourageshopperstovisitastore.Intheirresearch,theytiedthepurchaseofspecificitemstoadonationtocharity.Theypro-posedthatthedonationprovidedareasontopurchaseanitemthatwouldotherwisenotbepurchasedduetojustificationconcerns.Inlinewiththisresearch,webelievethattheprovisionofareasontopurchaseanimpulseitemduringanonlinepurchasingtransactionwillincreasethelikelihoodthatitemwillbepurchased.3DataWetestedourhypothesesbyofferingrealconsumerstheopportunitytopurchaseanimpulseitemduringthecheckoutprocessofarealon-linepurchasingtransaction.VisitorstotheofficialwebsiteoftheHuntsvilleHighSchool(HHS)100threunionusedthissitetogetinformationaboutthereunion,purchasetheirpassestothere-union,andpurchasesouvenirs.Itemsavailablewereclothing,household,andsport-inggoodsitemswhichrangedbetweenUS$6foraglasscoffeemugandUS$45forahoodedsweatshirt.AllitemsweremarkedwiththeHuntsvilleHighSchoolreunionlogo.WebelievethatconsumersmakingpurchasesontheHHSReunionwebsitewereplacingthesepurchasesinasingleReunionexpensecategoryandthereforeallpurchaseseachpersonmakesonthiswebsitewouldbeevaluatedinanaggregatemanner. 372S.A.Jeffrey,R.Hodge3.1ProcedureVisitorstothesiteinitiatedatransactionbyplacingtheproductstheywantedinavirtualshoppingcartbyclickingtheBuyNowbutton.Ifachosenproducthadattributes(e.g.colorandsizeforaT-shirt),theusershadtospecifytheirchoices.Whendoneshopping,thevisitorsproceededtothecheckoutbyclickingtheCheck-outbutton.Thesystemthenaskedthecustomerstoeithersignintoanexistingac-countorcreateanewaccountbyenteringcontactinformationonasinglepage.Oncesignedin,thecheckoutpageshowedasummaryofcontactinformation,shoppingcartcontentsandthetotalamountowed.Whentheusersviewedthecheckoutpagetheywereofferedtheopportunitytopurchaseoneofthreeadd-onimpulseitemsallcostingUS$5.00:amousepad,eyeglassclips,oraboxofchocolatetruffles.Thispricepointallowedfortheflexi-bilitytoofferqualityproductsatanamounthighenoughtohavedifferentimpactsontheexpectedsalesamountsofUS$30,$60and$110.TheUS$5priceforourimpulseitemsrepresenteda16.7percentincreaseinamountspentforaconsumerwhoboughttheUS$30bluepass,an8.3percentincreaseintheamountspentforaconsumerwhoboughttheUS$60goldpassanda4.5percentincreaseinamountspentforaconsumerwhoboughttheUS$110principlespass.Theseitemswerenotofferedinthesouvenirssectionofthewebsite,thereforethisitemshouldnothavebeenpartoftheinitialmentalbudget,makingitatrueadditionalimpulsebuy.Theconsumerconfirmedthecontentoftheshoppingcartandaccuracyofthecon-tactinformationbyclickingontheCheckoutbuttonagainandproceededtopayfortheorder.Followingpaymenttheuserwasaskedtoanswerdemographicquestions.Toencourageparticipantstoentertheirdemographicinformation,weofferedthemanentryintoalotteryforHHSReunionsouvenirs.Totestfortheeffectofprovidingareasontopurchase,arandomizationscriptassignedparticipantstoeitheradonationconditionoracontrolcondition(nodona-tion).ThoseinthedonationconditionasawthataUS$1.00donationtothesteppingstonefoundation1wouldaccompanythepurchaseoftheimpulseitem,whiletheothergroupofparticipantssawonlytheitemsforsale.NotethatpriceoftheitemwasstillUS$5.00,inthedonationcondition.Ifthecustomerpurchasedtheimpulseoffer,itwasaddedtotheshoppingcart,thecheckoutpagewasrefreshed,andthesamedona-tionconditionwasgeneratedagain.Whileitcouldbearguedthatpeoplepurchasedtheitemjustsothattheycouldmakeadonation,weruledoutthisalternativeexpla-nationbyprovidingaseparateplaceonthewebsitetomakeadonationwithoutanattachedpurchase.3.2MeasuresThedependentvariableforthisexperimentwaswhetherornotvisitorstotheweb-sitepurchasedanimpulseitem.Totestthis,wecollectedinformationregardinghowmuchwasspentpriortoseeingtheimpulseitem,aswellaswhetherornotadonation1ThesteppingstonefoundationwasorganizedbyHuntsvilleHighSchooltogiveuniversityscholarshipstoneedyanddeservingHHSstudents,makingitahighlysalientdonationforalumnioftheschool. Factorsinfluencingimpulsebuyingduringanonlinepurchase373waslinkedtothepurchase.Becauseotherfactorsmightaffectthewillingnesstopur-chaseanimpulseitem,wealsogatheredinformationregardinghouseholdincome,age,andgenderinformation.Webelievedthathigherincomevisitorswouldbemorelikelytopurchaseanimpulseitembecauseoftheadditionaldisposableincometheyhaveavailable.Weexpectednoeffectforageorgender.4ResultsDatawerecollectedonactualpurchasetransactionsontheHHSReunionwebsitefromMarch13,2004toMay19,2004.Duringthisperiod60,413pageswereviewed,3,488uniquevisitorsattendedthesiteand771transactionswereapprovedandprocessed.Ofthese,349ordersweresubmittedinpersonorbymailandprocessedusingaseparateinterfaceandthereforewerenotexposedtotheimpulseitem.Ofthe422remainingpurchasetransactions,twoofthesetransactionsweretesttransactionsthatwereperformedtoconfirmthatthedatacollectionandonlinepaymentprocess-ingwerefunctioningproperly.Asaresult,thesetworecordswereremovedfromthedatasetleaving420records.Totestfortheeffectofmoneyspent,werequireddemographicinformation,par-ticularlyregardingincome.Onehundredandeightconsumersvisitingthesitedidnotreplytothedemographicquestionswhichreducedthenumberofusabledatapointsto312.2Uponfurtherinspectionofthedemographicresponses,onevisitorindicatedthathewasage18orlesswith47childrenundertheageof12and74peopleliv-inginthehouseholdonanincomeofunderUS$20,000.Thisrecordwaseliminatedfromthedatasetforthefirsthypothesisleaving311purchasetransactionswhichwereusedinthefinaldataanalysis.Table1showsthemean,standarddeviation,andcorrelationsofthecontrolvariablesforthefinaldataset.Table1DescriptivestatisticsandcorrelationsofindependentvariablesVariableMeanStandardAmountAgeHouseholddeviationspentincomeAmountspent90.848.36Age48.49.800.282Householdincome77.021.170.164∗∗0.075GenderMalen=1280.210∗∗∗0.1110.236∗∗∗Femalen=183∗p<0.05∗∗p<0.01∗∗∗p<0.0012Therewasnodifferenceinthefrequencyofimpulsepurchasebetweenthosewhorespondedtothede-mographicquestionsandthosewhodidnot,t(420)=0.93,p=0.35. 374S.A.Jeffrey,R.Hodge4.1ResponsivenesstomoneyspentWeusedalogisticregressiontotesttheeffectsofpriorspendingonthelikelihoodofimpulsepurchase.Theempiricalmodelshownas(1)ispTlog=α+β1·amountspent+β·controlvariables+ε,(1)1−pwhereprepresentstheprobabilityofpurchasingtheimpulseitem,amountspentistheamountofmoneyspentonpurchasespriortobeingexposedtotheimpulseitem,andthecontrolvariablesareage,income,andgenderasoutlinedabove.TheinitialmodelreportedinTable2isthefullmodelwithallofthevariablesincludedwhilethefinalmodelretainsonlythecoefficientssignificantatp<0.05.Inthefinalmodel,theamountspentpriortoseeingtheimpulseitemhadasignificantandapositiveeffectonthelikelihoodtopurchase,β=0.0114,t(308)=2.79,p<0.01.Thiscoefficienttranslatesintoa1%increaseinpurchaselikelihoodforeachUS$10increaseinpriorspending.Householdincomealsohadasignificanteffectonthelikelihoodtopurchase,β=−0.0253,t(308)=2.28,p<0.05.EstimatesbasedonfinalmodelshowninTable2:p(impulsebuy)log1−p(impulsebuy)=−2.202+0.013·amountspent−0.027·householdincome.Table2LogisticregressionresultsDependentvariable:Logoddsofbuyingimpulseitem(StandardErrorinparentheses)VariableInitialmodelFinalmodelAmountspent0.014∗∗∗0.013∗∗∗(0.004)(0.004)Age0.176(0.245)Householdincome0.025∗0.027∗(0.012)(0.011)Gender0.465(0.583)Constant4.072∗2.202∗(2.035)(0.820)χ2(4,N=312)=13.7χ2(2,N=312)=12.6R2=0.12R2=0.11∗p<0.05∗∗p<0.01∗∗∗p<0.001 Factorsinfluencingimpulsebuyingduringanonlinepurchase375Fig.1Estimatedlikelihoodofpurchaseforspecifichouseholdincomelevels(35K,65K,95K)Toillustratetheeffectofincomeandpriorexpenditures,weestimatedthelike-lihoodofimpulsepurchasebasedonthefinalmodelshowninTable2.Theseesti-matesareshowninFig.1.Bylookingatthisfigure,thepositiveimpactofamountspentcanbeseenatalllevelsofincome.Thistablealsoshowsthenegativeinfluenceofhouseholdincome,astheinterceptoftheestimateisdroppingwithincomeatallexpenditurelevels.4.2ReasonstopurchaseInourinitialexperimentaldesign,wewantedtotesttheeffectsofpresentingtheimpulseiteminapop-upwindowvs.presentingtheimpulseofferembeddedinthecheckoutpage.Wefoundnoeffectforthismanipulationandsotheseconditionsweredroppedfromfurtheranalysis.Thisleft221observationswhichpresentedtheimpulseitemembeddedinthecheckoutpage,105withthedonation,and116withoutthedonation.Weusedtheseobservationstotestwhetherlinkingthepurchasetoadonationincreasedthelikelihoodofanimpulsepurchase.Whennodonationwaslinkedtothepurchase,2people(1.7%)offeredtheimpulseitempurchasedit.Ashypothesized,thisroseto9people(8.6%)whenadonationwaslinked,t(220)=2.33,p<0.05.5DiscussionThisexperimentfoundthatastheamountspentonotherpurchasesduringanon-linetransactionincreases,consumersaremorelikelytopurchaseanadditionalitemthattheydidnotintendtopurchasewhentheyfirstvisitedthesite.Whilethisiscertainlyconsistentwithmanyexperimentsonmentalaccounting,thishasadditionalexternalvalidityasitwasperformedonvisitorstoanactivee-commercesite,spendingtheirownfunds.Sincetheimpulseitemwasofferedafterallpurchasesweremade,theseresultsprovideevidencethatimpulsepurchasescanincrementallyincreaseonlinesaleswithoutcannibalizingotherproducts.Asaresult,impulsebuyingintheonlineenvironmentisanimportantissuethatshouldbestudiedfurther.Italsosuggests 376S.A.Jeffrey,R.Hodgeanotherimportantreasonwhywebsitedesignersshouldtrytomaximizeconsumerpurchaseswhileonthewebsite.Aswebsitevisitorsspendmoremoney,theyarelessresistanttospendingevenmore.Webelievedthatthosewithhigherincomeswouldbemorelikelytobuyimpulseitems,sincetheirdiscretionaryincomewouldbehigher.However,wefoundthatin-comeactuallyhadanegativeeffectonimpulsepurchases;asincomesrose,peoplebecamelesslikelytobuyonimpulse.OnetheorythatcouldpotentiallyexplainwhythisoccurredisthathigherincomeconsumersdonotbuyjunkyUS$5items,theymaybemorelikelytobuyahigherpriced,higherqualitysouvenirs.Highincomeconsumersmayalsoseelimitedvalueinamousepad,eyeglassclip,orchocolatetruffles.Whilethisisspeculation,webelieveitisimportantforfutureresearchtoex-aminetheimpactofincomeonimpulsepurchases.Thisfindingisconsistenthoweverwiththespeculationthatwealthyindividualsachievedtheirfinancialstatusbylivingwithintheirmeansandavoidingimpulsepurchasessuchastheoneweoffered[36].Thisstudyalsoshowsthatofferingareasontopurchaseisaneffectivewaytoincreaseimpulsepurchases.LinkingaUS$1donationtotheimpulseitem,therebyprovidingareasontopurchase,increasedthefrequencyoftheimpulsepurchasefromaround2%toalmost9%.Itmayseemthatthebaselineimpulsepurchaseof2%seemsquitelow,especiallyconsideringworkbyStern[1]thatshowedthatalmost50%ofpurchasesaredoneonimpulse.Webelievethisisareasonablenumberfortworeasons.First,weareonlystudyingpureandsuggestionimpulsepurchases[1,4].Othertypesofimpulsepurchases,notablyreminderandplannedmightaccountforalargenumberofimpulsepurchases.Inaddition,otherresearchhasshownthatlessimpulsepurchasesaremadeonlineduetotheabsenceofasalespersontopressurethesale[22].Whiletheincreasefrom2%to9%mayappearsmall,thisisapotentialincreaseofUS$5Bintotalrevenue,giventheUS$69Billionspentonlinein2004[6].Althoughwestudiedthisphenomenoninanonlineenvironment,weseenocom-pellingreasonwhywewouldnotseetheexactsameresultsinatraditionalretailen-vironment.Thediminishingsensitivitytoadditionalexpendituresmayevenbemorepronounced,sinceatriptoastoremaybemoreclearlysegregatedinaconsumersmind.Wealsobelievelinkingadonationtoimpulseitemsdisplayedattheregis-tershouldalsoinspireincreasedpurchaseofthoseitemsinbrickandmortarretaillocations.Infact,thisisprobablywhydonationboxesaregenerallylocatedatthecheckout.5.1ImportantdifferencesintheonlineenvironmentTherearesomeimportantdifferencesbetweenpurchasingonlineandpurchasinginabrickandmortarenvironment.Storelayoutsarequitedifferentthanawebsitesdesign.Thesitemustbeeasilynavigableandtheonlineretailermustbetrustworthy[37,38].Inadditiontothesebasicdifferences,onlineretailersdifferfromtheirbrickandmortarcounterpartswhenitcomestoimpulsebuying.Forexample,inthebrick-and-mortarenvironment,heavyorawkwarditemsarenottypicallyusedasimpulseitemsbecauseofthedifficultyconsumersmayenvisioningettingithome.Heavyorawkwarditemsmaybemoreeffectiveasonlineimpulseitemsbecauseallonlinepur-chasesaredeliveredtotheuserandtherefore,theconsumerwouldnothavetoworry Factorsinfluencingimpulsebuyingduringanonlinepurchase377abouttheweightorawkwardnatureofitems.Asaresult,theonlineenvironmentmayreducerestrictionsinthetypesofproductsthatcanbeofferedforimpulse.However,transportationcostssuchastheseareadoubleedgedsword.Onlineconsumersneedtoworryaboutshippingandhandlingcostswhichcouldsignificantlyraisethecostofthesegoods,reducingtheireffectivenessasanimpulseitem.Anadditionaldrawbacktoonlineimpulsepurchasesisthefactthattherearenoclerkstosuggestimpulseitems.Asaresult,influencingtheimpulsepurchasemaybeamoredifficulttaskonline[22].Inaddition,thedelayofreceivingthepurchasemaydissuadeshoppersfrommakingimpulsepurchasesonline,particularlysinceimpulsepurchasesaregenerallydesignedtosatisfyimmediatedesires.Forexample,whenaconsumerisstandinginlinewaitingtopayfortheirgroceries,thesightofgumatthecheckoutmaymaketheirmouthwater,increasingthevisceraldesirefortheitemandinducingthemtobuytheitemonimpulse[39,40].Thisisalsoatwo-edgedsword,sinceeventhoughthereisnoimmediatereductioninthedrivetoconsumetheimpulseitem,someitemsmayincreaseutilitythroughtheactofsavoringitpriortoitsarrival[41].5.2Limitations&futureresearchOnelimitationofthisexperimentwasthattheparticipantsrepresentedarestrictedsampleofthepopulation.Thiswasanot-for-profitwebsiteforahighschoolreunionwherethedonationwastargetedtoaschool-orientedfund.Becauseofthis,itislikelythatthischaritywasmoremeaningfultotheparticipants,increasingthelikelihoodofimpulsepurchasemorethanalessrelevantdonationwouldhavedone.Findingsuchasalientcharityforalargeraudiencewouldbemoredifficult,makingtheprovisionofareasonsuchasthismoredifficult.Furthermore,thisresearchwasnotconductedonatraditionalretailwebsiteandthismaylimitthegeneralizabilityofthefindings.Toresolvethis,similarmanipula-tionsshouldbeattemptedonamorepermanent,e-commercewebsite.Thisresearchisalsolimitedinthetypesofimpulsepurchasesstudied.Westudiedpureandsug-gestionimpulsepurchasing,futureresearchshouldexaminethefactorsthatleadtoreminderandplannedimpulsepurchases.Finally,weonlyexaminedmentalaccount-ingandreasontopurchasereasonsforimpulsepurchases,andthereareundoubtedlymanyothercauses.Wealsobelievethatfutureresearchshouldconsidertheprocessbywhichretailwebsitesselecttheproductstouseasimpulseitems.Largeon-lineretailerssuchasAmazon.comusepasttransactionhistorytopairupbooksfrequentlyboughttogether,thuscreatingopportunitiesforsuggestionimpulsepurchases.Futureresearchshouldexamineexistingworkonthecharacteristicsofimpulseitemsinbrickandmortarstores[1]andattempttoapplythemtotheonlineenvironment.6ConclusionThispaperexaminedimpulsepurchasingduringonlinetransactionswithrespecttotheamountspentonthesiteandtheprovisionofareasontopurchase.Consistent 378S.A.Jeffrey,R.Hodgewithmentalaccountingandthepsychophysicsofpricing,theresultsindicatethatthelikelihoodofanimpulsepurchaseispositivelycorrelatedwiththeamountspentonthewebsitepriortoseeinganimpulseitem.Themoresomeonespendsonawebsitepriortoseeinganimpulseitem,themorelikelytheyaretomakeanimpulsepurchase.Inaddition,thestudyreportedherefoundthatbylinkingtheimpulsepurchasetoarelevantcharitabledonationincreasedthefrequencyofthatimpulsepurchase.Wehypothesizethatthisisbecausethelinktoacharityprovidedaneasytojustifyreasontopurchasetheitem,eveniftheitemwasnotneeded.WithspendingonthewebatUS$69Bandgrowingat20%ayear,wefeelthatimpulsepurchasesine-commerceisanareathatdeservesmorescholarlystudy.Withanunderstandingofthefactorsthatinfluenceimpulsepurchasesintheonlineen-vironmentandminimaladditionalinvestment,onlineretailerscancapitalizeonthislucrativeopportunity.AcknowledgementsTheauthorswishtothanktheHuntsvilleHighSchool100thAnniversaryorganiz-ingcommitteefortheirsupport.References1.Stern,H.(1962).Thesignificanceofimpulsebuyingtoday.JournalofMarketing,26(2),5962.2.Bellenger,D.N.,Robertson,D.H.,&Hirschman,E.C.(1978).Impulsebuyingvariesbyproduct.JournalofAdvertisingResearch,18(6),1518.3.Benady,D.(2003).Sweettemptationleavessourtaste.InMarketingweek(p.23).4.Miller,R.(2001).In-storeimpactonimpulseshoppers.InMarketing(pp.2728).5.UnitedStatesDepartmentofCommerce.(2001).InE-commercemulti-sectorreport(pp.18).Wash-ington:UnitedStatesDepartmentofCommerce.6.UnitedStatesDepartmentofCommerce.(2005).InQuarterlyretaile-commercesales(p.2).Wash-ington:UnitedStatesDepartmentofCommerce.7.Cobb,C.J.,&Hoyer,W.D.(1986).Plannedversusimpulsepurchasebehavior.JournalofRetailing,62(4),384409.8.Kollat,D.T.,&Willet,R.P.(1967).Customerimpulsepurchasingbehavior.JournalofMarketingResearch,4(1),2131.9.Rook,D.W.(1987).Thebuyingimpulse.JournalofConsumerResearch,14(2),189199.10.Piron,F.(1991).Definingimpulsepurchasing.AdvancesinConsumerResearch,18,509514.11.Phillips,H.,&Bradshaw,R.(1993).Howconsumersactuallyshop:customerinteractionwiththepointofsale.InternationalJournalofMarketingResearch,35(1),5162.12.Iyer,E.S.(1989).Unplannedpurchasingknowledgeofshoppingenvironmentandtimepressure.JournalofRetailing,65(1),4057.13.Jones,M.A.,Reynolds,K.E.,Weun,S.,&Beatty,S.E.(2003).Theproduct-specificnatureofimpulsebuyingtendency.JournalofBusinessResearch,56(7),505511.14.Bayley,G.,&Nancarrow,C.(1998).Impulsepurchasing:aqualitativeexplorationofthephenom-enon.QualitativeMarketResearch,1(2),99117.15.Verplanken,B.,&Herabadi,A.(2001).Individualdifferencesinimpulsebuyingtendency:feelingandnothinking.EuropeanJournalofPersonality,15(1),S71S83.16.Agee,T.,&Martin,B.A.S.(2001).Plannedorimpulsepurchases?Howtocreateeffectiveinfomer-cials.JournalofAdvertisingResearch,41(6),3542.17.Palmer,J.W.(2002).Websiteusability,design,andperformancemetrics.InformationSystemsRe-search,13(2),151167.18.Kumar,M.,Rangachari,A.,Jhingran,A.,&Mohan,R.(2000).SalespromotionsontheInternet.YorktownHeights:IBMResearchDivision19.Ling,M.C.-H.,&Lawler,K.(2001).Internetadvertising,gametheory,andconsumerwelfare.Elec-tronicCommerceResearch,1(12),169181. Factorsinfluencingimpulsebuyingduringanonlinepurchase37920.Lee,P.(2002).Behavioralmodelofonlinepurchasersine-commerceenvironment.ElectronicCom-merceResearch,2(12),7585.21.Koufaris,M.,Kambil,A.,&LaBarbera,P.A.(2001).ConsumerbehaviorinWeb-basedcommerce:Anempiricalstudy.InternationalJournalofElectronicCommerce,6(2),115138.22.Madhavaram,S.R.,&Laverie,D.A.(2003).ExploringimpulsepurchasingontheInternet.AdvancesinConsumerResearch,31,59.23.Rook,D.W.,&Gardener,M.P.(1993).Inthemood:Impulsebuyingsaffectiveantecedents.InArnold-Costa,J.,&Belk,R.W.(Eds.).ResearchinConsumerBehavior(pp.128).Greenwich:JAIPress.24.Beatty,S.E.,&Ferrel,M.E.(1998).Impulsebuying:modelingitsprecursors.JournalofRetailing,74(2),169191.25.Thaler,R.H.(1980).Towardsapositivetheoryofconsumerchoice.JournalofEconomicBehaviorandOrganization,1(1),3960.26.Thaler,R.H.(1985).Mentalaccountingandconsumerchoice.MarketingScience,4(3),199214.27.Thaler,R.H.(1999).Mentalaccountingmatters.JournalofBehavioralDecisionMaking,12(3),183206.28.Shafir,E.,Simonson,I.,&Tversky,A.(1993).Reason-basedchoice.Cognition,49(12),1136.29.Kivetz,R.,&Simonson,I.(2002).Earningtherighttoindulge:effortasadeterminantofcustomerpreferencestowardfrequencyprogramrewards.JournalofMarketingResearch,39(2),155170.30.Henderson,P.W.,&Peterson,R.A.(1992).Mentalaccountingandcategorization.OrganizationalBehaviorandHumanDecisionProcesses,51(1),52117.31.Heath,C.,&Soll,J.B.(1996).Mentalbudgetingandconsumerdecisions.JournalofConsumerResearch,23(1),4052.32.Kahneman,D.,&Tversky,A.(1979).Prospecttheory:ananalysisofdecisionunderrisk.Economet-rica,47(2),263291.33.Simonson,I.(1989).Choicebasedonreasons:thecaseofattractionandcompromise.JournalofConsumerResearch,16(2),158174.34.Huber,J.,Payne,J.,&Puto,C.(1982).Addingasymetricallydominatedalternativesviolationsofregularityandthesimilarityhypothesis.JournalofConsumerResearch,9(1),9098.35.Strahilevitz,M.,&Myers,J.G.(1998).Donationstocharityaspurchaseincentives:howwelltheyworkmaydependonwhatyouaretryingtosell.JournalofConsumerResearch,24(4),434446.36.Stanley,T.J.,&Danko,W.D.(1996).Themillionairenextdoor:thesurprisingsecretsofAmerica’swealthy.Atlanta:LongstreetPress.37.Schlosser,A.E.,White,T.B.,&Lloyd,S.M.(2006).Convertingwebsitevisitorsintobuyers:howwebsiteinvestmentincreasesconsumertrustingbeliefsandonlinepurchaseintentions.JournalofMarketing,70(2),133148.38.Janiszewski,C.(1998).Theinfluenceofdisplaycharacteristicsonvisualexploratorysearchbehavior.JournalofConsumerResearch,25(3),290301.39.Loewenstein,G.(1996).Outofcontrol:Visceralinfluencesonbehavior.OrganizationalBehaviorandHumanDecisionProcesses,65(3),272292.40.Loewenstein,G.,&Thaler,R.H.(1989).Intertemporalchoice.JournalofEconomicPerspectives,3(4),181193.41.Loewenstein,G.(1987).Anticipationandthevaluationofdelayedconsumption.EconomicJournal,97(387),666684.ScottA.JeffreyisanassistantprofessorintheDepartmentofManagementSciencesattheUniversityofWaterloo,Canada.HehasaPh.D.inManagerialandOrganizationalBehaviorfromtheUniversityofChicagoGraduateSchoolofBusiness.Dr.Jeffreyconductsresearchonindividualdecisionmakingandemployeeperformance.RebeccaHodgeMASc.,receivedherMastersofAppliedSciencein2005fromtheDepartmentofMan-agementSciencesattheUniversityofWaterloo.SheiscurrentlyapartnerinthewebdesignfirmofE-XYNWebSolutions.E-XYNassistsclientsbyimplementingmoreeffectivewaystomanagebusinessandcustomersthroughtheefficientuseoftechnologyandtheweb.